Improved catalysts for the iridium-catalyzed asymmetric isomerization of primary allylic alcohols based on charton analysis

Article abstract of DOI:10.1002/chem.201001311

An improved generation of chiral cationic iridium catalysts for the asymmetric isomerization of primary allylic alcohols is disclosed. The design of these air-stable complexes relied on the preliminary mechanistic information available, and on Charton analyses using two preceding generations of iridium catalysts developed for this highly challenging transformation. Sterically unbiased chiral aldehydes that were not accessible previously have been obtained with high levels of enantioselectivity, thus validating the initial hypothesis regarding the selected ligand-design elements. A rationale for the high enantioselectivities achieved in most cases is also presented. Achieving enantioselectivity: An improved generation of chiral cationic iridium catalysts for the asymmetric isomerization of primary allylic alcohols is disclosed. The design of these air-stable complexes relies on preliminary mechanistic information and on Charton analyses using two preceding generations of iridium catalysts developed for this highly challenging transformation (see figure).

Full text of DOI:10.1002/chem.201001311

DOI: 10.1002/chem.201001311
Improved Catalysts for the Iridium-Catalyzed Asymmetric Isomerization of
Primary Allylic Alcohols Based on Charton Analysis
Luca Mantilli,[a] David Gꢀrard,[a] Sonya Torche,[a] Cꢀline Besnard,[b] and
Clꢀment Mazet*[a]
Abstract: An improved generation of
chiral cationic iridium catalysts for the
asymmetric isomerization of primary
allylic alcohols is disclosed. The design
of these air-stable complexes relied on
the preliminary mechanistic informa-
tion available, and on Charton analyses
using two preceding generations of iri-
dium catalysts developed for this
highly challenging transformation. Ster-
ically unbiased chiral aldehydes that
were not accessible previously have
been obtained with high levels of enan-
tioselectivity, thus validating the initial
hypothesis regarding the selected
ligand-design elements. A rationale for
the high enantioselectivities achieved
in most cases is also presented.
Keywords: asymmetric catalysis
·
Charton analysis · iridium · isomeri-
zation · ligand design
Introduction
ment.[3] In 2008, when we initiated a research program in
this challenging area, the best catalysts reported in the liter-
ature were incontestably the cationic planar chiral phospha-
ferrocene–rhodium complexes developed by Fu and co-
workers (Figure 1).[2b,4,5] Whereas catalyst A performed well
only for Z-configured primary allylic alcohols, catalyst B,
with bulkier o-tolyl phosphine substituents, appeared more
general and displayed interesting catalytic activities for both
E and Z isomers. Promising to satisfactory levels of enantio-
selectivity (50–80% ee) could be obtained for most sub-
strates surveyed, and very good enantioselectivities
The Rh-catalyzed asymmetric isomerization of primary allyl-
ic amines into the corresponding chiral enamines distin-
guishes itself as one of the very few transition-metal-cata-
lyzed asymmetric reactions that have been successfully
brought to the industrial scale. The high enantioselectivities,
mild reaction conditions, wide substrate scope, high turn-
over-numbers (TON), accessibility and recyclability of the
catalyst, and a detailed understanding of the reaction mech-
anism are parameters that have all favored the implementa-
tion of this reaction as the key step in the yearly production
of high tonnages of enantiopure (ꢀ)-menthol by the Takasa-
go Company in Japan.[1,2] In sharp contrast, and perhaps sur-
prisingly as both reactions were initially investigated by the
same research groups at the same time, the asymmetric iso-
merization of primary allylic alcohols into the corresponding
chiral aldehydes has not reached the same level of achieve-
AHCTUNGTREG(NNNU >90% ee) were obtained for the sterically more demanding
derivatives. Thorough mechanistic studies showed that this
rhodium-catalyzed asymmetric isomerization reaction pro-
ceeds through an intramolecular 1,3-hydrogen migration
pathway. Despite unprecedented reactivity and selectivity
patterns, these rhodium catalysts suffer from several impor-
tant limitations, which have hampered their dissemination
into the synthetic community: 1) practical quantities of
enantiomerically pure ligands were only accessible after iter-
ative preparative HPLC separations; 2) the rhodium com-
plexes needed to be pre-activated by hydrogenating off the
diolefin ancillary ligand, before the reaction was run at tem-
peratures higher than the solvent boiling point (typically at
100–1508C in THF); 3) despite these forcing conditions, the
catalyst activity remained relatively low, and typical reaction
times ranged from 24 to 48 h, with yields varying from 60–
90%; 4) with one exception, the substrate scope was limited
to primary allylic alcohols combining an aryl ring with an
alkyl substituent.
[a] L. Mantilli, D. Gꢀrard, S. Torche, Dr. C. Mazet
Department of Organic Chemistry
University of Geneva, Quai Ernest Ansermet 30
1211 Geneva 4 (Switzerland)
Fax : (+41)22-379-3215
E-mail: clement.mazet@unige.ch
[b] Dr. C. Besnard
Laboratoire de Cristallographie
University of Geneva, Quai Ernest Ansermet 24
1211 Geneva 4 (Switzerland)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001311.
12736
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12736 – 12745

FULL PAPER
Figure 1. Successful catalysts and ligands for the (asymmetric) isomerization of primary allylic alcohols (left). Appropriate experimental conditions for
either the exclusive hydrogenation or isomerization reaction using C (right).
With these clear limitations for the rhodium-catalyzed
asymmetric isomerization of primary allylic alcohols, we
sought to develop readily accessible catalysts that would
follow a fundamentally different mechanistic path, and
which would allow us to work under experimental condi-
tions more suitable for the development of an asymmetric
variant of the isomerization reaction. The basis of our rea-
soning originated from the general observation that olefin
isomerization is the most common side-reaction in (asym-
metric) hydrogenation processes, the undesired competing
isomerization being usually suppressed by increasing the
molecular hydrogen pressure.[6] Inspired by precedents in
the literature, we initially hypothesized that, under appropri-
ate experimental conditions, selected chiral hydrogenation
catalysts may be detracted from their initial task, and could
exclusively perform the asymmetric isomerization reactions
instead. In 1978, Baudry and Ephritikhine showed that the
cationic bis-phosphine–iridium hydrogenation catalyst
olefins lacking a proximal functional group to steer the hy-
drogenation reaction upon chelation to the metal center),
outperforming the most potent rhodium and ruthenium cat-
alysts for this particular class of substrates.[10] A few years
later, Stork and Kahne[11] and Crabtree and Davis[12] inde-
pendently demonstrated that the same iridium catalyst was
highly efficient in the directed hydrogenation of primary al-
lylic and homoallylic alcohols. A high degree of stereocon-
trol in the hydrogenation of various cyclohexenol derivatives
served to establish the two-point binding mode adopted by
the substrate during catalysis. Only low loadings of C were
usually necessary to reduce quantitatively a variety of di-
versely substituted allylic alcohols at room temperature.
