close

Вход

Забыли?

вход по аккаунту

?

Chiral Recognition in Complexes of Alkenes Aldehydes and Ketones with Transition Metal Lewis Acids; Development of a General Model for Enantioface Binding Selectivities.

код для вставкиСкачать
[(q5-C,Hs)Re(NO)(PPh3)]+ and analogons unsaturated chiral complexes (see element symbols along
the fingertips) can form two diastereomeric adducts
with prochiral alkenes, aldehydes, and ketones. The
ratio of these diastereomers is a measure of the chiral recognition, symbolized by the hands, of the
reaction partners. The rhenium complex mentioned is
depicted above the hands (Re: light gray, P: yellow,
N: light blue, 0: dark blue, Ph, Cp: dark gray). The
selectivity of alkene complexation can be analyzed
(and predicted) by considering the sizes of the substituents a 4 on the alkene and correlating them with the
steric conditions in the chiral complexes, which can
be represented with three-dimensional bar graphs.
REVINS
Chiral Recognition in n: Complexes of Alkenes, Aldehydes, and Ketones with
Transition Metal Lewis Acids;
Development of a General Model for Enantioface Binding Selectivities
J. A. Gladysz” and Brian J. Boone
Prochiral alkenes, aldehydes, and ketones constitute the most frequently
used starting materials for enantioselective organic syntheses. Protocols often
involve chiral binding agents or Lewis
acids that can give two diastereomeric
adducts, the ratios of which are measures of chiral recognition. With 7c adducts, the diastereomers differ in the
enantioface of the C = C or O=C group
bound to the Lewis acid. This review
provides the first comprehensive analy-
sis of such equilibria and related binding
phenomena with chiral transition metal
Lewis acids. An extensive body of data
from the authors’ laboratory for complexes of the pyramidal rhenium fragment [(y5--C5H,)Re(N0)(PPh3)]+
(I) affords particular insight. Literature data
for other complexes are also summarized. A general model for chiral recognition based upon the relative steric
properties of four quadrants is presented. This enables binding selectivities to
1. Introduction
Methodologies for enantioselective syntheses of chiral organic compounds from achiral precursors are developing hypertrophically.[’] Among the many classes of prochiral starting materials, alkenes, aldehydes, and ketones are the most common.
As illustrated in Scheme 1, these undergo a variety of 1,2-addition reactions that give chiral, but normally racemic, products.
Enantioselective versions of many such additions have been
crafted and continue to be actively sought. Despite the considerable mechanistic diversity of these reactions, some common attributes are apparent. For example, many protocols feature a
chiral binding agent or Lewis acidL2]that can give two
diastereomeric adducts. These adducts in turn evolve to products via two (or more) diastereomeric series of transition states.
Often, the Lewis acid activates the substrate towards attack of
an external agent, which fixes a product stereocenter. The dominant configuration or degree of “enantioselection” is determined by rate differences. Either the familiar Curtin-Hammett
or Winstein-Holness limits may apply.r31In the latter case,
there can be a direct link between binding selectivity and
product selectivity.
[*I
Prof. Dr. J. A. Gladysz, B. J. Boone
Department of Chemistry, University of Utah
Salt Lake City, UT 84112 (USA)
Fax: Int. code +(801)585-7807
e-mail: giadysz@rhenium.chem.utah.edu
Angew. Chem.Int. Ed. Engl. 1997, 36, 550-583
be individually and rationally optimized
for different classes of ligands. Electronic effects are also identified and correlated with specific structural properties.
Relationships between binding equilibria, reactivity, and product configurations are discussed.
-
Keywords: alkene complexes carbonyl
complexes * chiral recognition stereoselective syntheses * transition metal
Lewis acids
-
Regardless of the exact
origin of enantioselection,
“chiral recognition” is in
some context required. For
the purposes of this review,
chiral recognition will be defined as the selective formation of one of two possible
diastereomeric adducts from
B-X
X-B
I c-enantiomers
I
two chiral or prochiral spec i e ~ . [Of
~ ] course, there is an
important distinction between kinetic and thermodyI = HP. HSiR3, “0,
1,3dienes, etc.
namic selectivities. The latter
Scheme 1. Enantioselective syntheses
will be implied unless stated
from achiral alkenes, aldehydes, and
otherwise. Also, competing
ketones.
transition states are often
easily analyzed from the standpoint of chiral recognition.
This review provides a detailed and comprehensive analysis of
chiral recognition in yz 7[: complexes of prochiral alkenes, aldehydes, and ketones with chiral transition metal Lewis acids.
Surprisingly, this subject has only received fragmentary treatment in the literature previously. It should be emphasized in
passing that such phenomena can also play key roles in reactions that do not give chiral or nonracemic products. For example, in the polymerization of propene by chiral metallocene
catalysts, the tacticity of the polypropylene depends upon the
0 VCH Verlugsgesellschuft mbH, D-6945/
-
a
Weinheim,f997
0570-0833/9713606-055/$ /7.50+ S0/0
551
REVIEWS
enantioface of the propene C=C bond that inserts into the
metal-carbon bond of the growing polymer.[’]
This review begins with a discussion of the general properties
of transition metal Lewis acids and alkene, aldehyde, and ketone TC complexes. Adducts of the chiral rhenium Lewis acid
[ (4’-C,H,)Re(NO)(PPh,)]+ (I) are then analyzed in detail. This
extensive body of data has been acquired in the authors’ laboratory over the last decade and affords numerous new insights.
Next, a general model for chiral recognition in TC complexes is
presented. This is based upon the relative steric properties of
four quadrants and enables binding selectivities to be individually and rationally optimized for different classcs of ligands. In
the fourth part of this review, literature data for .n complexes of
other chirdl transition metal Lewis acids are summarized. Finally, relationships between binding equilibria, reactivity, and
product configurations are discussed.
2. Transition Metal Lewis Acids
Transition metal Lewis acids are ubiquitous and can be subdivided into various classes. Sixteen-valence-electron complexes
are probably the most common. Some are isolable, whereas
others are easily generated by the dissociation of weakly bound
ligands from coordinatively saturated, eighteen-valence-electron precursors. There are obvious parallels to six-valence-electron species such as trivalent boron Lewis acids and carbocations. However, unlike boron Lewis acids, transition metal
Lewis acids with other valence electron counts can also be generated or isolated.
A transition metal Lewis acid need not exist as a discrete
entity. For example, a metal fragment might move from Lewis
base to Lewis base strictly through associative mechanisms.
Nonetheless, organic substrates may be activated, and relative
binding strengths can be quantified, as with Lewis acids that can
exist in unassociated forms.
J. A. Gladysz and B. J. Boone
Chiral transition metal Lewis acids can be divided into two
classes: 1) those in which the Lewis base binding sites are
stereocenters.[61 often described as “chiral-at-metal”, and
2) those that have ligand-based stereocenters such as in chiral
phosphane ligands. The former have no counterparts among
common main-group-element Lewis acids. For example, BF,
and AICI, have trigonal-planar binding sites and can be rendered chiral only with chiral substituents.
One distinguishing feature of transition metal Lewis acids
deserves emphasis. Namely, except in the case of do species such
as TiTVor Ta”, occupied d orbitals are present. These are normally the highest energy orbitals and capable of backbonding to
acceptor orbitals on Lewis bases. Thus, most transition metal
Lewis acids are more accurately viewed as amphoteres. As a
consequence, the relative binding affinities of various Lewis
bases can differ markedly from those of archetype Lewis acids
such as H’, Li’, and BF, .[71 In these cases, donor orbitals are
either absent or are very low in energy.
It should also be noted that backbonding can be an important
determinant of ligand conformation. Such interactions are almost always of TC symmetry, resulting in a purely electronic
component to the barrier for rotation about the metal-ligand
axis. A familiar illustration is the Dewar- Chatt-Duncanson
model for bonding in alkene complexes. Further, in an asymmetric environment, all d orbital degeneracies are removed.
Thus, for the types of TC complexes in this review, it will generally
be assumed that only one electronic conformational energy minimum need be considered (derived from the d orbital HOMO).
Backbonding has other types of structural consequcnces, several of which are examined in this article.
Finally, the unique electronic properties of transition metal
Lewis acids also allow the binding and/or activation of many
small molecules such as H, and CO. Thus, transformations that
would not be possible with main-group-element Lewis acids can
be effected (hydrogenation, hydroformylation, etc.) .
John A. Gladysz is a native of the Kalamazoo area in Michigan
( U S A ) . He obtained a B. S. degree from the University of Michigan (1971) and a Ph.D. degree from Stanford University (1974)
for research conducted with E . E. van Tamelen. He subsequently
moved to the University of California, Los Angeles as an Assistant
Professor. In 1982 he accepted an appointment at the University of
Utah, where he is now Projessor of Chemistry. He is a jellow of the
Alfved P. Sloan Foundation (1980- 1984) anda Camille and Henry
Dreyfus Teacher-Scholar (1980-1985). He received an Arthur C.
Cope Scholar Award in 1988, the University of Utah Distinguished
Research Award in 1992, the ACS Award in Organometallic ChemJ. A. Gladysz
B. J. Boone
istry in 1994, and a Research Awardjor Senior Scientists from the
Alexander von Humboldt Foundation in 199s. Since 1984 he has
been Associate Editor of Chemical Reviews. Ire has authoredover 22Spublications and his research spans a wide range of topics
in the general area of synthetic and mechanistic organometallic chemistry.
Brian J. Boone was born in 1968 in Botwood, Newfoundland (Canada). He obtained a B. Sc. (Honors) degree jrom the
Memorial University of Newfoundland (1990) and twice received the National Science and Engineering Research Council
Award for Undergraduate Studies. He is currently a graduate student at the University of Utah, where he was awarded a Henry
Eyring Research Fellowship.
552
Angew. Chem. Int. Ed. Engl. 1997, 36, 550 - 5 8 3
Chiral Recognition in n Complexes
REVIRNS
3. Classes of Alkene, Aldehvde.
and Ketone 7c Complexes
"
A. Ethylene
I
3.1. Alkene Complexes
Consider a chiral transition metal Lewis acid of fixed absolute
configuration, and simple achiral alkenes with all possible substitution patterns; Scheme 2 depicts the principal binding modes. As analyzed in this section, these include both configurational and conformational diastereomers. To help convey
structural and conceptual similarities to three-membered-ring
organic compounds, metallacyclopropane resonance forms are
utilized. Many instructive relationships are apparent, and thereC. Symmetrically substituted trans Alkenes
fore each class of alkene ligand is individually discussed.
A) Ethylene: ethylene is not prochiral."] The n: faces of the
C=C bond are homotopic or identical, and the C=C termini
bear identical substituents. Consequently, adducts with chiral
transition metal Lewis acids can only exist as a single isomer.
(identical)
Nonetheless, the four vinylic protons and two carbon atoms of
the double bond become inequivalent. Thus, four =CH D. Symmetrically substituted cis
'H NMR signals and two =CH, I3C NMR signals will be observed in the slow-exchange limit. This renders ethylene a useful
probe ligand for the determination of dynamic properties, as
will be described later.
homotopic
B) Monosubstituted alkenes : such alkenes are prochiral. The
B :
L------------.-----------------------------~
C=C faces are enantiotopic,['] and the C=C termini bear different substituents. Thus, four isomeric complexes are possible. E. Symmetrically substituted Geminal Alkenes
First, two Configurational diastereomers can exist, which differ in
the C=C enantioface bound to the metal and can interconvert
by a "face flip".['] The =CHR carbons are stereocenters, and
I
their configurations can be designated by the RIS conventions
analogous to those used for three-membered-ring organic comIldentlcal)
~------_________________________________-----------------------pounds." O,
Second, for each configurational diastereomer,
two conformutional diastereomers or M -(C=C) rotamers F. Unsymmetrically substituted Geminal Alkenes
can exist. These differ by a 180" rotation in the C=C n nodal
plane and can be distinguished with terms such as scjac or
sp/ap.[l2]
C) Symmetrically substituted trans alkenes: such alkenes are
enantiotopic
prochiral. The C = C faces are enantiotopic and thus, as with
monosubstituted alkenes, two configurational diastereomers G. Unsymmetrically substituted cans Alkenes
are possible. However, the C=C termini bear identical substituents, which are exchanged by a 180" rotation about the
M-(C=C) axis. Hence, each diastereomer has only one
M-(C=C) rotamer. Importantly, in these complexes the trans
alkene ligands have two =CHR stereocenters. The configurations are fixed relative to each other, as a single inversion would H. Unsymmetrically substituted cis Alkenes
give a cis alkene ligand. A reciprocal situation exists with cis
alkene ligands.
