COMMUNICATION
Expanded scope for the iridium-catalyzed asymmetric isomerization of
primary allylic alcohols using readily accessible second-generation
catalystsw
´
Luca Mantilli and Clement Mazet*
Received (in College Park, MD, USA) 29th September 2009, Accepted 19th November 2009
First published as an Advance Article on the web 3rd December 2009
DOI: 10.1039/b920342g
A second generation of chiral (P,N)-iridium catalysts—readily
accessible from inexpensive L-serine—displays expanded scope
for the asymmetric isomerization of primary allylic alcohols.
with commercially available protected dialkylphosphines
(see ESIw for details). Nonetheless, when comparing the limited
number of natural amino acids with the virtually infinite
catalogue of carboxylic acids, ligand scaffold L-2 offers a
potentially higher structural diversity at R2. In order to access
a small but structurally relevant library of six iridium pre-
catalysts, we limited ourselves to two sets of complexes having
a trialkylphosphine moiety on the left-hand-side (R1 = Cy,
t-Bu, 1-Ad) and either an sp2 or sp3 hybridized carbon at R2
(R2 = Ph for 2a–c and t-Bu for 2d–f, respectively).
In modern asymmetric catalysis, at the first stage of their
development, chiral catalysts are primarily evaluated based
on their intrinsic ability to deliver the targeted products in
high yields and enantioselectivities. Optimization of the
reaction conditions (catalyst loading, solvent choice, optimal
temperature etc.) and investigation of the substrate scope
are considered next.1 Catalyst accessibility—which ultimately
relies on an efficient, short and modular synthetic route using
inexpensive starting material from the chiral pool—is often
underappreciated. In homogeneous transition-metal-based
asymmetric catalysis, this has sometimes led to paradoxal
situations where the organic chiral ligands are more expensive
by a few order of magnitude than the precious metal salts.
We have recently disclosed that chiral (dialkyl)phosphanyl-
alkyloxazoline-iridium catalysts promote the asymmetric
isomerization of primary allylic alcohols to the corresponding
aldehydes under mild reaction conditions.2–4 The most selective
and general candidate 1a was unfortunately derived from
expensive, non-natural, L-tert-leucine. Although they exclusively
varied the oxazoline substituent R2 while keeping the diphenyl-
phosphine moiety unchanged, Burgess and co-workers have
convincingly demonstrated the highly modular nature of
ligand L-2 in the context of Pd-catalyzed asymmetric allylic
substitution (Fig. 1).5
Ligand deprotection, iridium complexation and subsequent
chloride abstraction with NaBArF were performed in a one-pot
process combining standard literature procedures affording
2a–f in good yields (Scheme 1).7 Each of the air-stable iridium
complexes 2a–c (R2 = Ph) are systematically obtained as a
mixture of two non-interconverting isomers as evidenced by
1H and 31P{1H} variable-temperature NMR measurements
performed between À50 1C and +60 1C. Interestingly, the
ratio decreases as the size of the phosphorus substituents
increases. We assume these isomers are conformers. Their
formation is tentatively attributed to interactions between
the phosphorus substituents R1 and/or the phenyl group at
R2 with the cyclooctadiene moiety. Complexes 2d–f (R2 = t-Bu)
are virtually obtained as a single isomer.
In order to evaluate the potential of catalysts 2a–f in
the asymmetric isomerization of primary allylic alcohols, a
comparative study using model substrate (E)-4-methyl-3-
phenylpent-2-enol 3a was carried out (Table 1). The reactions
were performed in THF at room temperature using 5 mol%
catalyst loading. The solution was degassed after activation by
molecular hydrogen to prevent competing hydrogenation.
Whereas catalysts 2d–f performed poorly, delivering the
Inspired by this work, we report herein a second generation
of iridium catalysts for the asymmetric isomerization of primary
allylic alcohols. This second generation of catalysts is readily
available from dialkylphosphine precursors and inexpensive
L- or D-serine.6 They display similar reactivity and selectivity
to catalyst 1a for most substrates and improved performance for
the more challenging 3,3-dialkyl primary allylic alcohols.
The synthetic routes leading to the protected ligands L-1
and L-2 have been described in the literature and compare
relatively well in terms of efficiency (number of steps, overall
yield). Both are equally modular since in each case a final SN2
reaction allows one to combine the oxazoline building block
University of Geneva, Organic Chemistry Department,
Quai Ernest Ansermet 30, 1211-Geneva, Switzerland.
E-mail: clement.mazet@unige.ch; Fax: +41 (0) 22 3793215;
Tel: +41 (0) 22 3796288
w Electronic supplementary information (ESI) available: Full experi-
mental data and characterization for all new compounds. See DOI:
10.1039/b920342g
Fig. 1 Comparative synthetic routes leading to ligands L-1 and
L-2 (Cy = cyclohexyl; t-Bu = tert-butyl; 1-Ad = 1-adamantyl,
BArF
=
tetrakis-[3,5-bis-(trifluoromethyl)phenyl]borate; COD
=
1,5-cyclooctadiene).
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 445–447 | 445