Owing to the very high catalytic activity, and, moreover, to
the virtual insensitivity of the catalyst to the degree of sub-
stitution of the double bond of the allylic alcohols, this
useful methodology has become a standard tool for the syn-
thetic chemist, as emphasized by its numerous utilizations in
the elaboration of complex molecules since its discovery.[13]
Building on these precedents, we recently established that
Crabtreeꢂs catalyst is indeed a very general catalyst for the
isomerization of primary allylic alcohols, given that an ex-
perimental protocol analogous to that developed by Baudry
and Ephritikhine is employed.[14] Much like in the directed
hydrogenation reaction, the scope of the related isomeriza-
tion reaction turned out to be extremely broad, and allylic
alcohols having a di-, tri-, or tetrasubstituted double bond
were isomerized quantitatively under very mild conditions
using low loadings of C. While optimizing the experimental
protocol to promote exclusively the isomerization pathway,
we identified crucial electronic features responsible for the
catalytic activity. Any variation from the original structure
of Crabtreeꢂs catalyst, which combines a trialkylphosphine
with an sp2-hybridized N-donor ligand, led to a complete
loss of catalytic activity. The use of the less coordinating,
bulky, and lipophilic BArF anion provided enhanced catalyst
activity compared to the original PF6 analogue (BArF =
tetrakis-[3,5-bis-(trifluoromethyl)-phenyl]borate). The high
[(Me(Ph)2P)2IrACHTUNGTRENNUNG(cod)]PF6 was promoting exclusively the iso-
merization of some primary allylic alcohols to aldehydes
(cod=1,5-cyclooctadiene).[7] In their experimental proce-
dure, the iridium precatalyst had to be activated by molecu-
lar hydrogen, and the solution was subsequently degassed to
avoid competing hydrogenation. The catalyst displayed in-
teresting TON for substrates having a mono- or disubstitut-
ed double bond, but was ineffective for allylic alcohols with
higher substitution patterns (i.e., potential prochiral sub-
strates). More recently, Fehr and Farris observed that Crab-
treeꢂs hydrogenation catalyst C (Figure 1) promoted an ex-
clusive isomerization/lactolization sequence rather than the
initially expected hydrogenation, even though a hydrogen
atmosphere was maintained.[8,9] The observation by Baudry
and Ephritikhine that C was also competent at promoting
the isomerization of allylic alcohols, albeit with reproducibil-
ity issues, suggested the isomerization/lactolization sequence
discovered by Fehr and Farris may be very substrate specif-
ic. Crabtreeꢂs catalyst C was initially discovered in the late
1970s for the hydrogenation of unfunctionalized olefins (i.e.,
Chem. Eur. J. 2010, 16, 12736 – 12745
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12737

C. Mazet et al.
degree of reproducibility and the substrate scope generality
contrasted strongly with earlier observations, and were at-
tributed to the increased chemical stability of the catalyst
also provided by the counteranion.[15,16] These electronic
prerequisites were instrumental in identifying two genera-
tions of isomeric chiral (dialkyl)phosphinoalkyloxazoline iri-
dium complexes for the asymmetric version of the isomeri-
zation reaction.[17,18] Ligand scaffolds 1 and 2 introduced by
Helmchen[19] and Burgess,[20] respectively, in the context of
Pd-catalyzed asymmetric allylic alkylations, were tuned ac-
cording to the design elements identified with the achiral
catalyst C (Figure 1). Under similarly mild reaction condi-
tions, using the same activation protocol, the best candidates
of each generation (1a and 2a) displayed catalytic activities
approaching that of C for 3,3-disubstituted primary allylic
alcohols. This was accompanied by unprecedented levels of
enantioselectivity in the case of E-configured substrates
combining a large alkyl substituent with an aryl group.
When a smaller alkyl substituent was combined with the
aryl group, lower enantioselectivity along with competing E/
Z isomerization of the starting material were observed. Al-
lylic alcohols with a Z configuration were much less reac-
tive, and delivered the chiral aldehyde of opposite absolute
configuration with substantially lower enantioselectivities.
Concomitant isomerization of the allylic alcohol double
bond was also observed for this substrate class. Catalyst 2a
was found to outperform catalyst 1a in the isomerization of
the more challenging 3,3-dialkyl primary allylic alcohols.
Tetrasubstituted or 2,3-disubstituted primary allylic alcohols,
homoallylic alcohols, and secondary allylic alcohols were
not isomerized with these iridium catalysts.
of NMR analyses, labeling and cross-over experiments, an
unprecedented intermolecular dihydride mechanism was
proposed to be operative (Figure 2). The conformational
binding of the allylic alcohol to the active dihydride iridium-
1
N
troscopy is crucial in the general outcome of the reaction. In
the productive isomerization pathway, the migratory inser-
tion at C2 is believed to be both the rate- and the enantio-
determining step. Competing migratory insertion at C3 is
postulated to be responsible for the E/Z isomerization ob-
served for E-configured 3,3-disubstituted allylic alcohols
with a small alkyl substituent and for Z-configured sub-
strates. The lower enantioselectivities measured in these
cases were attributed to the difference in rate between the
multiple competing isomerization pathways (E!Z!E; E!
chiral aldehyde; Z!ent-chiral aldehyde). Similar observa-
tions were made during isomerization of the model substrate
when ligands of type 1 or 2 with small alkyl groups on the P
atom were employed, which emphasize the strong depend-
ency of the isomerization reaction with respect to steric fac-
tors arising from the ligand structure and/or the substrate
structure.[21]
Despite these significant advances in the asymmetric iso-
merization of primary allylic alcohols, in particular regarding
the accessibility of the catalyst and the mild reaction condi-
tions, expanding the scope to substrates that are not biased
with a large alkyl group remains particularly challenging. In
this article, we describe in detail the rationale we followed
for the development of a new generation of iridium catalysts
for the highly enantioselective isomerization of primary al-
lylic alcohols that were not successfully isomerized using the
first generations of catalysts.