D) Symmetrically substituted cis alkenes : such alkenes are
not prochiral. The C=C faces are homotopic or identical, and
the C=C termini bear identical substituents. Nonetheless, two
isomeric complexes are possible. These can be viewed as config- 1. Symmetrically trisubstituted Alkenes
urational diastereomers that can interconvert by a C=C face
flip, or conformational diastereomers that can interconvert by a
rotation
R M R
180" rotation about the M-(C=C) axis. Regardless, such isomers can lead to enantiomeric organic products. For example,
enantiotopic
consider the 1,a-addition of an achiral species A-B from the
side Opposite to the
and in a regiospecific fashion (that is
Scheme 2. Classification of complexes formed from chiral transition metal Lewis
A always adds to C = c and B to C=C). One isomer will give one
acids and achiral alkenes [XI.
,
.
-
Angew. Chem. Int. Ed. Engl. 1997. 36, 550-583
-
553
REVINS
product enantiomer, and the other the opposite enantiomer.
Nonracemic mixtures would be expected.[l3I
E) Symmetrically substituted geminal alkenes: such alkenes
are not prochiral. Although the C=C faces are homotopic or
identical, the C=C termini bear different substituents. Hence,
two M-(C=C) rotamers are possible. However, in this case (as
well as for ethylene), 1,2-addition of an achiral species cannot
give a chiral organic product.
F) Unsymmetrically substituted geminal alkenes: such alkenes are prochiral. The C=C faces are enantiotopic, and the
C=C termini bear different substituents. Thus, both configurational diastereomers and M-(C=C) rotamers are possible, in
analogy to monosubstituted alkenes.
G) Unsymmetrically substituted trans alkenes: such alkenes
are prochiral and give isomers that are analogous to those analyzed in F ) but having two =CHR stereocenters.
H) Unsymmetrically substituted cis alkenes: such alkenes are
prochiral and give isomers analogous to those analyzed in G).
I) Trisubstituted alkenes : all trisubstituted alkenes are
prochiral, regardless of substitution pattern. When the three
substituents are identical, the isomers are analogous to those
analyzed in B). Otherwise (except when the geminal substitucnts
are identical), there will be two = C stereocenters, as in G)
and H).
J) Tetrasubstituted alkenes: the preceding examples can be
extended to tetrasubstituted alkenes in which all four substituents are identical (see A)), three substituents are identical
(see B)), both pairs of trans, cis, or geminal substituents are
identical (see C), D), and E)) , and one pair of geminal, trans, or
cis substituents is identical (see F), G), and H)). Tetrasubstituted alkenes with four different substituents would be analogous
to trisubstituted alkenes with three different substituents.
3.2. Aldehyde and Ketone Complexes
The preceding analysis is easily applied to 7c complexes of
achiral aldehydes and ketones. Formaldehyde and symmetrically substituted ketones will give isomeric complexes analogous to
those of symmetrically substituted geminal alkenes (E) . Other
aldehydes will give isomeric complexes analogous to those of
monosubstituted alkenes (B), and other ketones will give isomeric complexes analogous to those of unsymmetrically substituted geminal alkenes (F).
3.3. Related Classes of Complexes
Two extensions of Scheme 2 deserve emphasis:
1) Consider adducts of achiral transition metal Lewis acids
and prochiral alkenes (B), C), F)-I)), aldehydes, and ketones.
Carbon stereocenters will be generated, resulting in chiral complexes. The enantiomers will normally form in equal amounts.
If the achiral transition metal Lewis acid is also prochiral, then
metal stereocenters will be concurrently generated. The metal/
carbon configurational diastereomers will normally form in unequal amounts.
2) Most chiral alkenes have diastereotopic[’l C=C faces (exceptions would include chiral alkenes with C, symmetry). Con554
J. A. Gladysz and B. J. Boone
sider as an example a monosubstituted alkene with one allylic
stereocenter. In any type of 7c complex, a second carbon stereocenter will be generated. Thus, achiral (nonprochiral) transition
metal Lewis acids will give either two or four configurational
isomers, depending upon whether the alkene is enantiomerically
pure or racemic (two diastereomers, or two diastereomers and
their enantiomers). Similarly, chiral transition metal Lcwis acids
will give up to eight configurational isomers. In both cases, the
diastereomers will normally form in unequal amounts.
Chiral alkenes will be treated only in passing in this review.
However, the development of chiral transition metal Lewis acids
that will selectively bind or promote the reaction of one enantiomer of a racemic chiral alkene is of obvious importance. The
bonding models developed here provide a basis for rational
approaches.
4. Alkene Complexes;
General Bonding Considerations
Many books and reviews have been devoted to transition
metal alkene c o m p l e ~ e s . [ ~Their
~ l structural and electronic
properties have been analyzed in detail. These include several
subtle features, and aspects of this extensive earlier literature
will be introduced as needed.
Chiral receptors for alkenes have sometimes been analyzed in
terms of the steric properties of “quadrants” corresponding to
each vinylic position.[’ 5 , l 6 I A quadrant lettering system (A-D)
that will be used throughout this review is given at the top of
Scheme 3. Such diagrams and extensions introduced in Sections
22-24 allow control elements in chiral recognition to be easily
conceptualized.
Considerable effort has gone into the quantification of steric
properties of organometallic fragments and ligands.[l7I Steric
properties of organic groups are often probed by measurements
of various types of diastereomeric equilibria. One familiar example would be the “A values” associated with axial substituents in cyclohexanes.[”I However, with alkene complexes
of chiral metal fragments, it was apparently not recognized that
measurements of binding equilibria can provide a semiquantitative map of the relative steric environments of the quadrants.
For example, consider the configurational diastereomers of
the complex of a symmetrically substituted trans alkene (C in
Scheme 3). The equilibrium constant will reflect the relative
steric environments of quadrants (A C) and (B D). Similarly, the equilibrium constant for the M-(C=C) rotamers of
the complex of an unsymmetrically substituted trans alkene (G
in Scheme 3) will reflect the relative steric environments of
quadrants A and C (for one configurational diastereomer), and
of B and D (for the other configurational diastereomer).
In principle, such equilibria might be established by keeping
the configurations of the alkene carbons constant and inverting
the chiral transition metal Lewis acid. This pathway would give
the enantiomers of the products depicted in many of the schemes
in this review. However, conclusions regarding the relative steric
environments of the quadrants would not be affected. Such
mechanisms are highly improbable for most Lewis acids with
ligand-based chirality. For example, any carbon stereocenters in
chiral phosphane ligands would have to undergo inversion. Re-
+
+
Angew. Chem. Int. Ed. Engl. 1997.36, 550-583
Chiral Recognition in a Complexes
REVIEWS
A. Ethylene
8. MonosubmutedAlkenes
D
A
C vsA
A vsB
B vsD
?
L_____-___-___-___________
C. Symrnetncally substiMed trans Alkenes
0. Symmetricallysubstituted cis Alkenes
(A
+ 0 ) vs (6 + C)
E. Symmetrically substituted Geminal Alkenes
Fl Fi
(C + 0) vs (A + B)
F. Unsymmetrically substituted Geminal Alkenes
M-(C=Q
,-,
A
C=C-
0
TGxiT
R/
rotation
A
k3
G. Unsymmetrically substituted trans Alkenes
H. Unsymmetncallysubstituted CISAlkenes
rotahon
A
w
A
Scheme 3. Relationships between alkene binding equilibria and the relative steric environments of the quadrants of the Lewis acid
receptors.
Angen,. Chem. Inr. Ed. EngI. 1997, 36, 550-583
gardless, if nonracemic adducts are available, the
epimerizing stereocenter is
usually easy to identify.
Accordingly, the chiral-atmetal complexes studied in
the authors’ laboratory
and described in subsequent sections of this review have been shown to
equilibrate with retention
of configuration at the
metal and inversion at carbon.
Any analysis of the relative steric environments of
quadrants carries several
potential sources of nonideality. Most importantly, chiral receptors are not
necessarily structurally invariant. Some have many
degrees of freedom and
might offer a different
rictus to different types
of alkenes. Alternatively,
consider a chiral metal
fragment with two orthogonal d orbitals that are
close in energy, both of
which can backbond with
an alkene C=C n* acceptor orbital. Some alkene
ligands may, for steric or
other reasons, adopt M(C=C) conformations orthogonal to others.
Also, alkene ligands
have a surprisingly large
number of degrees of freedom. First, the C=C linkage can be “rocked’
around either of the two
axes separating the four
quadrants in Scheme 3, as
well as the axis perpendicular to the plane of the paper. These correspond to
the roll, pitch, and yaw axes in aircraft navigationan appropriate metaphor
in view of the similarities
between a docking maneuver and Lewis acidlbase
complexation.
Alkene ligands can furthermore undergo “slippage”) 91 the primary
manifestation of which is
’
555
3. A. Gladysz and
REVIEWS
pressed interstice between the NO and PPh, ligands is the most
congested; that between the two smallest ligands, NO and
C,H,, is the least congested;[25]that between the two largest
ligands, PPh, and C,H,, is actually intermediate. These conclusions follow from numerous studies of conformational and
diastereomeric equilibria.[' 6*"1
The electronic properties of 1 and/or its Lewis base adducts
have been studied both theoreticallyrz2,'*I and experimentally.[291These data establish that I is a strong n donor, with the d
orbital HOMO shown in Figure 1. When unsaturated ligands
bind to I, they commonly adopt conformations that allow high
degrees of overlap of their acceptor orbitals with this d orbital,
even when this is sterically unfavorable.[30] The high HOMO
energy is derived from the poor x-accepting capabilities of the
PPh, ligand.[28a]The two orthogonal d orbitals that can overlap
with the nitrosyl ligand acceptor orbitals are much lower in
energy.
the lengthening of one metal-carbon bond relative to the other.
These bond elongations are particularly favored by strongly
carbocation-stabilizing substituents on the C=C unit, such as
nitrogen or oxygen donors. A slippage parameter can be defined
that has two intuitively logical limiting values: 1) 0 % when the
perpendicular from the metal to the C=C bond intersects the
bond at the bond midpoint, as in an equilateral triangle, and
2) 100% when the perpendicular from the metal to the C=C
bond intersects the bond at a carbon atom.[201
As expected from the Dewar- Chatt -Duncanson model and
backbonding, the ligating carbons in alkene complexes rehybridize, adopting considerable sp3 character. Thus, substituents
move out of the x nodal plane of the free ligand and away from
the metal fragment. This attenuates steric interactions between
substituents and the metal fragment. There are several ways to
quantify this feature.@'] In this review "bend-back angles" will
be employed. These are computed b y first defining a plane that
contains the C = C linkage but is perpendicular to the metallacyclopropane plane. The angle of the carbon-substituent bond
with this plane is then the bend-back angle. Bend-back angles
involving hydrogen atom positions determined by X-ray crystallography are not usually reliable.
6. Functional Equivalents of the Rhenium
Lewis Acid I
Adducts of I and an extremely broad spectrum of organic
Lewis bases have been isolated. Preparations begin with commercial [Re,(CO),,], which is converted to the chiral methyl
complex [(q5-C,H,)Re(NO)(PPh,)(CH,)] (rac-1) in four steps
and 57% overall yield.[3'] Resolution is easily effected en route
in two steps and 76%
As shown in Scheme 4 , 1 is then
5. The Chiral Rhenium Lewis Acid I
The chiral,
d6 sixteen-valence-electron rhenium fragment [(q5-C,H,)Re(NO)(PPh3)]+ (I) has, in various
synthetic and mechanistic contexts, captivated the attention of
the senior author for almost two decades.[23]Mass spectra show
that it can exist as a discrete species, but as described in Section 6
it is probably rarely generated in solution.[241Yet it can pass
from one organic molecule or Lewis base to another, or between
functional groups in a single molecule, always with retention of
configuration at rhenium.[''] Some structural features of I and its
adducts I1 and 111with generic Lewis bases are given in Figure 1.
B.J. Boone
C
Scheme 4. Generation of functional equivalents of the chiral Lewis acld
[(s5-c,H,)RefNOf(PPh,)1+ (1).
900
Figure I. Key steric and electronic features of the chiral rhenium Lewis acid
[(q5-C,H,)Re(NO)(PPh,)lt (I) and its Lewis base adducts I1 and 111.
Several steric properties of I deserve emphasis. First, the
PPh,, cyclopentadienyl, and nitrosyl spectator ligands differ
greatly in size (large, medium, and small, respectively). Second,
the coordination geometry at rhenium is formally octahedral;
the cyclopentadienyl ligand occupies three coordination sites.
Thus, the idealized ON-Re-P bond angle is 90", and angles from
the cyclopentadienyl centroid to the other three ligands are 125".
Third, from the perspective of a Lewis base, the angularly com556
transformed to the substitution-labile dichloromethane complex [ (q5-C,H ,)Re(NO)(PPh,>(ClCH,Cl)]+ BF, (2+ BF,)r321
or a related chlorobenzene complex (3' BF,).13,] These react
with a variety of neutral and anionic Lewis bases to give adducts
of I with overall retention of configuration at rhenium.
The chlorohydrocarbon complexes 2 + BF, and 3' BF, have
fascinating properties, many of which have been studied and
described in detail. Both the formation and substitution reactions of 2+BF; occur with retention at r h e n i ~ m . [ ~ ~ . ~ ~ ]
Above - 20 "C, enantiomerically pure 2+ BF, decomposes to
an enantiomerically pure bridging chloride complex. Consequently, it is impossible to measure a racemization rate, since
thermal decomposition occurs more rapidly. The chlorobenzene
complex 3+ BF, is a mixture of linkage and constitutional isom e r ~ . [Although
~~'
the relative populations of isomers vary as
samples are warmed from - 45 "C to 100 "C, all are converted to
Angew. Chem. Int. Ed. Engl. 1997.36,550-583
Chiral Recognition in
K
Complexes
REVIEWS
Lewis base adducts. Thus, the mixtures of isomers of 3' BF;
that undergo substitution will be different with strongly nuckophilic and weakly nucleophilic Lewis bases. In any event, adducts form with overall retention at rhenium regardless of reaction temperature.
Kinetic binding selectivities are a reflection of the mechanism
of adduct formation. Accordingly, the mechanisms of the reactions of 2' BF, and the indenyl (q5-C,H,) analog with cyclohexanone have been investigated in
In both cases, the
corresponding o ketone complexes form. Rate measurements
show that these substitutions are associative.The data also indicate that slippage of the cyclopentadienyl ligand does not occur
prior to or during the rate-determining step. Although there is
not a compelling choice among the remaining mechanistic possibilities, one plausible pathway is shown in Scheme 5.
(9-2' BFL
4' BFT
Scheme 5 One pos\ihle mechanism for dichloromethane substitution in 2 + BF;.
7. Complexes of I with Alkenes;
General Bonding Considerations
From the frontier orbital properties noted in Section 5,
alkene complexes of I should adopt the idealized Re-(C=C)
conformation shown in I1 (Figure 2). The labels used for the
U
II
I
m
I
IV
I
substituent positions in I1 (a-d) are maintained throughout the
review. It was anticipated that positions a and b, which are anti
to the bulky PPh, ligand, would be less congested than posit i o n s ~and d, which are syn. It was further anticipated that
position b, in which steric interactions with the cyclopentadienyl
ligand should be possible, would be more congested than
position a, which is proximal to the small nitrosyl ligand.
Finally, it was thought that position d, which projects into
the small interstice between the PPh, and nitrosyl ligands,
would be more congested than position c. Note that in principle,
the PPh, ligand contributes equally to congestion in positions c
and d.
Ten crystal structures of alkene complexes of I have been
executed. These closely approximate the idealized structure 11,
and key features are summarized in Figure 3. In 11, the angle
formed by the rhenacyclopropane plane and the Re-P bond is
0". The values in the crystal structures range from 2 to 23.5",
with an average of 16". In all cases the rhenacyclopropane plane
is rotated slightly counterclockwise from its position in 11. Thus,
there is a modest degree of freedom (ca. 20 ) about the
Re-(C=C) axis. The counterclockwise deviation might serve to
relieve steric interactions involving position d. Other geometric
measures of this feature are possible (for example the angle
formed by the rhenacyclopropane plane and the plane defined
by the cyclopentadienyl centroid, rhenium, and the C=C midpoint), and give comparable values and trends.
The bend-back angles involving the non-hydrogen C=C
substituents range from 13 to 28", with an average of 20.3'
(Figure 3 ) . There d o not appear to be any significant trends.
Slippage values are generally small.[20]Other features of these
crystal structures are analyzed in Section 18. Structures of these
complexes in solution have commonly been assigned from
NMR data. Some key trends are summarized in Figure 2.
The ethylene complex of I has been
The two C=C
13C NMR signals coalesce at 96.2"C (CDC1,CDClJ. Alkene
ligands normally dissociate from I only at much higher temperatures. Hence, a 180' rotation about the Re-(C=C) axis is
required to render the NMR signals equivalent. .4 barrier of
16.4 kcalmol-' (AG*, 369 K) can be calculated. Also, a
[DJethylene complex with the CD, label in positions c and d
can be generated by a [Re=CD,]+/CH,N, coupling reaction
at -95 0C.[361 The CH,/CD, termini rapidly scramble
above -4O"C, consistent with the barrier determined by I3C
NMR spectroscopy. No equilibrium deuterium isotope effect
was detected.
8. Complexes of I with Monosubstituted Alkenes
(R = Alkyl, Aryl, Silyl)
8.1. Kinetic Binding Selectivities
PPh,
ON
VI
6=.5.1 - 5.2
Figure 2 . Key properties ofalkene complexes ofI.11) Idealized Re-(C=C) conformation: 111) trends in 'H and I3C NMR chemical shifts for vinylic or allylic groups;
IV) trends in NMR coupling constants; V) 'H NOE enhancements, R = H, C H , R ;
VI) the effect of an aryl group in position b on the 'H NMR signal ofthe cyclopentadienyl ligand.
Anpm (%mr
lnr Ed Ennl 1997, 36. 550-583
In typical experiments, the dichloromethane complex 2+ BF,
and five to ten equivalents of simple monosubstituted alkenes
were combined at -80°C (Scheme 6).[35,371 Workup at
room temperature gave the corresponding alkene complexes
5a-c,e-h+ BF, in 63-99% yields as mixtures of configurational diastereomers. The diastereomer ratios of both crude and
purified products were assayed. These ranged from 63:37
557
J. A. Gladysz and B. J. Boone
REVIEWS
(RSSR)-Sc+ BF<
(RSSR>Sp BP;
(RS+7R>Si+BFr
C&d-
CH(CH&+
RIR'
i
(RR.SS)-Sc+ BPT
(RR.SS)-Sc+ PFC
CHXHCHd-
CJV-
CHzC,H+
18"
19.3"
w
n
Angle formed by the Re-(C=C)
plane and the Re-P bond
150
23.5"
C-R bend-back angle
15.4"
16"
19=
Slippage
7.6%
8.5%
16.4%
Ref.
1551
1371
1491
(RSRJRS>6m+ BF<
R R
(RSRSRS)-7.*
CH3/cH3
m
Angle formed by the Re-(C=C) 12.3"
plane and the Re-P bond
C-Rbend-backangle
Slippage
Ref.
PFC
i
(RSSSRRpJai+ B F i
19.4"
25"
6.5%
1.4%
1551
[351
i (RSRS,SRSR)-13+ TfOCHK(=O)CHzCH, i2-C&NHCH(CrjH~)-
(RSSSRR)-lO++ BFT
-(CHz)Aal
CHKK
8.8"
22.6"
21.9"
i
25.7. 19.3'
22.28"
18.2. 16.8"
18.1, 12.7"
-12.7%
1.4%
-2.8%
-12.0%
1451
1461
(451
(501
Figure 3. Key structural parameters of alkene complexes of 1. [a] $-C,H,CH,
1
20.7, 27.5"
i
1.0%
1531
instead of qs-C,HS
(RS,SR/RR,SS)[381
to 75 :25, except for the bulkier tert-butylethylene complex 5gf BF,, which gave a higher value (84:16).
The chlorobenzene complex 3' BF; afforded alkene complexes
with comparable diastereomer ratios, except for 5g+ BF,,
which gave an even higher value (96: 4). Throughout this review,
isomer ratios are normalized to 100, and error limits on each
integer are & 2 (e.g. 63 :37 = (63 & 2): (37 f2)) or less.[391
not measured) ,[401 and other compounds described in Sections
15 and 16. With Sa-h+ BF;, the equilibrium ratios range from
90: 10 to >99: 1 (RS,SR/RR,SS). Similar values are obtained
in CHCl,CHCl,, from syntheses conducted directly at 95100"C and from thermolyses of diastereomerically pure complexes. In all cases, mass balance is excellent. Rate measurements give AH' and AS* values of 29 kcalmol-' and
2 cal mol- K ' for the isomerization of the less stable styrene
(CD,ClCD,Cl;
AG*(369 K)
complex (RR,SS)-Se+BF,
28 k ~ a l r n o l - ' ) . [ ~ ~ ~
~
8.2. Thermodynamic Binding Selectivities
Configurational diastereomers of 5 + BF, equilibrate over
the course of 6-24 h in chlorohydrocarbon solvents at 95100°C. Scheme 7 gives data for all of the compounds in
Scheme 6 (C,D,Cl; ca. 0.01 M, lOOT), a 1,Cpentadiene
complex prepared subsequently (5d BF, ; kinetic selectivity
+
558
8.3. Interpretation
The kinetic binding selectivities for Sa-c,e-h+ BF, in
Scheme 6 are lower than the thermodynamic binding selectiviAngew. Chem. Int. Ed. Engl. 1997, 36, 550-583
Chiral Recognition in
Complexes
REVIEWS
Yield (RS,SFURR,SS ratio)
from Z+ B F i from 3’ EFi
a. R = CH3
b, A = CH2CH2CH3
C, R = CH&H5
e, R = GH5
1, R = CH(CH3)2
8. R = C(CH3)s
h, R = Si(CH3)3
95% (66:34) 90% (68:32)
94% (63137‘) 95% (6733)
88% (70:30) 80% (6733)
93% (75:25) 94% (80:20)
99% (64:36) 99% (62:38)
73% (8436) 8% (96:04)
63% (69:31) 9% (69:31)
Scheme 6. Synthesis of complexes of I with monosubstituted alkenes; kinetic binding selectivities [37]
R
ON
H PPh3
a, R
b. R
c, R
d, R
0.
- CH3131
131
-- CH2CH-C&1371(401
CH2CH&H3
97:W
CH,C.H~
97-03
98.02
-
90110
R CeY I31
f, R
--
CH(CH,), 1371
~99:l
>9%1
1, R C(CH& 131
h, R SI(CH& 131
1, R
1,
~99:l
CH-CHCH, 1491
R C(CH3)-CY I491
k, R
%a4
C(-0)H [So]
90110
98.02
DW1
I, R-C(-O)CHS[B]
94106
m,R C(=O)CyCH, [Sol
%:W
%:W
n, R - C(-O)C.HS 1361
0,
R
-
C(-O)OCkl$H,
(361 93:07
Scheme 7 Complexes of I with monosubstituted alkenes; thermodynamic binding
selectivities. The PF; \ a h of 5n‘ and So+ were equilibrated.
ties in Scheme 7. The rhenium and carbon stereocenters should
be much closer together in the complexes than in the progenitor
transition states. Consequently, the differences in energy between the diastereomeric transition states (AAG *) should be less
than those between the diastereomeric complexes (AG), giving
Angeu, c‘hem In!. Ed EngI. 1997, 36. 550-583
lower kinetic binding selectivities. Other alkene complexes of I
exhibit analogous trends, and kinetic isomer ratios are not further tabulated in this review. The fascinating mechanism by
which the equilibrations in Scheme 7 occur has been studied in
detail.r4’I In brief, alkene ligands with cis and truns deuterium
labels do not undergo scrambling. This establishes that the entire C = C enantioface is exchanged, not just that of the CHR
terminus. Surprisingly, equilibration occurs without alkene dissociation, and an intermediate ‘‘0bond complex” of I and a
vinylic hydrogen has been implicated.