Mechanism: Preliminary mechanistic investigations have
shed light on the mechanism at play in the asymmetric iso-
merization of primary allylic alcohols using chiral (dialkyl)-
phosphinoalkyloxazoline iridium complexes.[17] On the basis
Figure 2. Mechanism for the isomerization of primary allylic alcohols using iridium catalysts with ligands of type 1 or 2.
12738
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12736 – 12745

Iridium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
Results and Discussion
Ligand and catalyst design: Miller and Sigman have recently
described the possibility to elaborate the linear free energy
relationship (LFER) between steric parameters (Charton
value) and enantiomeric ratios (e.r.) for a variety of well-es-
tablished catalytic asymmetric reactions, thus demonstrating
that the substituent effects on the catalyst and/or the sub-
strate can be correlated.[22,23] These Charton analyses pro-
vide insight into the transition state with respect to steric
factors and offer an attractive complement to Hammet
plots, which only give a measurement of the electronic ef-
fects. More importantly, Miller and Sigman anticipated
these correlations may ultimately serve as a new platform
for catalyst structure optimization. Indeed, despite impor-
tant advances in computational methods, ligand design in
asymmetric catalysis remains a challenging task that relies
mostly on screening and intuition, in particular in the ab-
sence of any mechanistic information. Figure 3 shows Char-
ton analyses for the isomerization of two different classes of
E-configured 3,3-disubstituted primary allylic alcohols ob-
tained using our first two generations of iridium catalysts 1
and 2.[17,18] The excellent correlation in the first plot indi-
cates that E-configured 3,3-disubstituted primary allylic al-
cohols, combining a phenyl substituent with alkyl groups of
various volumes, are isomerized via a similar transition state
using the most selective catalyst of the first generation 1a.
The value of the sensitivity factor (i.e., slope; Y=2.32) re-
veals that variation in e.r. as a function of the steric demand
is strongly marked for this substrate class. Consequently, for
alkyl groups smaller than isopropyl, the enantiomeric ratio
is sharply reduced (ee=90%!log
ACHTUNGTREN(NUGN 95/5)=1.278). The
second plot shows that a good correlation is also obtained
for 3,3-dialkyl substituted primary allylic alcohols using the
optimal structure of second generation 2a, suggesting that
again the reaction follows a similar mechanism for all sub-
strates investigated. Importantly, 3,3-dialkyl-substituted pri-
mary allylic alcohols are notoriously more challenging sub-
strates for isomerization. They barely reacted using the first
generation of catalysts (type 1). Although the enantioselec-
tivity values are lower in this case, the lower sensitivity
value (Y=0.90) indicates a less pronounced dependence of
this substrate class on steric variations using catalyst 2a.
On the basis of these Charton analyses, and the observa-
tion that an increased steric demand on the substrate and/or
the ligand steers the isomerization reaction towards the pro-
ductive pathway, we sought to design new catalysts that
would perform equally well for sterically biased and un-
biased allylic alcohols by transferring the steric demand ex-
clusively on the ligand structure. Nevertheless, we anticipat-
ed that a potential risk in following such an approach would
be the complete hampering of substrate coordination to the
metal. Starting from the more general scaffold 2, we rea-
soned that bringing the P-alkyl groups and the oxazoline
substituent closer to the coordination sites where the reac-
tion takes place (by a formal homologation of the ligand
bridge) would have a beneficial impact on the enantioselec-
Figure 3. Charton analysis for the isomerization of aromatic primary al-
lylic alcohols using 1a (top) and 3,3-dialkyl primary allylic alcohols using
2a (bottom).
tivity of the isomerization reaction, while maintaining
enough flexibility in the ligand backbone to ensure substrate
coordination (Scheme 1).
Both ligand scaffolds 2 and 3 were originally designed by
Burgess and co-workers in the context of Pd-catalyzed
asymmetric alkylation reactions.[19,24] Ligands L3 (JM-phos
ligands) were also used later in the Ir-catalyzed asymmetric
Scheme 1. Conceptual approach for catalyst optimization based on the
reaction mechanism rationale and the Charton analyses.
Chem. Eur. J. 2010, 16, 12736 – 12745
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12739

C. Mazet et al.
hydrogenation of unfunctionalized alkenes.[25] Exploiting the
virtually infinite catalogue of carboxylic acid precursors, sys-
tematic variation of the exocyclic substituent of the oxazo-
line ring served to demonstrate the highly modular nature
of these structures. Nevertheless, only diarylphosphine moi-
eties were implemented on the left-hand side of the ligand.
The presence of a formal trialkylphosphine along with an
sp2-hybridized N-donor being a necessary requirement to
ensure catalytic activity in the isomerization of primary al-
lylic alcohols,[14,17,18] we focused on the synthesis of a struc-
turally relevant library of nine new iridium complexes, all
possessing a formal trialkyl-P donor (R1 =Cy, tBu, 1-ada-
mantyl (1-Ad)) combined with oxazoline rings having either
C-sp2 or C-sp3 exocyclic substituents (R2 =1-Ad or aryl)
(Scheme 2).
dition of a slight excess of NaBArF followed by vigorous stir-
ring for one hour. All iridium complexes were purified by
standard silica gel chromatography, and isolated in pure
form in average to excellent yields depending on the R1/R2
combination. The red–orange solids were all bench-stable,
but were kept in a ꢀ358C freezer under an inert atmos-
phere. No loss of catalytic activity was observed even after
several months.
Catalyst structure optimization: In a preliminary survey, the
six new iridium complexes 3a–f were evaluated for their
ability to catalyze the isomerization of (E)-4-methyl-3-phe-
nylpent-2-enol, a model substrate that has proven relevant
in previous studies. The reactions were performed using the
standard protocol for the isomerization of primary allylic al-
cohols recently developed in our laboratory. The iridium
complexes were activated by bubbling molecular hydrogen
directly through a solution of the complexes in THF for five
minutes. The substrate was added only after complete de-
gassing of the excess of hydrogen to avoid undesired hydro-
genation. The reactions were run for 4 to 14 h at room tem-
perature using 5 mol% of iridium precatalyst 3a–f (Table 1,
entries 1–8).
The reactions with catalysts 3a–c, which displayed very
low reactivities, were run for 14 h to obtain sufficient mate-
rial to facilitate the subsequent analyses. In this series (R1 =
alkyl, R2 =1-Ad), both the yield and the enantiomeric
excess decreased as the size of the substituents on the phos-
phorus atom increased. Indeed, whereas catalyst 3a deliv-
ered the chiral aldehyde in 41% yield and 93% ee, catalysts
3b and 3c gave the isomerization product in substantially
reduced yields (28 and 17%, respectively) and lower enan-
tioselectivity values (84 and 26%, respectively). Catalysts
3d–f (R1 =alkyl, R2 =Ph) were more reactive, and the reac-
tions were run for only 4 h. Again, the yields were reduced
as the steric demand of the P-alkyl substituents was in-
creased, but the enantioselectivity followed an inverse
trend. All three catalysts gave
Protected ligand and iridium complex syntheses: Access to
diversely substituted tosyl-oxazoline derivatives was ach-
ieved using the most recent four-step procedure reported by
Burgess and co-workers, starting from inexpensive l-aspartic
acid.[26] The modular combination of these five oxazoline
electrophiles with three different borane-protected dialkyl-
phosphines readily gave access to a library of six new (P,N)
ligands (L3a–f). Ligands L3g–i were synthesized after initial
evaluation of the catalysts in the asymmetric isomerization
of our model substrate (vide infra). SN2 displacement of the
tosyl anion by reaction with in situ generated borane-pro-
tected lithium dialkylphosphanide precursors in THF at
ꢀ788C yielded the protected ligands as air-stable white
solids in moderate to good yields after 48 h. The one-pot de-
protection/complexation sequence recently optimized in our
laboratory was carried out next. Deprotection of the phos-
phine was secured by reacting the borane adducts with an
excess of diethylamine at reflux in a sealed tube. After evap-
oration of the volatiles, half an equivalent of [{IrClACHTNUTRGNEUNG(cod)}2]
was added in dichloromethane and heated at reflux for a
further two hours. Halide abstraction was performed by ad-
the desired product with excel-
lent to almost perfect enantio-
selectivity levels (90, 97, and
98% ee, respectively); catalyst
3e (R1 =tBu, R2 =Ph) offered
the best balance between reac-
tivity and selectivity. Because
the flexibility of the general
synthetic route developed by
Burgess and co-workers allows
straightforward structural opti-
mization, modification of the
electronic and steric properties
of the phenyl group (R2) in 3e
was carried out next, while
maintaining tert-butyl substitu-
ents on the P atom. Complexes
3g–i were prepared in modest
Scheme 2. Synthetic route for the synthesis of the iridium precatalysts 3a–i.
to moderate yields (Scheme 2)
12740
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12736 – 12745

Iridium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
Table 1. Catalyst evaluation through asymmetric isomerization of (E)-4-
Table 2. Asymmetric isomerization of E and Z-configured aromatic pri-
methyl-3-phenylpent-2-enol.[a]
mary allylic alcohols.[a,b]
Catalyst[b]
[(P,N)Ir(cod)]BArF
t
Yield
[%][c]
ee
G
G
[h]
[%][d]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
(S)-1a
(S)-2a
(S)-3a
(S)-3b
(S)-3c
(S)-3d
(S)-3e
(S)-3 f
(S)-3g
(S)-3h
(S)-3i
(S)-3i
(S)-3i
(S)-3i
22
4
14
14
14
4
4
4
4
4
75
73
41
28
17
60
57
30
<5
11
43
65
52
64
97 (S)[e]
R1
R2
Olefin
config
Yield
[%][c]
ee
96 (R)[f]
93 (R)
[%][d]
1
2
3
4
5
6
7
8
9
10
11
12
13
tBu
iBu
Cy
iPr
Et
Me
iPr
iPr
Ph
Ph
Ph
Ph
Ph
Ph
E
E
E
E
E
E
E
E
E
E
Z
Z
Z
55
50
85
64
24
26
87
77
50
12
60
<5
19
99 (R)
96 (R)
98 (R)
98 (R)
90 (S)
80 (S)
99 (R)
98 (R)
97 (R)
92 (S)[e]
97 (S)
nd[f]
84 (R)
26 (R)
90 (R)
97 (R)
98 (R)
nd[g]
ꢀ
4-OMe C6H4
4-Me C6H5
4-Cl C6H4
2-Me C6H4
Ph
Me
iPr
89 (R)
ꢀ
4
4
4
22
>99 (R)
>99 (R)[h]
96 (R)[i]
98 (R)[j]
ꢀ
iPr
ꢀ
Me
SiMe3
Ph
[a] Average of at least two runs. [b] Reaction on a 0.2 mmol scale using
5 mol% of catalyst. [c] Determined by 1H NMR spectroscopy using an
internal standard and/or GC analysis. [d] Determined by chiral GC.
[e] See reference [17]. [f] See reference [18]. [g] Not determined.
[h] Using 10 mol% of catalyst. [i] At 508C. [j] Using 7.5 mol% of catalyst
at 358C.