The high RS,SR/RR,SS equilibrium ratios follow logically
from the steric environments of positions a-d. As shown in
Scheme 6, both diastereomers would be expected to favor
Re-(C=C) rotamers in which the larger CHR termini are
anti to the bulky PPh, ligands (RS,SR: VII>>IX; RR,SS:
VIII >> X). To date, no evidence has been found for a second
rotamer of any diastereomer of 5a-h+ BF, in low-temperature
NMR studies. However, Re-(C=C) rotamers are easily detected for complexes of I with disubstituted alkenes as described
below. Hence, the RS,SR/RR,SS equilibrium ratios will very
closely match the VII/VIII ratios (Scheme 7) and can be viewed
as measures of the relative steric environments of positions a
and b.C4’1
Several trends are evident in the RS,SR/RR,SS equilibrium
ratios for Sa-h+ BF,. For example, sp3-hybridized substituents
on the C=C unit that are unbranched at the allylic carbon
(Sa-d+ BF,) give binding selectivities of 96:4-98:2. With sp3hybridized substituents that are branched at the allylic carbon,
or trimethylsilyl ( S f - h + BF,), binding selectivities increase to
>99: 1. However, with a sp2-hybridized phenyl substituent
(5e+BF,), the binding selectivity decreases to 90: 10. For convenience, this and similar phenomena below will be termed
“small phenyl” effects.
Several rationales have been considered for the lower selectivity with Se+BF;. For example, a phenyl substituent might
enhance slippage relative to an alkyl substituent, increasing
the distance between the rhenium and CHR stereocenters.
This should diminish chiral recognition. The crystal structures of both diastereomers of Se+ BF, have been determined,
as well as of two analogs with alkyl substituents ((RR,SS)5c+ PF,, (RS,SR)-Sf+BF,; see Figure 3 ) . However, the
error limits on the bond lengths are too large for a n y conclusion. Regardless, data below show that other factors must contribute.
In this context, phenyl groups exhibit smaller effective steric
sizes than methyl groups in many types of equilibria.143]Also,
there is increasing evidence for attractive interactions between
sp2 carbon-hydrogen bonds and n electron clouds of unsaturated functional groups.r441This would uniquely stabilize the
higher energy diastereomer (RR,SS)-Se+B F , and is furthermore consistent with the ‘ H N M R shielding effect in VI
(Figure 2). Indeed, Brunner has previously proposed that the
hydrogen atoms of the cyclopentadienyl ligand can attractively interact with the n electron clouds of phenyl rings.[44a]
He is also, to our knowledge, the first to invoke this type
of phenomenon as a factor in a diastereomeric equilibrium.
This intriguing “small phenyl” effect will be further analyzed
as other equilibria are presented in subsequent sections (10- 12,
15, 19, and 20).
559
3. A. Gladysz and B. J. Boone
REVIEWS
9. Complexes of I with Symmetrically Substituted cis
Alkenes
9.1. Binding Data
Complexes of I with cis-2-butene, cis-3-hexene, cis-stilbene,
and cis-I ,Zdichloroethylene (6a-df BF;) can be prepared by
procedures analogous to those in Scheme 6.[451As summarized
in Scheme 8, NMR spectra recorded at sufficiently low temperatures show two diastereomers (RSR,SRS/RRS,SSR or XI/
XII) .[38b* Equilibrium ratios are lowest for cis-1,2-dichloroethylene complex 6d' BF, (59:41), highest for cis-stilbene complex 6c'BF; (93:7), and intermediate for 6a,b'BF;. With
6a+ BF,, a significant solvent effect is observed.
R 'PPh3
70~30(-60 'C. CD&)
8436 (-70°C,CDC13)
b. R CHzCH3 8535 (-62 "C, CDU3)
C. R iC&5
93:07 (-70 "c, CDCl3)
d. R = CI
59:41 (22 "C, CDC13)
a, R = CH3
p
XI'
7+8Fi
11.1
11.0-11.1
12.6-12.8
13.2
>17.5
XXI'
a, n+4 = 5
Ratio (CDCQ.-70 "C)
95:05
b, n+4 = 6
>98:02
d, n+4 = 8
85:15
c, n+4 = 7
246
241-245
270-274
280
2349
m a
Scheme 8. Complexes of 1 with symmetrically substituted cis alkenes; thermodynamic binding selectivities [45,46].
Complexes of I with cyclopentene, cyclohexene, cycloheptene, and cyclooctene (7a-d+ BF,) have also been prepared.[461
As summarized in Scheme 8, these give comparable mixtures
of diastereomer~.[~'I 2D NMR spectra show that the
diastereomers interconvert by Re-(C=C) rotations as opposed
to C=C "face flips".[451The former exchanges the substituents
of the C=C unit in positions a and c, and b and d, whereas the
latter exchanges substituents in positions a and b, and c and d.
9.2. Interpretation
The binding equilibria for the cis-2-butene and cis-3-hexene
complexes 6a,bf BF, in Scheme 8 (XIjXII) reflect the relative steric congestion in positions (a + d) vs (b + c) (cf. D in
Scheme 3). As analyzed in the preceding section, binding equilibria for complexes with monosubstituted alkenes 5a,b+ BF,,
560
which have comparably sized sp3-hybridized substituents on the
C=C unit, reflect the relative steric congestion in position a vs
b. Thus, if position d were to be more congested than position c,
binding selectivities for 6a,b+ BF; should be lower than those
for 5a,b+ BF,. Accordingly, 6a,bf BF; give lower isomer ratios than 5a,b+ BF; (84:16-85:15 in CDC1, at I - 62°C vs
96:4-97:3 in C,D,CI at 100 "C). Hence, the CHR termini of cis
alkenes are mismatched with respect to the binding properties of
the chiral receptor I.
The data in Scheme 8 also require that there be a larger difference in steric congestion for positions a and b than for posit i o n s ~and d. Otherwise, the stability order of the isomers
(XI > XII) would be reversed. Thus, for both 5a,b+ BF; and
6a,b+ BF, the dominant diastereomer is controlled by the difference in congestion in positionsa and b. However, for
6a,b+ BF;, the dominant diastereomer is destabilized with respect to the minor diastereomer because position d is more
congested than position c. This leads to the corollary that complexes of I with trans alkenes should show higher binding selectivities than 5a,b+ BF; , which is tested below. Curiously, the
cis-stilbene complex 6c+ BF, gives a higher binding selectivity
than the styrene complex 5e'BF; (93:7 vs 90:lO). However,
the difference is very slight, and additional anomalous trends
involving phenyl substituents are described in Sections 10- 12,
and 15.
The barriers to Re-(C=C) rotation given in Scheme 8 are
also easily rationalized. For 6a-c+ BF,, there is a slight increase with the size of the CHR substituent (CH,/CH,CH,/
C6H, 11.O- 11.1/12.6- 12.8/13.2 kcalmol- I ) , subject to the ambiguities with phenyl groups noted above. However, the
cis-l,2-dichloroethylenecomplex 6d+ BF, gives a much higher
barrier ( > 17.5 kcal mol- I ) . Accordingly, vinylic chloride substituents enhance the x acidities of alkenes, and x backbonding
diminishes along the reaction coordinate. Thus, the ground
state will be stabilized more than the transition state, raising the
barrier to rotation. The small size of the chloride substituents in
6df BF; accounts for the lower binding selectivity.
The data for 6a' BF; in Scheme 8 show that the solvent can
appreciably affect conformational equilibria that exchange positions of the substituents on the C=C unit. It would logically
follow that there could be solvent effects upon configurational
equilibria that exchange positions of C=C substituents. However, none have been found to date. One factor may be the
limited range of solvents suitable at the equilibration temperatures required (e.g. 95-100°C for Scheme 7). For example, a
significant amount of alkene ligand displacement occurs in acetonitrile. Alkene complexes of I are usually insoluble in hydrocarbons and ethers. Hence, only a few halohydrocarbons have
been employed as solvents.
10. Complexes of I with Symmetrically Substituted
trans Alkenes
10.1. Binding Data
Complexes of I with trans-2-butene, trans-3-hexene, and
trans-stilbene (8a-c+ BF,) can be prepared by procedures
analogous to those described above.r451However, these complexes form at much slower rates than the corresponding cis
Angeu.. Chern. Int. Ed. En@. 1997, 36, 550-583
Chiral Recognition in
7c
REVIEWS
Complexes
25
-
C6H5C'
R = CH3 AGS = 18.6 kdlmol(397 K, C605Cl)
fl = C6H5 AG* > 17.6 kcaVmol(393K,C6D&I)
(RSS.SRFI)
8+ BF;
b, R = CH2CH3
C, fl = C6H5
R
C6H5. AGS = 11.6 kcal/mol (258 K,CDC13)
(RRR,SSS)
case, one substituent on the C=C
unit is a methyl group. Although in
principle Re-(C=C) rotamers are
possible, equilibria should be even
more biased than with the complexes of monosubstituted alkenes
5+BF,. Hence, only the more
stable rotamer (sc) of each configurational diastereomer is depicted
(RS,SR: XV; RR,SS: XVI).
>99:1
98~02
Scheme 9. Complexes of I with symmetrically substituted trans alkenes; thermodynamic binding selectivities [45].
alkene complexes 6a-c+ BF;, indicating decreased alkene nucleophilicity. Mixtures of two configurational diastereomers are
obtained (89: 11-76:24). Importantly, as summarized in
Scheme 9,8a-c+ BF; exhibit thermodynamic binding selectivities (RSS,SRR/RRR,SSS or XIII/XIV >99: 1-98:2) much
higher than those of 6a-c+ BF;. In each series of diastereomers, Re-(C=C) rotational barriers can be determined by
NMR spectroscopy.
10.2. Interpretation
The binding equilibria for 8a-c+ BF; in Scheme 9 (XIII/
XIV) reflect the relative steric congestion in positions (a + c) vs
(b + d) (cf. C in Scheme 3). In accord with the prediction in the
previous section, the isomer ratios are greater than those of
related complexes with monosubstituted alkenes 5a,b,e+BF, .
Thus, the steric congestion in position d must be greater than
that in positionc, and the CHR termini of trans alkenes are
matched with respect to the binding properties of the chiral
receptor I.
The barriers to Re-(C=C) rotation in the more stable
RSS,SRR diastereomers (XIII; > 17.6 kcalmol-') are much
higher than those of the corresponding complexes with cis alkenes (Scheme 8; 11.1-13.2 kcalmol-'). Importantly, exchange
of the CHR termini in XI11 requires passage of a CHR moiety
over the bulky PPh, ligand. In contrast, the isomeric cis alkene
complexes XI and XI1 can interconvert without passage of a
CHR moiety over the PPh, ligand. However, the barrier to
Re-(C=C) rotation in the less stable RRR,SSS diastereomer
of trans-stilbene complex 8c+BF; (XIV) is much lower
(11.6 kcalmol-I). Since the phenyl substituents in this compound occupy the most congested positions on each C=C terminus, strain in the ground state likely lowers the barrier to rotation. Accordingly, (RRR,SSS)-8c+ BF; slowly isomerizes to
(RSS,SRR)&+ BF; at room temperature-much faster than
the isomerizations of complexes of monosubstituted alkenes in
Scheme 7.
11. Complexes of I with Geminally Disubstituted
Alkenes
11.1. Binding Data
Data for three complexes of I with geminally disubstituted
aikenes P a - c + BFT) are summarized in Scheme 10.[4sJIneach
Angen. Chem. In! Ed
Engl. 1997,36. 550-583
95 "C
xv
H3C
S+ BF;
(SC-RSSR)
b. R = CH2CHzCH3
XVI
(sc-RRSs)
68:32
Scheme 10. Complexes of I with unsymmetrically substituted geminal alkenes;
thermodynamic binding selectivities [48]. The higher energy uc Re-(C=C) rotamers are not depicted.
11.2. Interpretation
The XVjXVI equilibrium ratios can be viewed as direct measures of the relative steric environments of positions a and b. In
contrast to the similar situation with 5+ BF,, a non-hydrogen
substituent must occupy position b. As would be expected, the
data for the 2-methyl-1-butene and 2-methyl-I-pentene complexes 9a,b' BF; show that methyl groups partition into position b in preference to ethyl or propyl groups. However, the
differences in sizes are not sufficiently large to give high binding
selectivities.As the bulk of the non-methyl substituent increases
(for example to tert-butyl), binding selectivities are certain to
increase.