Ph
61 (S)
[a] Average of at least two runs. [b] Reaction on a 0.1 mmol scale using
7.5 mol% of catalyst. [c] Determined by 1H NMR spectroscopy using an
internal standard and/or GC analysis. [d] Determined by chiral GC or
SFC. [e] Catalyst 3 f was used. [f] Not determined.
and subsequently evaluated in the asymmetric isomerization
of our test substrate, giving rise to contrasting results
(Table 1, entries 9–13). Catalysts 3g and 3h proved much
less reactive than their parent structure 3e, suggesting that
the introduction of substituents on either the ortho or meta
position may interfere with substrate coordination (Table 1,
entries 9 and 10). Catalyst 3i (R2 =p-anisyl) turned out to
be slightly less reactive than 3e but more enantioselective,
delivering the aldehyde with virtually perfect enantioselec-
tivity. Further optimization showed that the yield can be im-
proved if either the catalyst loading or the temperature is in-
creased, albeit at the expense of the enantioselectivity in the
latter case (Table 1, entries 11–13). Running the reaction in
a thermostated bath at 358C using 7.5 mol% of catalyst of-
fered a convenient compromise, and these conditions were
uniformly used for the isomerization of other primary allylic
alcohols.[27] Under these conditions, 3i reached performan-
ces similar to those displayed by the best catalysts of the
first and second generation, that is, 1a and 2a, respectively.
tries 5 and 6) were isomerized in modest yields but with a
significant improvement of the enantioselectivity value com-
pared to the first and second generations of catalysts. Impor-
tantly, no competing E/Z isomerization was detected when
1
monitoring the reaction by H NMR spectroscopy and GC.
These results validate the structural modification of the
phosphinoalkyloxazoline ligands (2!3) based on our mech-
anistic proposal and Charton analyses. They further suggest
that the increased steric demand in the vicinity of the reac-
tive sites funnels the reactions exclusively through the pro-
ductive isomerization pathway. The improved catalytic per-
formances of 3i over 1a are better visualized on the compa-
rative Charton analysis plotted on Figure 4. The reduced
value of the sensitivity factor Y (1.75 vs. 2.32) clearly indi-
cates that a change in the enantiomeric ratio as a factor of
steric demand is less pronounced with 3i than with 1a.
We also observed that a gradual increase of the electronic
density of the aromatic ring of the model substrate had a
beneficial impact both on the activity and the enantioselec-
tivity, whereas a decrease led to a slightly reduced yield and
ee value (compare entries 4 and 7–9 in Table 2). Substitution
at the ortho position was barely tolerated in terms of cata-
lytic activity, but the aldehyde was still obtained with excel-
lent enantioselectivity (Table 2, entry 10). This observation
may be paralleled with the detrimental effect on the catalyt-
ic activity of ortho and meta substitution of the aryl group of
the oxazoline in the ligand structure (Table 1, entries 9 and
10). The aromatic primary allylic alcohol with R1 =SiMe3,
which has a formal Z-configured double bond, was isomer-
ized with good yield and enantioselectivity (60% yield,
97% ee), whereas other Z-configured allylic alcohols, in
Asymmetric isomerization of E- and Z-configured aromatic
primary allylic alcohols: To explore the effect of the ligand
bridge homologation, a series of aromatic primary allylic al-
cohols with alkyl groups of various sizes at C3 were subject-
ed to the optimized conditions developed for the asymmet-
ric isomerization using 3i (Table 2). Substrates with large
alkyl substituents (R1 =tBu, iBu, Cy, iPr) delivered the cor-
responding aldehydes in moderate to good yields, with ex-
tremely high levels of enantioselectivity (96–99% ee;
Table 2, entries 1–4). Remarkably, aromatic primary allylic
alcohols with small alkyl groups (R1 =Me, Et; Table 2, en-
Chem. Eur. J. 2010, 16, 12736 – 12745
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12741

C. Mazet et al.
sponding to the isomerization of the E-configured 3,3-di-
AHCTUNGTREGaNNUN lkyl allylic alcohols is plotted in Figure 5. A linear, free-
energy relationship can be traced for the substrates, for
Figure 4. Comparative Charton analysis for aromatic primary allylic alco-
hols using 1a and 3i (corresponding to entries 1–6 in Table 2).
which the positions of the phenyl and alkyl groups are re-
versed, remain challenging targets (Table 2, entries 11–13).
Interestingly, in these reactions, again no competing E/Z iso-
merization was observed.
Figure 5. Comparative Charton analysis for 3,3-dialkyl primary allylic al-
cohols using 2a and 3i (corresponding to entries 1–4 in Table 3).
Asymmetric isomerization of 3,3-dialkyl primary allylic alco-
hols: Catalyst 3i was further evaluated in the asymmetric
isomerization of the more challenging 3,3-dialkyl primary al-
lylic alcohols using the optimized reaction conditions
(Table 3).
All substrates investigated were efficiently isomerized, de-
livering the chiral aldehydes in average to good yields with
promising enantioselectivity levels. In any case, parasite E/Z
isomerization was detected. Remarkably, geraniol was iso-
merized into citronellal with an unprecedented enantioselec-
tivity value, which, coincidentally, matches the value of nat-
ural citronellal (82% ee; Table 3, entry 4). Isomerization of
nerol turned out to be more challenging, and citronellal was
obtained in much lower yield and enantioselectivity (26%
yield, 31% ee; Table 3, entry 5). A Charton analysis corre-
which 3i clearly outperforms 2a (R1 =CH2CH2CH=C(Me)2,
Cy, tBu; Y 0.28, R2 =0.911) provided that the derivative
with a neopentyl substituent is excluded from the analysis.
According to Miller and Sigmanꢂs hypothesis, the apparent
kink in the correlation could be tentatively attributed to a
change in the conformation of the substrate and/or the
ligand. Alternatively the reaction may follow a different
mechanistic pathway. In the present case, we believe the in-
creased flexibility provided by the methylene homologation
in ligands of type 3 may account for the observed result.
Attempts to promote the asymmetric isomerization of tet-
rasubstituted or 2,3-disubstituted primary allylic alcohols
with catalysts 3 were met with
failure. Homoallylic alcohols
Table 3. Asymmetric isomerization of (E)-and (Z)-3,3-dialkyl primary allylic alcohols.[a,b]
and secondary allylic alcohols
were not isomerized either with
these iridium catalysts.
X-ray analysis and mechanistic
model: Crystals of suitable
quality for an X-ray crystallo-
graphic analysis of 3i were ob-
tained by indirect diffusion of
hexanes into a saturated solu-
tion of the iridium complex in
toluene.[28] Figure 6 contains a
comparison of the crystallo-
graphic structures of iridium
complexes (R)-1b[17] (R1 =1-
R1
R2
Olefin
config
Yield
[%][c]
ee
[%][d]
1
2
3
4
5
Me
Me
Me
Me
tBuCH2
tBu
Cy
E
E
E
E
Z
60
43
96
49
26
70 (R)
87 (S)
86 (S)
82 (R)
31 (S)
[e]
AHCTUNGTRENNUNG
[e]
ACHTUNGTRENNUNG(CH3)2C=CHCH2CH2
[a] Average of at least two runs. [b] Reaction on a 0.1 mmol scale using 7.5 mol% of catalyst. [c] Determined
by 1H NMR spectroscopy using an internal standard and/or GC analysis. [d] Determined by chiral GC. [e] A
Charton value of 0.70 was used.
12742
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12736 – 12745

Iridium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
Figure 6. Side and front views of the molecular structures of complexes (R)-1b and (S)-3i. The counteranion in both views and the CH2 cod in the front
views have been omitted for clarity (left). Sense of the oxidative addition of H2 to C1-symmetric [(P,N)Ir
the preferred binding of E-configured substrates to the catalysts (bottom right).
(cod)]+ (top right); quadrant diagrams to explain
ACHTUNGTRENNUNG
Ad, R2 =Ph) and (S)-3i, along with quadrant diagrams and
a rationale for the preferred two-point binding of the pri-
mary allylic alcohols to the catalysts. Both d8 complexes
adopt a distorted square planar arrangement around the
metal center, with a wider bite angle for the six-membered
chelate of 3i than for the five-membered chelate of 1b
(87.88(11)8 versus 80.43(17)8, respectively). In both struc-
tures the cyclooctadiene ligand is severely twisted out of the
coordination plane, counterclockwise for 1b and clockwise
for 3i. Consequently, one carbon of each double bond is
nearly in plane with the N, Ir, and P atoms. The strong trans
influence of the phosphorus atom of the C1-symmetric li-
bridge homologation impose a fundamentally different ori-
entation of the oxazoline ring. Whereas the oxazoline is in
plane with the chelate in the more rigid structure of 1b, the
increased flexibility in 3i places the heterocycle almost per-
pendicular to the coordination plane, thus projecting the
anisyl substituent above this plane. Although these struc-
tures give only insight into the thermodynamics of precata-
lysts for the asymmetric isomerization of primary allylic al-
cohols, they can be exploited to develop a rationale for the
preferred two-point binding of the substrate to the metal
center. The quadrant diagrams established on the basis of
the molecular structures of (R)-1b and (S)-3i clearly indi-
cate that only the south-eastern quadrant is spatially accessi-
ble. Furthermore, it is now well accepted that the oxidative
addition of molecular hydrogen to (P,N)–iridium complexes
is steered by the C1-symmetric nature of the ligand.[29]
Hence, after complete hydrogenation of the ancillary diole-
ꢀ
gands is well reflected by the noticeably longer Ir Ccod
bonds, in particular for the Ccod in plane with the other
donor atoms. (see Table 4). The different connectivity of the
oxazoline (Csp2 in 1b versus Csp3 in 3i) and the ligand
finic ligand, highly reactive iridiumACTHNUTRGNEUNG(III) intermediates in
Table 4. Comparison of selected structural features for (R)-1b and (S)-
3i. Selected NMR data for (S)-3i (500 and 126 MHz in CDCl3).
which the two hydrides are located cis to the phosphorus
are generated (Figure 6, top right). If chiral (P,N) ligands
are used, up to two isomeric complexes may be generated,
leaving in any case the two eastern quadrants trans to the
phosphorus electronically open.[17,30] The conjunction of the
electronic and steric discriminations suggest that binding
mode II will better accommodate the substituents of the E-
configured primary allylic alcohol on the open quadrant
(Figure 6, bottom right).
For aromatic primary allylic alcohols the phenyl ring will
occupy the free quadrant, whereas the alkyl group will inter-
act with the P substituent lying in the south-western quad-
rant. As the size of the alkyl increases this interaction may
become more pronounced.[31] For 3,3-dialkyl primary allylic
alcohols, the larger alkyl substituent may also be better ac-
Selected bond lengths [ꢃ]
Selected NMR data for (S)-3i
(500 MHz or 126 MHz; CDCl3)
d [ppm]
and angles [8]
(R)-1b
(S)-3i
ꢀ
N Ir
ꢀ
P Ir
2.110(6)
2.345(5)
2.138(7)
2.125(8)
2.202(8)
2.178(8)
1.40(1)
2.071(3)
2.373(1)
2.143(4)
2.165(4)
2.163(4)
2.205(4)
1.400(7)
1.399(7)
87.88(11)
Ha
5.19
4.06
4.60
3.15
70.5
63.0
Hb
Hc
Hd
Ca
Cb
Cc
Cd
ꢀ
Ir Ca
ꢀ
Ir Cb
ꢀ
Ir Cc
ꢀ
Ir Cd
90.6 (2JCP =16.0 Hz)
82.4 (2JCP =8.3 Hz)
ꢀ
Ca Cb
ꢀ
Cc Cd
1.40(1)
80.43(17)
N-Ir-P
Chem. Eur. J. 2010, 16, 12736 – 12745
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12743

C. Mazet et al.
commodated in the sterically open quadrant. This model is
in agreement with the stereochemical outcome of the reac-
tion, because both catalysts deliver preferentially the same
enantiomer of the aldehyde, although they have an opposite
absolute configuration. Interestingly, the south-eastern
quadrant for 3i appears only semi-hindered if compared to
that of 1b. This observation, the increased flexibility of the
ligand, and the wider bite angle in 3i are likely to be deter-
minant parameters coming into play to explain the im-
proved reactivity and selectivity observed for substrates with
smaller substituents in the asymmetric isomerization reac-
tion. Using this model it is understandable that repulsive
steric interactions may account for the absence of catalytic
activity in the isomerization of tetrasubstituted and 2,3-di-
terms of enantioselectivity, these new catalysts perform well
for sterically biased and unbiased allylic alcohols, thus not
only significantly expanding the substrate scope, but also
validating the initial hypotheses. In addition, the detrimental
competing E/Z isomerization of the substrate, a reaction
that was assigned to weak steric interactions between the
substrate and the catalyst, has been suppressed. In terms of
reactivity, a parameter that could not be predicted using the
Charton correlation, substrates having a small alkyl substitu-
ent are still obtained with unsatisfactory yields in some
cases. A model in which steric and electronic effects are re-
sponsible for the high enantio-inductions measured in the
asymmetric isomerization of primary allylic alcohols using
these iridium catalysts has been proposed. Current investiga-
tions are underway to gain additional mechanistic insights
into the isomerization reaction, and to develop more reac-
tive catalysts, which further expand the scope of this trans-
formation, in particular for Z-configured primary allylic
alcohols.