In contrast to the situation with the propene and styrene
complexes 5a,e+ BF;, the a-methyktyrene complex 9c' BF,
allows an internal as opposed to external comparison of the
relative sizes of methyl and phenyl substituents. Importantly,
any slippage induced by the carbocation-stabilizing phenyl
group should be equal in both diastereomers. Interestingly,
9c'BF;
gives a binding selectivity opposite to that of
9a,b+ BF,, with the methyl group preferentially in position a.
Hence, either a phenyl group exhibits a smaller effective size
than a methyl group in positions a and b, or there is a special
attractive interaction in position b.
12. Complexes of I with Unsymmetrically
Substituted trans Alkenes
12.1. Binding Data
Data for three complexes of I with methyl/alkyl- and methyl/
phenyl-substituted trans alkenes (10a-c' BF;) are given in
Scheme 11.r481As found for complexes 8' BFi with symmetri561
J. A. Gladysz and B..J. Boone
REVIEWS
effective size than the methyl
group. This suggests that a special attractive interaction may be
associated with position b.
95 "C,
XVII
XVIII
XIX
xx
(sc)
(ad
R = CH2CH3 76:24 (CDC13,20 "C)
R = C6H5 71:29 (CD2C12,-80 "C)[C]
R = CH(CH3)2
>99:1 (all conditions)
4555 (C~HSCI,20-95 "C)[b]
R = C6H5
R = C(=O)H
c1:99(all conditions)
R = C(=O)CH2CH3 29171 (CH2C12,29 "C)
13. Complexes of I with
Unsymmetrically
Substituted cis Alkenes
13.1. Binding Data
10' B F i
Data for two methyl/alkyl-substituted complexes of I with cis
b, R = CH(CH&
>99:1
alkenes (lla,b+ BF;) are given in
C, R =&H5
d, R = C(=O)H
>99:1
at the top of Scheme 12.[481One
configurational diastereomer is
slightly more stable than the othAGS [kcal/mol] = [a] 17.0 (363 K, C&,CI)
[b] >18.0 (393 K, CDC12CDC12)
er (RSR,SRS/RRS,SSR 59:41[C] 10.3 (218 K, CD2C12)
79:21). Two Re-(C=C) roScheme 11. Complexes of I with unsymmetrically substituted frans alkenes; thermodynamic binding selectivities
tamers
are observed for each con[48.50]
figurational diastereomer of the
cis-2-pentene complex 1 la' BF,
cal trans alkenes (Scheme 9), one configurational diastereomer
(XXIjXXII; XXIIIjXXIV) These interconvert with barriers
is much more stable than the other (RSS,SRR/RRR,SSS
comparable to those of the cis-2-butene and cis-3-hexene com2 99: 1). However, each configurational diastereomer can now
plexes 6a,b+ BF, . The other classes of complexes in Scheme 12
give two Re-(C=C) rotamers (XVIIjXVIII; XIXjXX). These
(12' BF,, 13' BF,) are treated in Section 17.2.
interconvert with barriers comparable to those of the corresponding isomers of 8' BF;. The other complexes in Scheme 11
are analyzed below.
13.2. Interpretation
(RSS,SRRj
12.2. Interpretation
For the more stable RSS,SRR diastereomers of 10a-c' BF,,
the rotameric equilibria (XVIIjXVIII) reflect the relative steric
congestion in positions a and c (cf. G in Scheme 3). In the trans2-pentene and trans-4-methyl-2-pentene complexes 10a,b+BF, ,
the methyl group preferentially occupies position a. Selectivities
are modest with 10a' BF, (65 :35, CH, vs CH,CH,), but high
with lob' BF, (>99:1, CH, vs CH(CH,),). The first ratio is
very close to the binding selectivity observed for the geminally
disubstituted alkene complex 9a' BF, (68: 32), in which methyl
and ethyl groups are partitioned between positions a and b.
Thus, the steric environments of positions b and c appear to be
comparable.
The /I-methylstyrene complex 1Oc' BF, gives a reversed selectivity, with a slight preference for the methyl group in position a and the phenyl group in position c. Thus, either a phenyl
group exhibits a smaller effective size than a methyl group in
positions a and c, or there is a special attractive interaction in
position c, for example between the phenyl rings of the PPh,
ligand and the phenyl ring of the alkene.
For the less stable RRR,SSS diastereomers, the rotameric
equilibria (XIXjXX) reflect the relative steric congestion in positions b and d. In the two complexes of which sufficient samples
could be obtained (lOa,c' BF,), there are moderate preferences
for methyl groups in the more congested position d. In the case
of 10c+BF,, the phenyl group no longer exhibits a smaller
562
(RRR,SSS)
These binding equilibria are complicated to analyze. Consider
first the more stable RSR,SRS diastereomer of the cis-2-pentene
complex lla' BF,. In the major rotamer (XXI), the larger ethyl
and smaller methyl groups are matched into the less congested
position a and more congested position d, respectively. In the
minor rotamer (XXII), the larger ethyl group is mismatched into
the slightly more congested position c. Thus, the XXIjXXII
equilibrium ratio (73: 27) is slightly greater than the XIjXII ratio
for the cis-2-butene complex 6a' BF, (70: 30), when measured
in the same solvent.
Conversely, in the major rotamer of the less stable RRS,SSR
diastereomer of lla' BF, (XXIII), the larger ethyl and smaller
methyl groups are mismatched into the more congested position d and less congested position a, respectively. In the minor
rotamer (XXIV), the ethyl group is matched into the slightly less
congested position b. Thus, the XXIIIjXXIV equilibrium ratio
(60: 40) is less than the XIjXII ratio for 6a+ BF;.
Rotamers of the configurational diastereomers of the cis-4methyl-2-pentene complex l l b + BF, cannot be detected by
low-temperature NMR spectroscopy. However, with the more
stable diastereomer, this is certainly because the XXIjXXII
equilibrium ratio is much higher than that of lla+BF,, as
would be expected from the greater difference in size between
the isopropyl and methyl substituents. Thus, a >99: 1 mixture
is indicated in Scheme 12. We speculate that with the less
stable diastereomer, rotamer XXIV may dominate. This should
occur when the non-methyl substituent becomes sufficiently
large.
Angen. Chem. Inr. Ed. Engl. 1997,36, 550-583
REVIEWS
Chiral Recognition in 7c Complexes
Scheme 12 is too small. Minor internal inconsistencies are not unexpected in view of the many
caveats discussed above and below.
I
14. Complexes of I with Trisubstituted
Alkenes
14.1. Binding Data
(RSR,SRS)
(RRS.SS@
Complexes of I with trisubstituted alkenes
having only alkyl or aryl substituents on the
C=C unit are much less stable. Parallel trends
are found for alkene complexes of other metal
Consequently, only the 2-methyl2-butene complex 14' BF; shown in Scheme 13
has been ~tudied.[~*l
To avoid decomposition,
equilibrations were conducted over extended periods at room temperature or below. Thermodynamic binding selectivities (RS,SR/RR,SS) were
bracketed as 91:6-94:9.
c-c
face-ffip
6090'C
neat
ligand
-
rotation
13' TfO-
J
-80°C
RaM
21:79
-1W"C
-30°C
2o:W
Solvent
C@Cl,
THF
CD3NOz
XXXI
(sc-RS~)
xxx
(acRSRS,SRS@
14'BFL
(,91:9-94:6)
XXW
(SC-RRSS)
Scheme 13. Complexes of I with trisubstituted alkenes; thermodynamic binding selectivities [48]. The higher energy ac Re(C=C) rotamers are not depicted.
4753
Scheme 12. Complexes of I with unsymmetrically substituted cis alkenes; thermodynamic binding
selectivities [48,50.53]
14.2. Interpretation
Additional generalizations about binding selectivities in this
class of compounds are difficult to make. As the non-methyl
substituent on the C=C unit becomes larger, XXI should become increasingly favored relative to the other three isomers
XXII-XXIV. However, with two very large substituents
(R, > RL.), an isomer analogous to XXIV might be the most
stable. This would avoid occupancy of the most congested position d. Receptors that give more clear-cut predictions for binding selectivities of unsymmetrically substituted cis alkenes are
illustrated below.
A XXI/XXIIjXXIII/XXIV equilibrium ratio can be calculated
from the data in
for the cis-2-pentene complex Ila'BF;
Scheme 12 (43:16:25:16). However, this is formulated from
measurements at markedly different temperatures and must be
analyzed with caution. For example, the XXI/XXIII ratio, 45: 25
or 63: 37 (normalized), should reflect the partitioning of methyl
and ethyl groups between positions a and d. The XXIIjXXIV
ratio, 16: 16 or 50: 50, would be an analogous measure for positions b and c. However, Schemes 10 and 11 give equilibrium
constants that directly reflect the partitioning of methyl and
ethyl groups between positions alb, a/c, and bJd. Careful analysis of these suggest that the XXI/XXIII ratio calculated from
Angeu. Chem. Int. Ed. Engl. 1997. 36, 550-583
It is assumed that XXXI and XXXII (Scheme 13) constitute
the dominant Re-(C=C) rotamers of each configurational
diastereomer. Hence, the RS,SR/RR,SS equilibrium ratio reflects the relative steric environments of positions c and d, and
shows the latter to be much more congested. The slightly greater
equilibrium ratio for the propene complex 5a' BF, (96:4,
Scheme 7) indicates that the difference in congestion between
positions c and d is less than that between positions a and b, in
accord with other data above.
These data further indicate that high binding selectivities can
be expected for complexes of I with trisubstituted alkenes
whenever the size of the geminal substituent that is trans to the
CHR group is greater than or equal to that of the other geminal
substituent. This allows the larger groups on each C=C terminus to occupy positions a and c.
15. Complexes of I with 1,3-Dienes
15.1. Binding Data
Unsymmetrically substituted 1,3-dienes are bifunctional
Lewis bases. Both trans-piperylene and isoprene complexes of I
563
J. A. Gladysz and B. J. Boone
REVIEWS
have been prepared ( 5 i j + BF,, Scheme 7) .[491 Consistent with
literature precedent>'4a1the monosubstituted C=C unit show
much higher thermodynamic binding affinities than the disubstituted C = C units (>99:1 and 97:3). As summarized in
Scheme 7, the monosubstituted C=C units give enantioface
binding selectivities (RS,SR/RR,SS)of 90: 10 and 98:2, respectively. These ligands, and others with X=C-C=X moieties below, can also exist as s-cisls-trans isomers. Such equilibria play
important roles in many enantioselective syntheses, and it is
frequently possible to assign the dominant isomer from spectroscopic data. These data, which are outside the theme of this
review, are given in the original papers cited.[49]
15.2. Interpretation
The trans-piperylene and styrene complexes 5i+ BF; and
5e+ BF, give identical binding selectivities (RS,SR/RR,SS
90: 10). In each case, the substituent on the C=C unit is spZ-hybridized. Alkene complexes 5a-d+ BF,, which have sp3-hybridized substituents, give higher selectivities. Thus, sp2-hybridized substituents may have intrinsically smaller effective
steric sizes. Alternatively, attractive cyclopentadienyl C-H . . . n
interactionsr441may be operative in the RR,SS diastereomers of
5e,i+ BF, .
As shown in Figure 3, in crystalline (RS,SR)-Si+BF, the
Re-CH, bond is shorter than the Re-CHR bond (2.16(1) vs
2.23(1) A). The resulting slippage value (16.4%) is the largest to
date for a complex of I with a monosubstituted alkene. As
analyzed above, this could also contribute to the reduced binding selectivity. However, both rhenium -carbon bonds appear
to be shorter than those of the other alkene complexes in Figure 3. This should enhance binding selectivity, contrary to the
trend observed. The contracted bonds may arise from the enhanced 71 basicity and n acidity of 1,3-dienes, as predicted from
simple Hiickel n MO theory.
When the unbranched I-propenyl substituent on the C=C
unit in 5i+ BF, is replaced by the a-branched 2-propenyl substituent in isoprene complex 5j+ BF,, the binding selectivity
increases to 98:2 (Scheme 7). However, this ratio is still less than
that of isopropylethylene complex 5 f +BF, (> 99: I), which has
an a-branched but sp3-hybridized substituent on the C=C unit.
16. Other Complexes of I with Monosubstituted
Alkenes (R = C(=O)X)
Regardtion metal Lewis acids that are stronger n
less, as summarized previously in Scheme 7, binding selectivities
have been measured for five complexes of monosubstituted
alkenes having C(=O)X substituents (5k-o+ X-).L36,501
16.2. Interpretation
The three a$-unsaturated ketone complexes 51,m' BF, and
5nt PF; give binding selectivities of 94:6-96:4 (RS,SRI
RR,SS).[36.