ACHTUNGTRENNUNGsubstituted primary allylic alcohols, as well as in the isomeri-
zation of secondary allylic alcohols. The lower enantioselec-
tivities observed for Z-configured allylic alcohols can be at-
tributed to the orientation of either the phenyl substituent
or the larger alkyl group in the south-western quadrant for
aromatic primary allylic alcohols and 3,3-dialkyl primary al-
lylic alcohols, respectively. Multinuclear mono- and bidimen-
sional NMR analyses indicated that the structure of 3i ob-
served in the solid state remained nearly identical in solu-
tion. Both the cyclooctadiene twist and the manifestation of
Acknowledgements
We thank the University of Geneva and the Swiss National Foundation
(project 200021-121546/1) for financial support. We thank Professors A.
Alexakis and E.-P. Kꢄndig (University of Geneva) for giving us access to
their analytical facilities, and Johnson Matthey for providing us with a
generous amount of the iridium precursors. We also thank Professor E.-P.
Kꢄndig and Dr. Charles Fehr (Firmenich SA) for stimulating discussions.
1
the trans influence are particularly visible in the H and 13C
chemical shifts and the JCP coupling constants (Table 4). The
signals of the two carbon atoms in the coordination plane
(Ca, Cc) and the signal of the two corresponding protons
(Ha, Hc) are significantly deshielded with respect to the
nuclei out of plane (Cb, Cd, Hb, Hd), thus reflecting the
better electronic communication with the heteroatom locat-
ed in the trans position. The average chemical shift differen-
[1] a) S. Akutagawa, in Comprehensive Asymmetric Catalysis, Vol. 2
(Eds: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer, Berlin,
1999, Chapter 23; b) S. Akutagawa, in Comprehensive Asymmetric
Catalysis, Vol. 3 (Eds: E. N. Jacobsen, A. Pfaltz, H. Yamamoto),
Springer, Berlin, 1999, Chapter 41.4.
[2] Seminal contributions: a) K. Tani, T. Yamagata, S. Otsuka, S. Akuta-
gawa, H. Kumobayashi, T. Taketomi, H. Takaya, A. Miyashita, R.
Noyori, J. Chem. Soc. Chem. Commun. 1982, 600–601; b) K. Tani,
Pure Appl. Chem. 1985, 57, 1845–1854; for mechanistic studies, see:
c) S. Inoue, H. Takaya, K. Tani, S. Otsuka, T. Sato, R. Noyori, J.
Am. Chem. Soc. 1990, 112, 4897–4905.
[3] Recent reviews: a) R. C. Van der Drift, E. Bouwman, E. Drent, J.
Organomet. Chem. 2002, 650, 1–24; b) R. Uma, C. Crꢀvisy, R. Grꢀe,
Chem. Rev. 2003, 103, 27–51; c) G. C. Fu, In Modern Rhodium-Cat-
alyzed Organic Reactions, Ed: Evans, P. A.; Wiley-VCH: Weinheim,
2005, Chapter 4; d) V. Cadierno, P. Crochet, J. Gimeno, Synlett 2008,
1105–1123.
[4] a) K. Tanaka, S. Qiao, M. Tobisu, M. M.-C. Lo, G. C. Fu, J. Am.
Chem. Soc. 2000, 122, 9870–9871; b) K. Tanaka, G. C. Fu, J. Org.
Chem. 2001, 66, 8177–8186.
1
ces are 1.13 and 1.29 ppm in the H NMR spectrum, and 7.5
and 8.2 ppm in the 13C{1H} NMR spectrum. The better elec-
2
tronic communication is also emphasized by the greater JCP
2
2
value for Cc (Cc: JCP =16.0 Hz; Cd: JCP =8.2 Hz). The ortho
and meta protons of the anisyl substituent are equivalent, in-
2
2
ꢀ
dicating rotation around the Csp Csp aryl–oxazoline bond
on the timescale of the NMR experiment. The intense NOE
contacts between Hd and the anisyl ortho protons, and be-
tween Hc and the proton at the stereogenic center, indicate
the maximized stereochemical communication between the
chiral oxazoline unit and the coordination site where the
double bond of the primary allylic alcohol is expected to
bind.
[5] For other relevant examples, see: a) C. Botteghi, G. Giacomelli,
Gazz. Chim. Ital. 1976, 106, 1131–1134; b) C. Chapuis, M. Barthe,
J.-Y. de Saint Laumer, Helv. Chim. Acta 2001, 84, 230–242; c) F.
Boeda, P. Mosset, C. Crꢀvisy, Tetrahedron Lett. 2006, 47, 5021–5024;
d) A. Doppiu, A. Salzer, Eur. J. Inorg. Chem. 2004, 2244–2252; e) S.
Bovo, A. Scrivanti, M. Bertoldini, V. Beghetto, U. Matteoli, Synthe-
sis 2008, 2547–2550; for Ru-catalyzed asymmetric isomerization of
secondary allylic alcohols, see: f) M. Ito, S. Kitahara, T. Ikariya, J.
Am. Chem. Soc. 2005, 127, 6172–6173.
[6] a) R. R. Schrock, J. A. Osborn, J. Am. Chem. Soc. 1976, 98, 2134–
2143; b) C. S. Chin, J. H. Shin, C. Kim, J. Organomet. Chem. 1988,
356, 381–388; c) Y. Sun, R. N. Landau, J. Wang, C. LeBlond, D. G.
Blackmond, J. Am. Chem. Soc. 1996, 118, 1348–1353.
Conclusion
A new generation of well-defined chiral (P,N)-iridium cata-
lysts for the asymmetric isomerization of primary allylic al-
cohols to aldehydes has been elaborated, based on the
linear free energy relationship between steric parameters
and enantiomeric ratios. Using a preceding ligand genera-
tion, the steric demand has been brought closer to the coor-
dination sites where the reaction takes place by a formal ho-
mologation of the bridge linking the P and N donors. In
12744
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12736 – 12745

Iridium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
[7] D. Baudry, M. Ephritikhine, H. Felkin, Nouv. J. Chim. 1978, 2, 355–
356.
steric effects. See reference [22] for a detailed discussion on this
matter.
[8] C. Fehr, I. Farris, Angew. Chem. 2006, 118, 7058–7061; Angew.
Chem. Int. Ed. 2006, 45, 6904–6907.
[24] a) D.-R. Hou, K. Burgess, Org. Lett. 1999, 1, 1745–1747; b) D.-R.
Hou, J. H. Reibenspies, K. Burgess, J. Org. Chem. 2001, 66, 206–
215.
[25] D.-R. Hou, J. H. Reibenspies, T. Colacot, K. Burgess, Chem. Eur. J.
2001, 7, 5391–5400.
[9] Guillou, Bonjoch, and Evans have independently observed that cat-
alyst C promoted isomerization reactions although with lower de-
grees of selectivity and predictability. See: a) M. Krel, J.-Y. Lalle-
mand, C. Guillou, Synlett 2005, 2043–2046; b) D. Solꢀ, X. Urbaneja,
J. Bonjoch, Org. Lett. 2005, 7, 5461–5464; c) Y. Kavanagh, C. M.
Chaney, J. Muldoon, P. Evans, J. Org. Chem. 2008, 73, 8601–8604.
[10] R. Crabtree, Acc. Chem. Res. 1979, 12, 331–337.
[11] G. Stork, D. Kahne, J. Am. Chem. Soc. 1983, 105, 1072–1073.
[12] a) R. H. Crabtree, M. W. Davis, Organometallics 1983, 2, 681–682;
b) R. H. Crabtree, M. W. Davis, J. Org. Chem. 1986, 51, 2655–2661.
[13] a) D. A. Evans, L. Kværnø, T. B. Dunn, A. Beauchemin, B. Raymer,
J. A. Mulder, E. J. Olhava, M. Juhl, K. Kagechika, D. A. Favor, J.
Am. Chem. Soc. 2008, 130, 16295–16309; b) S. Gao, Q. Wang, C.
Chen, J. Am. Chem. Soc. 2009, 131, 1410–1412.
[14] L. Mantilli, C. Mazet, Tetrahedron Lett. 2009, 50, 4141–4144.
[15] B. Wꢄstenberg, A. Pfaltz, Adv. Synth. Catal. 2008, 350, 174–178.
[16] a) M. Brookhart, B. Grant, A. F. Volpe, Organometallics 1992, 11,
3920–3922; b) S. P. Smidt, N. Zimmermann, M. Studer, A. Pfaltz,
Chem. Eur. J. 2004, 10, 4685–4693.
[17] L. Mantilli, D. Gꢀrard, S. Torche, C. Besnard, C. Mazet, Angew.
Chem. 2009, 121, 5245–5249; Angew. Chem. Int. Ed. 2009, 48, 5143–
5147.
[26] a) M. C. Perry, X. Cui, M. T. Powell, D.-R. Hou, J. H. Reibenspies,
K. Burgess, J. Am. Chem. Soc. 2003, 125, 113–123; b) L-aspartic
acid is half as expensive as l-serine, thus validating the increased ac-
cessibility of such ligands as discussed in references [17] and [24].
[27] If the conditions used for catalysts 1a and 2a are employed
(5 mol%, 238C), the ee values reported in Tables 2 and 3 are virtual-
ly identical, but the yields are much lower. Therefore, it is still rea-
sonable to compare the three generations of catalysts as represented
in Figures 4 and 5.
[28] Crystals of suitable quality for X-ray analyses were grown using the
PF6 analogues of complexes (R)-1b and (S)-3i. CCDC-774476 con-
tains the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[29] a) T. G. Appleton, H. C. Clark, L. E. Manzer, Coord. Chem. Rev.
1973, 10, 335–422; b) R. H. Crabtree, H. Felkin, T. Fillbeen-Kahn,
G. E. Morris, J. Organomet. Chem. 1979, 168, 183–195; c) B. F. M.
Kimmich, E. Soomsook, C. R. Landis, J. Am. Chem. Soc. 1998, 120,
10115–10125; d) R. H. Crabtree, R. Uriarte, Inorg. Chem. 1983, 22,
4152–4154; e) C. Mazet, S. P. Smidt, M. Meuwly, A. Pfaltz, J. Am.
Chem. Soc. 2004, 126, 14176–14181.
[18] L. Mantilli, C. Mazet, Chem. Commun. 2010, 46, 445–447.
[19] J. Sprinz, G. Helmchen, Tetrahedron Lett. 1993, 34, 1769–1772.
[20] a) A. M. Porte, J. Reibenspies, K. Burgess, J. Am. Chem. Soc. 1998,
120, 9180–9187; b) K. Burgess, A. M. Porte, Tetrahedron: Asymme-
try 1998, 9, 2465–2469.
[30] In agreement with our previous labeling experiments (see refs. [14]
and [17]) and on the basis of the mechanism we propose, we believe
these hydrides are exchanging rapidly in such species and hence the
two isomers are interconverting at a faster rate than the rate-limit-
ing migratory insertion.
[21] A Charton analysis for ligands of type 1 is available in the Support-
ing Information.
[22] a) J. J. Miller, M. S. Sigman, Angew. Chem. 2008, 120, 783–786;
Angew. Chem. Int. Ed. 2008, 47, 771–774; b) M. S. Sigman, J. J.
Miller, J. Org. Chem. 2009, 74, 7633–7643.
[23] In this study, we focused on the use of Charton steric parameters be-
cause they do take into account the van der Waals radii of the alkyl
groups, and allow the exclusion of inductive and resonance effects.
In contrast, the Taft steric parameters have been the subject of
some controversy because they were not an exclusive measure of
[31] A phenyl substituent has two distinct Charton values (0.57 and
1.66). Consistent with our rationale, the phenyl group has a larger
value than an Me (0.52) or an Et (0.56) substituent in any case. Nev-
ertheless, the largest value is preferred for phenyl groups lying out
of plane, as is presumably the case for the isomerization reaction
studied herein.
Received: May 14, 2010
Published online: September 15, 2010
Chem. Eur. J. 2010, 16, 12736 – 12745
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12745