501 Consistent with other trends noted above, these
ratios appear to be slightly lower than those of alkene complexes
with comparably sized sp3-hybridized substituents on the C=C
unit. The a$-unsaturated ester complex So+ PF, gives the lowest binding selectivity in this series (93 :7).
Curiously, (RS,SR)-5o+PF, is the only complex of I with a
monosubstituted alkene for which Re-(C=C) rotamers have
been detected.r36]As expected, the equilibrium ratio is high
(VII/IX 93:7; Scheme 6). In view of the small effective size of
the ethoxycarbonyl substituent on the C=C unit, this ratio is
likely much higher for other complexes (RS,SR)-5+X - . Surprisingly, the NMR signals do not coalesce in spectra recorded
up to 160°C (CHCI,CHCl,). This bounds the AG* value for
Re-(C=C) rotation as 221.4 kcalmol-' (433 K). Since both
isomers give diagnostic JHpand Jcpcoupling constants for the
=CH 'H and 13CNMR signals (see IV, Figure 2), assignments
are unequivocal. The high barrier is provisionally attributed to
the electron-withdrawing nature of the ethoxycarbonyl g r o u p
similar to the effect of the chlorine substituents in the cis-1,2dichloroethylene complex 6d' BF, (Scheme 8). The normalized equilibrium ratio for the three observable isomers
(VIIjVIIIjIX 86: 7: 7; Scheme 6) indicates that the steric congestion in positions b and c is comparable.
Interestingly, the acrolein complex 5 k + BF, has the smallest
substituent on the C=C unit in this series of compounds, but
shows the highest binding selectivity (>99: 1) .r50* 5 2 1 Although
a simple explanation is not apparent, an electronic effect is certainly implicated. Also, 5 k - o + X- can give s-czs/s-trans isomers
about the O=C-C=C linkages. Perhaps these equilibria, which
will depend upon the C(=O)X group, can affect the RS,SR/
RR,SS ratios.
17. Other Complexes of I with Alkenes
Having Unsaturated Substituents
16.1. Binding Data
17.1. Complexes with trans Disubstituted Alkenes
Unsaturated carbonyl compounds such as enals and enones
constitute heterobifunctional Lewis bases. Interestingly, the carbony1 moieties are much more nucleophilic towards the
dichloromethane complex 2+BF, in all cases studied to
date.[501
Thus, O=C adducts of I are the exclusive kinetic products. However, isomerizations to C=C adducts occur at higher
temperatures. For all ligands except cycloalkenones, only C=C
adducts remain at equilibrium. Thus, kinetic and thermodynamic binding selectivities are completely divergent. Although
the bases for these phenomena are beyond the scope of this
review,[511the stability trend is even more pronounced in transi-
Data for two complexes with trans methyl and C(=O)X
(X = H, CH,CH,) substituents, lOd,e' BF;, are given in
Scheme 1l.[501
As observed for the other complexes of trans
alkenes, one configurational diastereomer is much more stable
than the other (RSS,SRR/RRR,SSS 2 99: 1 ) . The rotameric
equilibria (XVIIjXVIII) show that the C(=O)H moiety in crotonaldehyde complex lOd+ BF; has a much smaller effective
steric size than a methyl group, and the C(=O)CH,CH, moiety
in lOe+ BF; has a slightly smaller size than a methyl group.
Interestingly, the latter compound crystallizes as the less stable
Re-(C=C) rotamer (Figure 3).
564
Angew. Chem. In!. Ed. Engl. 1997. 36, 550-583
Chiral Recognition in II Complexes
REVIEWS
17.2. Complexes with cis Disubstituted Alkenes
Data for cycloalkenone complexes 12a,bt BF, are given in
Scheme 12 (middle)
Two configurational diastereomers are
observed but equilibrium ratios are modest (RRS,SSR/
RSR.SRS 63 :37-83 : 17), paralleling those of complexes
1 la,b+ BF, with unsymmetrically substituted cis alkenes. In the
more stable configurational diastereomers, Re-(C=C) rotamers that direct the smaller carbonyl substituent syn to the
PPh, ligand are more stable.
In connection with another project, a C = C adduct of I
and a 1.2-dihydroquinoline ligand, (RSRS,SRSR)-13+ TfO(Scheme 12, bottom), was prepared and crystallographically
characterized (Figure 3) .[531 With this compound, the relative
stability of the two Re-(C=C) rotamers was opposite to those
of other cis alkene complexes (XXX > XXXI). Also, the equilibrium ratio in CD,NO, differed considerably from those in
CD,CI, or THE Apparently, the two large substituents on the
C = C unit are most readily accommodated in positions b and c
which have intermediate steric congestion. Furthermore, there
is the possibility of an attractive cyclopentadienyl C - H ' 7
interaction[451involving the aryl substituent on the C = C unit in
t .
xxx.
17.3. Complexes with Allenes
Several allene complexes of I have been prepared.[541With the
simplest prochiral allene, H,C=C=CHCH,, very high binding
selectivity for one diastereoface of the unsubstituted H,C=C
moiety is observed. These somewhat more complicated equilibria can be analyzed by extensions of the bonding models
outlined above.[541
Figure 4. Structures of the cations of styrene complexes (RS,SR)-Se+ BF; (top left)
and (RR,SS)-5etBF; (top right). and a superposition of the two (bottom).
~
Figure 5 shows the structures of the cations of the cis-2butene complex (RSR,SRS)-6a+ BF; and the trans-2-butene
complex (RSS,SRR)-Sa+ BF, (Figure 3 and XI, XI11 in
Schemes 8 and 9).[451In principle, these differ only in the position of a methyl substituent on the C = C unit. However, in
contrast to the example in Figure 4, this involves the more
congested positions c and d . Accordingly, the structure of the
fragment I varies slightly more than in Figure 4. In particular,
the conformation of the PPh, ligand appears to respond to the
methyl group in position c, with the closest phenyl ring rotating
away in a clockwise direction. There is a reciprocal response of
a phenyl ring (pointing down in Figure 4) to a methyl group in
position d. Interestingly, the cyclopentadienyl ligands adopt
18. A Second Look at the Steric Properties
of Lewis Acid I
The preceding binding data for alkene complexes of I provide
a detailed map of the relative steric environments of positions a-d. However, an interlude to reassess key assumptions
or approximations is instructive. First, the crystallographic
response of I to various types of alkene ligands will be
scrutinized.
Figure4 shows the structures of the cations of the
diastereomeric styrene complexes (RS,SR)-Se+ BF,
and
(RR.SS)-Se' BF, (Figure 3 and VII, VIII in Scheme 7).["] The
eleven-atom ($-C,H,)Re(NO)(PPh,)(C=C) units in each are
virtually superimposable. Thus, the only significant difference is
the position of the phenyl substituent on the C = C unit. The
bend-back and Re-CH-C,H, bond angles in (RR,SS)-Se+ BF,
are slightly greater than those in (RS,SR)-Se+BF; (19.4" vs
15.4'; 120.1(5)0 vs 116.8(6)"), suggestive of net repulsive steric
interactions between the phenyl group and the cyclopentadienyl
ligand in the former. However, when the structure is viewed on
a stereoscopic viewing screen with atomic radii set at van der
Waals radii, the phenyl/cyclopentadienyl spatial overlap is
small. Accordingly, the AG value, 1.64 kcalmol- as calculated
from the 90: 10 equilibrium ratio at 100 " C , is modest.
Angru C l i m i . In1 Ed. EngI. 1997, 36,
550-583
Figure 5. Structures of the cations of crs-2-butene complex (RSR,SRS)-6at BF;
(top left) and rrans-2-butenecomplex (RSS,SRR)-Sa+ BF; (top right), and a superposition of the two (bottom)
565
REVIEWS
orthogonal conformations (Figure 5, bottom) that differ by approximately 36” (360 “/lo) rotations about the rhenium-cyclopentadienyl centroid axis.
Importantly, PPh, ligands have “propeller” chirality.[56]This
additional stereogenic element generates another set of
diastereomers. Such diastereomers normally interconvert very
rapidly in solution, and equilibrium ratios are likely comparable
for structurally similar types of ligands. However, only one type
of PPh, rotor will commonly crystallize. The structures in Figure 4 have identical propeller chirality, and the structures in
Figure 5 have identical propeller chirality. However, the helical
senses are opposite. The allylbenzene and piperylene complexes
(RR,SS)-Sc+PF, and (RS,SR)-5iCBF, also crystallize with
propeller chiralities like that in Figure 4, whereas the other
alkene complexes in Figure 3 crystallize with chiralities like that
in Figure 5. There is not an obvious relationship between alkene
type and PPh, chirality. Fortunately, this degree of freedom
does not appear to be able to invert the relative steric congestion
of any pair of positions (a-d). However, it remains a source of
nonideality, and may be a factor in some of the trends involving
phenyl or other unsaturated substituents on the C=C unit in
positions c and d.
J. A. Gladysz and B. J. Boone
ON
a. Ar-
effect of
Ar group
effect of
counteranion
configuration
effect of
concentration
19. Complexes of I with Aromatic Aldehydes;
Electronic Effects
Aldehydes and monosubstituted alkenes differ only at one
X=C terminus (X = 0 vs X = H,C) and are approximately
isosteric. Accordingly, the complexes of I with aromatic
aldehydes
I(q5-C,H,)Re(N0)(PPh,)(q2-0=CHAr)]+
BF;
(15’ BFq)[”] give configurational diastereomers and Re(X=C) rotamers analogous to those of the complexes with
monosubstituted alkenes 5+BF; (Scheme 6). In brief, equilibrium constants are similar. However, diastereomers of 15+BF;
rapidly interconvert below room temperature. The isomerization mechanism involves intermediate o complexes-a type of
energy minimum not available to alkene complexes 5 + BF; .1581
With p-methoxybenzaldehyde, the o isomer is actually more
stable at room temperature.
Under standard conditions (0.00071 M, CH,Cl,, 173 or
183 K), complexes 15a-g+ BF, give the binding selectivities (RS,SR/RR,SS or XXXIIIjXXXIV) summarized in
Scheme 14.[591Multiple determinations have been made, and
the small standard deviations are listed elsewhere.r591A profound electronic effect is obvious. Electron-withdrawing aryl
substituents, which enhance n acidity, give higher selectivities
(up to 97 :3). In contrast, electron-donating aryl substituents,
which diminish n acidity and enhance CJ basicity, give diminished selectivities (as low as 73:27). Hammett plots are linear,
with p values of 0.60 (183 K) and 0.46 (173 K).
We thought, perhaps optimistically, that these electronic
trends might reduce to a simple, one-parameter, structural explanation. Specifically, the distances between the rhenium and
carbon stereocenters might decrease in adducts of the more
x-acidic aldehydes.r601This would enhance destabilizing steric
interactions between the cyclopentadienyl ligands and aryl
groups in the RR,SS diastereomers (XXXIV), giving higher
RS,SR/RR,SS ratios and chiral recognition as observed. The
566
183K
973x3
86112
83:17
7a:22
73:27
75125
CFc
cr
effect of
temperature
1
lSe*BFJ
-
PPh,
173K
97103
a%11
8456
m:m
-
76:24
79121
7426
83117
84116
15d‘PF;
74126
76124
lSe*PFc
7030
7426
0.000709M
0.W743M
0.0179M
73127
77%
79:21
ai:is
83:17
76124
EX21
81119
83117
0.0558U
0.156M
rpq-
203
193
183
173
163
153
7212a
74126
74:Z
m:z4
-
85:15
76124
78122
a2:w
83x7
[a] CYcl,, 04074 M
(b] CHCCF, 0.00071M
Scheme 14. Complexes of I with aromatic aldehydes; thermodynamic binding selectivities 1591
rhenium-carbon bonds would not necessarily be equal in the
two diastereomers. However, they would undergo parallel
changes as substituents are varied. With the diastereomeric
styrene complexes 5e’ BF; (Figures 3 and 4), the Re-CHC,H,
bond lengths differ only slightly (2.258(9) vs 2.284(7) A).
Accordingly, the crystal structures of five complexes were
determined at room and/or low temperature (Figure 6). Each
compound crystallized as the more stable RS,SR diastereomer.
The rhenium-carbon bond lengths are plotted vs the RS,SR/
RR,SS ratios in Figure 7. Importantly, as the rhenium-carbon
bonds contract from 2.199(6)-2.184(5) A ((RS,SR)-lSf+PF;)
to 2.161(9)-2.157(5) 8, ((RS,SR)-15a+PF,), the binding selectivities increase monotonically from 75:25-79:21 (15f+BF,,
183 or 173 K) to 97:3 (15a’ BF,). Figure 7 can be viewed as a
“crystallographic map” of a chiral recognition event. By the
commonly employed criterion of three times the estimated standard deviation, the bond lengths in adjacent pairs of compounds
in Figure 7 are not significantly different. However, there is a
statistically rigorous overall correlation with the RS,SR/RR,SS
ratios (x2 test) .[591
The crystal structures recorded at room temperature are overlaid in Figure 8. Although some variations are evident, the posiAngew Chem. h t . Ed. Engi. 1997,36,550-583
Chiral Recognition in
71
Complexes
REVIEWS
C/
(R$SR)-lBc* PFS
1.331(6)A
T YCl
-80
16
16
-125
16
20.6%
20.1%
28.0%
29.9%
30.9%
m
Angle between the Re-(O=C)
plane and the Re-P bond
2.5"
1.9"
17.2"
20.0"
19.7O
C-Ar bend-back angle
20.5"
20.2O
19.2"
19.3O
20.0°
168.5(4)",-6.3(7)O
169.0(7)",-6.7(11)O
173.8(5)0,-2.5(8)0
171.9(4)",-3.3(6)0
171.6(5)0, -4.q7)O
Slippage
O=C-CzC torsion angles
PF;
(R$SR)-lSf' PF;
(RS,SR)-l5d*SbF;
1.3248)A
T I"C1
1.313(11)A
-80
16
-80
16
30.7%
31.6%
29.5%
33.0%
m
Angle between the Re-(O=C)
plane and the Re-P bond
12.8"
13.2"
6.6"
6.0°
C-Ar bend-back angle
17.5"
17.4"
17.5"
17.4'
176.0(6)0, -5.1(9)0
177.0(13)O,-5.3(21)O
164.9(5)", -13.7(8)O
166.6(12)", -12.3(19)'
Slippage
OS-C=C torsion angles
Figure 6. Key structural parameters of complexes of I with aromatic aldehydes [59]
tions of the ten-atom ($-C,H,)Re(NO)(PPh,)(O=C) segments
are virtually superimposable. Figure 6 shows that the bend-back
angles (20.5-17.4') and O=C-C=C torsion angles (165 to 177"
and - 14 to - 3') are also similar, indicating comparable aryl
group conformations. In all cases, the rhenium-oxygen bonds
(2.083(5)-2.046(3) A) are shorter than the rhenium-carbon
bonds (2.199(6)-2.157(5) A). This reflects the greater affinity of
the more electronegative oxygen for the electropositive rhenium, and the greater ability of carbon to support a partial positive charge. Although the slippage values (20-33 %) are larger
A n p u Chem Int Ed. Engl. 1997, 36. 550-583
than for complexes of I with monosubstituted alkenes (1 - 16 YO;
Figure 3), it should be noted that the rhenium-carbon distances
generally remain shorter.
Figure 8 also shows that complexes (RS,SR)-lSa-d,f+ X crystallize with two types of PPh, ligand conformations with
opposite propeller chiralities. Curiously, the propeller chiralities of the p-trifluoromethylbenzaldehyde and p-chlorobenzaldehyde complexes (RS,SR)-lSb,c+PF; differ from all other
complexes of I with aldehydes, thioaldehydes, and selenoaldehydes that have been structurally characterized to date
567
J. A. Gladysz and B. J. Boone
REVIEWS
"'i
t
901
'".\
15
70
X[%I^^
70'
2.16
-
2.17
2.18
d(Re-C) [A]
\
2.19
-.. 2.20
Figure 7. A "crystallographic map" of chirai recognition in complexes 15ad,f+ X-. The y axis gives the proportion of the RS,SR diastereomer in mol YO.The
bond lengths plotted along the x axis are obtained from crystal structure analyses
data
conducted at 16°C (top plot) and at low temperature (bottom plot). 0 (-):
recorded in solution at 183 K, n(---): data recorded in solution at 173 K. Error
bars correspond to one standard deviation.
(BF, > PF, > SbF,). This shows that chiral recognition can
be influenced by species formally exogenous to the Lewis acid/
Lewis base pair. However, there are no close contacts of the
anions and cations in the crystal structures. Significant counter
anion effects have been observed with adducts of chiral ammonium salts and chiral crown ethers.@''
Unexpectedly, binding selectivities are markedly concentration dependent (Scheme 14). One rationale would involve the
formation of aggregates in more concentrated solutions. However, racemic and enantiomericallypure samples of benzaldehyde
complex 1 9 ' BF,, which should aggregate differently, give identical diastereomer ratios. Also, the polarity of a medium becomes
increasingly affected by the solute at higher concentrations. Unfortunately, diastereomer ratios can only be measured in a small
number of solvents, all of which are chlorinated, due to a combination of freezing point limitations, insolubility (ethers, hydrocarbons), and reactivity (isopropanol). All of these parameters are
deserving of further study and interpretation, particularly in
view of the desirability of accurately comparing binding data
from different types of transition metal Lewis acids.
Surprisingly, electronic effects in chiral recognition have not
been investigated very extensively.[61JFurthermore, several fascinating electronic effects in transition metal mediated enantioselective syntheses have recently been reported.[62]In most
cases, the mechanistic bases for the observed trends are poorly
understood. The preceding data provide what we believe to be
the first electronic effects involving a common step in metal-catalyzed enantioselective reactions (x complexation) that can be
easily and intuitively interpreted!601 These findings lead to the
important general prediction that chiral recognition should be
enhanced when either the x acidity of the ligand or the x basicity
of the metal fragment is increased. To our knowledge, this represents a new approach to the optimization of chiral receptors,
which are most commonly designed and then modified based
upon steric principles.
20. Complexes of I with Aliphatic Aldehydes
Figuie 8 A superposition of the structures of the cations of the aromatic aldehyde
complexes shown in Figure 6.
(see Sections 20 and 21). However, we presume that the equilibrium ratios of propeller diastereomers are similar for all compounds in solution. We speculate that the different propeller
chiralities may disrupt monotonic trends in some of the geometric parameters in Figure 6. For example, the oxygen-carbon
and rhenium-oxygen bond lengths and the angles of the
rhenium-oxygen-carbon planes with the rhenium- phosphorus
bonds might have been expected to correlate with the aldehyde
ligand x acidity. However, these features involve the aldehyde
oxygen, which is proximal to the PPh, ligand and perhaps
slightly affected by the propeller chirality. Other possible contributing factors have been discussed. [ 5 b1
Since the configurational diastereomers of aromatic aldehyde
complexes 15' X- could be more easily equilibrated than those
of alkene complexes 5' X-, effects of other variables were examined. As exemplified in Scheme 14, binding selectivities increase somewhat at lower temperatures, as would be intuitively
expected. Selectivities also vary slightly with the counter anion
568
Binding selectivities for complexes of I with aliphatic aldehydes, 16+ BF,) ,[631are particularly difficult to determine.
[($-C,H,)Re(NO)(PPh,)(~2-O=CHR)]+ BF,
16' BF,
Measurements are complicated by the low temperatures required for decoalescence of NMR signals and the very small
quantities of the less stable RR,SS diastereomers present at
equilibrium. Data acquired to date are summarized in
Scheme 1S.[59b1 These establish the expected steric effect upon
binding selectivities, with RS,SR/RR,SS (XXXVjXXXVI) ratios increasing from 99.0: 1.0 for acetaldehyde complex
16a' BF, (unsubstituted tl carbon) to 99.8:0.2-99.5:0.S for
propionaldehyde and butyraldehyde complexes 16b,c+ BF;
(unbranched a carbons), to > 99.9:O.l for isobutyraldehyde
and pivalaldehyde complexes 16d,e' BF, (branched tl carbons). However, propionaldehyde appears to give a higher
diastereomer ratio than butyraldehyde.
Regardless, aliphatic aldehydes bind much more selectiveIy to
I than aromatic aldehydes. As summarized in Figure 9, the
crystal structures of the acetaldehyde, propionaldehyde, and
Angew. Chem. In!. Ed. Engi. 1997,36,550-583
REVIEWS
Chiral Recognition in n: Complexes
&
A
bonds than isopropylethylene and allylbenzene complexes
phenylacetaldehyde
complexes
(RS,SR)(RS,SR)-5f+ BF, and (RR,SS)-Sc+ PF;, which give higher
rapid
binding selectivities. Other ligands for which rhenium-carbon
16b,c,ff PF; have been
52yc
R
- H
determined.[631 These
bond lengths do not correlate with binding selectivities are noted in Section 21. Therefore, we attribute the lower binding selecON
PPh3
exhibit
a
very
high
dePPhi
ON
gree
of
homology,
as
iltivities
with aromatic aldehydes to a “small phenyl” effect,
XXXV
16’BF:
XXXVI
derived from slight stabilizing interactions between the aryl
(ac-RS+SR)
(OC-RRSS)
lustrated by the tabulat(a, R = CH, (591
99.0:1.o
ed data and overlaid
substituent and the cyclopentadienyl ligand in the RR,SS
b, R = CkCH, [59]
99.8:0.2
structures. The PPh,
diastereomers and/or steric properties (see Sections 8, 11, 15).
C, R = CHzCHzCH, [59] 9950.5
ligand conformations
d, R = CH(CH&..1591 >99.9:0.1
S9.9:O.l
are close to those of the
a, R
______________.____.--.--= C(CH& [59]
g, R = CH=CHz [So]
>99:1
aromatic
aldehyde
21. Other Complexes of I with Aldehydes or Ketones
(RS,SR)complexes
Scheme 35. Complexes of I with aliphatic and
15a,d,f+x-in Figure 6.
Additional types of aldehyde complexes of I have been studother aldehydes; thermodynamic binding
The distances between
ied. First, an O=C adduct of acrolein, 16g+BF,, has been
selectivities.
the rhenium and carbon
isolated (Scheme 15).[501 This is a linkage isomer of (and precurstereocenters range from 2.15(1) to 2.155(12) A. Importantly,
sor to) the C=C adduct 5k’ BF, in Scheme 7. By analogy to
these constitute three independent determinations and allow an
complexes of I with monosubstituted alkenes having sp2-hy“average” rhenium-carbon bond length of 2.15 A, which is
bridized vinyl substituents on the C=C unit (5i,j+ BF,,
statistically shorter than that in the aromatic aldehyde complexes,
Scheme 7) the binding selectivity of 16g’ BF, would be expectto be confidently assigned. When this value is extrapolated on the
ed to be lower than that of the propionaldehyde complex
“crystallographic map” in Figure 7, > 99: 1 RS,SR/RR,SS equi16b+ BF,. However, careful NMR analyses under a variety of
librium mixtures are predicted.
conditions show no evidence for a second isomer. Thus, we
presume that 16g+ BF, exists as a > 99: 1 RS,SR/RR,SS equilibrium mixture. The linkage isomer 5k’ BF& also exhibits an
anomalously high binding selectivity. Hence, some selectivityenhancing factor is apparently associated with acrolein (pre2.155( 12) A
sumably electronic) that is applicable to both O=C and C=C
binding
modes.
R
PFG
A glyoxal complex of I, 16h+ BF,, has also been prepared
(RS,SR)-lGfPFG
(Scheme 15).[641 In solution, 16h+ BF, always occurs as a 95 : 5
R = CH2CH3
mixture of RS,SR/RR,SS diastereomers. This is presumed to
R = CHzCH2CHs
R = CH&HS
be an equilibrium value. However, the NMR signals do not
coalesce at 100°C (CD,CN). This bounds the AG* value for
isomerization as 220.0 kcalmol-’ (373 K). The high barrier is
O=C bond
1.35(1)Al 1.338(5)A11.318(11)A
likely due to the superior TC acidity of glyoxal. Isomerization
angle, Re-O=C plane
17.0’1 20.5O 122.8’
barriers for aromatic aldehyde complexes 15’ BF, exhibit a
pronounced
electronic effect, with values increasing from
19.2”1 19.3’1 19.0’
C-Rbend-back angle
8.5 - 8.6 kcal mol- for the p-methoxybenzaldehyde complex
15g’ BF, to > 16.6 kcalmol-’ for the pentafluorobenzaldehyde complex 15a’ BF,.r58, 591
Diastereomerically pure (RS,SR)-16h+ BF, can be crystallized. The crystal structure shows a rhenium-carbon bond slightly shorter than those of the other aldehyde complexes in Figures 6, 7, and 9 (Re-CHR 2.129(5) A, R e - 0 2.056(3) A). This
would be expected to enhance the binding selectivity-contrary
to the data. As discussed for similar cases (see Sections 16, 17), a
smaller effective steric size of the sp2-hybridized C(=O)H substituent, or attractive interactions involving the cyclopentadienyl
ligand in the RR,SS diastereomer, can be invoked. However, the
Figure 9. Summary of crystallographic data for complexes of I with aliphatic aldeapproximately isosteric acrolein complex 16g BF, behaves difhydes. [63]
ferently.
Although the preceding correlation gives a good agreement
Formaldehyde, thioformaldehyde, and selenoformaldehyde
with experiment, there are several hints that it is fortuitous. For
complexes of I have also been synthesized and their crystal
example, the crystal structures of the alkene complexes in Figstructures determined.r651The Re-(X=C) conformations are
ure 3 do not show any obvious bond length trends. The styrene
analogous to those in the aldehyde complexes above. The rhenium-heteroatom bond lengths progressively increase, as would
and piperylene complexes (RS,SR)-5e+ BF,,
(RR,SS)5e’ BF,, and (RS,SR)-5i’ BF,, which give lower binding selecbe expected from size and diminishing electronegativity (Retivities in solution, do not exhibit longer rhenium-carbon
CH,2.108(18), 2.199(8), 2.173(6) A;Re-X2.036(11), 2.381(2),
I
II
I
’
+
Angen. Chem. Int. Ed. Engl. 1997,36, 550-583
569
J. A. Gladysz and B. J. Boone
REVIEWS
2.522(1) A). Recently, Schenk has isolated thiobenzaldehyde,
p-chlorothiobenzaldehyde, and p-methoxythiobenzaldehyde
complexes of I.[661Low-temperature NMR spectra show only
one set of resonances, and a crystal structure of the thiobenzaldehyde complex establishes relative rhenium/carbon configurations identical to those in the more stable complexes (RS,SR)15+X- (Re-CHC,H, 2.207(9) A, Re-S 2.376(3) A). However, two isomers of the analogous thioacetaldehyde complex
can be detected at - 60 “C (96:4). This suggests that isomers of
the thiobenzaldehyde complexes might be observed at still lower
A 71 complex of I and a symmetrical ketone, 1,3-difluoroacetone, has been isolated and crystallographically characterized (Re-C(CF,), 2.17(1) A; Re-0 2.044(6) A).[671However,
due to steric and electronic factors analyzed elsewhere, nearly all
ketones give (T as opposed to n adducts of I. Regardless, based
upon the data in Scheme 10 for complexes with unsymmetrically
substituted geminal alkenes (9’ BF,), modest binding selectivities would be expected, except when the carbonyl substituents
differ markedly in size. Early transition metal Lewis acids and
fragments that are stronger n donors than I show a much greater
preference for n ketone complexes.[51b.
22. General Analysis of Chiral mType Receptors for
Alkenes, Aldehydes, and Ketones
The approach detailed above provides a paradigm for mapping the steric properties of chiral receptors for q 2 n ligands. It
would be of obvious interest to compile similar binding data for
other chiral transition metal Lewis acids and compare selectivities for various types of ligands. Some relevant equilibria have
been reported and are summarized below. However, it is of
equal or greater importance to generalize the key concepts embodied in the preceding analysis. In particular, we seek predictive and testable models for binding selectivities with other
classes of chiral receptors.
Thus, we turn from the rhenium Lewis acid I and the accompanying positions a-d, to the generic Lewis acid in Scheme 3
with the attendant quadrants A-D. Importantly, only six types
of chiral receptors are possible for q2 n ligands, as classified by
the relative steric environments of the four quadrants. These are
illustrated in A-F in Figure 10. A lower bar indicates a less
congested quadrant, and a higher bar indicates a more congested quadrant. A simple integral step function is utilized for the
relativc heights (I-4), giving a sum of ten arbitrary units. In
molecular mechanics parlance,[17b1
the heights of the bars height
can be viewed as “steric energy”. To facilitate comparisons,
quadrant A is defined as the least congested. The bars in the
other quadrants then differ in the pattern of ascent. Enantiomeric receptors exist but are not depicted.
Although the analysis in Figure 10 is simple, we are unaware
of any previous comprehensive treatments that encompass all
six cases. Many refinements can be envisioned. For example, the
relative steric environments of the quadrants certainly do not
have to correspond to integral step functions. However, even at
this primitive level, a number of testable predictions are easily
generated. For instance, receptors C and D, in which the two
least congested quadrants are trans, should give higher binding
570
/-
I,,,,(
trans 3 v s 7
trans 5 vs 5
p q
trans 4 v s 6
Figure 10. Types of chiral receptors for alkenes, classified by the relative steric
properties of each quadrant; the values in the boxes are relative energies of the
isomers.
constants with trans-2-butene than cis-2-butene-contrary to
the generalization in most textbooks.[’4a1
Consider receptors for symmetrically substituted trans alkenes, which will give two configurational diastereomers but degenerate Re-(C=C) rotamers (Schemes 2 C and 3C). Binding
selectivities should decrease in the order C,D > A,F > B,E. In
the most selective set of receptors (C, D), the two least congested
quadrants are trans (sum of bar heights = 3), favoring one
diastereomer. The two most congested quadrants are also trans
(sum of bar heights = 7), disfavoring the other (energy difference of four units). In the least selective receptors (B, E), the
least and most congested quadrants are trans (sum of bar
heights = 5 ) , and the sterically intermediate quadrants are trans
(sum of bar heights = 5). Thus, binding either enantioface results in comparable steric interactions.
Receptors for symmetrically substituted cis alkenes can be
analyzed in the same way, and binding selectivities should decrease in the order E,F > B,D > A,C. In the most selective set of
receptors (E, F), the two least congested quadrants are cis (sum
of bar heights = 3) and the most congested quadrants are cis
(sum of bar heights = 7). In the least selective receptors (A, C),
the least and most congested quadrants are cis (sum of bar
heights = 5), and the sterically intermediate quadrants are cis
(sum of bar heights = 5 ) .
Unsymmetrically substituted alkenes require slightly more involved analyses. With monosubstituted alkenes, the most stable
of the four possible configurational diastereomers and Re(C=C) rotamers will always be that with the CHR substituent
in the least congested quadrant A. Thus, some degree of binding
selectivity is always expected. To optimize enantioface binding
selectivity, the key energy difference to maximize is that between
the more stable Re-(C=C) rotamer of one configurational
diastereomer and the more stable rotamer of the other. Here,
receptor D provides a particularly favorable architecture. In one
diastereomer, the substituent can occupy either of the two least
congested quadrants (A or C; bar heights = 1, 2). In the other,
the substituent can occupy either of the two most congested
quadrants (D or B; bar heights = 3, 4). Thus, there will be a
energy difference of two units (1 vs 3, quadrants A vs D).
An identical result is obtained with receptor C, in which the
steric environments of quadrants D and B are reversed. At first
Angew. Chem. Int. Ed. EnxI. 1997. 36, 550-583
REVIEWS
Chiral Recognition in R Complexes
glance, receptor D might seem to
be superior, as the least and most
congested quadrants (A, D) have
ON
PPhQ
a geminal relationship. However,
this only serves to maximize the
energy difference between the
more stable Re-(C=C) rotamer
of the more stable configurational diastereomer and the corresponding rotamer of the least
not normalized
stable diastereomer. Similar
or
not normalized
analyses of the other receptors
give energy differences of only
one unit. However, binding selectivities can still be high, as with I
(see also Section 23). Parallel
conclusions are easily reached for
aldehyde ligand receptors.
Figure 11. Some modifications of the rhenium Lewis acrd I: steric consequences
With unsymmetrically disubstituted alkenes, the relative size
of the substituents (RL,R,) must be considered. As R, becomes
infinitely large relative to R,, the binding situation asymptotically approaches that analyzed for monosubstituted alkenes.
Although quantitative treatments of intermediate cases are
many interesting qualitative observations are possible. For example, consider cis alkenes and receptors C, D, and
F. In one configurational diastereomer, the R, and R, groups
can partition into the less congested and more congested quadrants in both Re-(C=C) rotamers. However, in the other configurational diastereomer, there will be mismatches in both rotamers. Thus, receptor C will always give enantioface binding
selectivities superior to those of its counterpart A. Similarly, D
will be superior to B, and F will be superior to E. With geminally
disubstituted alkenes, a similar situation exists for receptors A,
C, and D, and counterparts B, E, and F.
Figure 12. Space-fillingmodels of 1(top) and I-Me, (bottom) with atoms at van der
Although many other relationships and predictions can be
Waals radii, taken from the crystal structures of propionaldehyde complexes
derived from Figure 10, the preceding are representative. The
( R S , S R ) - ~P~ ~ +
;and (RS,SR)-~~~-M~. + PF;, respectively
application to rhenium Lewis acid I and derivatives thereof are
examined next.
trans alkenes with even higher enantioface binding selectivities
(minimum 98:2). Many types of trisubstituted alkenes should
also be bound with high selectivities. However, cis and unsym23. Experimental Evolution of the Receptor Model;
metrically substituted geminal alkenes generally give lower
Modification of Lewis Acid I
binding selectivities.
23.1. General Features of Lewis Acid I
With I, a “small phenyl” effect is observed in equilibria that
involve
position b. Although steric interactions of the cyclopenThe binding selectivities given above clearly show that the
tadienyl
ligand and groups in position b are unfavorable, there
chiral rhenium Lewis acid I belongs to receptor type A. Howis
apparently
an ameliorating interaction involving cyclopentaever, some data suggest that the steric environments of posidienyl
C-H
bonds
and the R electron clouds of aryl and other
tions b and c are likely quite close, as illustrated in the modified
unsaturated
s
u
b
~
t
i t u e n t sThis
. ~ ~ phenomenon
~~
also seems to be
quadrant diagram in Figure 11. Furthermore, as elaborated in
manifested
to
varying
degrees
in
positions
c
and
d, presumably
Section 24.4, such diagrams can give misleading impressions of
involving phenyl C-H bonds of the conformationally flexible
sterically homogeneous environments within each quadrant.
PPh, ligand.
For reference, space-filling representations of I, taken from
si
the crystal structure of propionaldehyde complex (RS,SR)16b+ PF; (Figure 9), are given in Figure 12.
Of the receptors in Figure 10, A clearly represents a good
choice for giving appreciable binding selectivities with a broad
range of alkenes. For example, I binds monosubstituted alkenes
with high enantioface binding selectivities(minimum 90: 10) and
Angen‘. Chem. Ini. Ed. Engl. 1997, 36, 550-583
23.2. Modifications of I
The preceding analysis suggests several structural modifications of I that may give improved binding selectivitiesfor certain
types of alkenes. In particular, the cyclopentadienyl ligand
571
J. A. Gladysz and B. J. Boone
REVIEWS
could be replaced by the bulkier and more electr~n-releasing[~~~dominant. Thus, to some extent a “small phenyl” effect seempentamethylcyclopentadienyl ligand. Electronically, the resultingly operates in position b of the Lewis acid I-Me,, where
ing Lewis acid I-Me, would be a somewhat weaker acceptor and
cyclopentadienyl C-H . . n interactions are not possible. Restronger donor, but with frontier orbitals analogous to those of
gardless, the steric environment in quadrant B has become com1. However, quadrant B would now be much more congested,
parable to that in quadrant D.
and quadrant C would be slightly more congested (Figure 11).
Modifications of I-Me, can in turn be considered. For exAlthough it would be difficult to predict the relative steric
ample, the substitution of PPh, by PMe,, which has a smaller
congestion in quadrants B-D in advance, it is immediately obcone angle, would give “I-Me,” (Figure 11). This Lewis acid
vious that binding selectivities for monosubstituted alkenes and
would be less congested in quadrants C and D. In the event that
unsymmetrically substituted geminal alkenes should be enquadrant D is in turn less congested than C, the two most spahanced. Data for the 1-pentene and styrene complexes of I-Me, ,
cious quadrants are cis (A, D), but maintain an appreciable
5b,e-Mel BF,, are given in Scheme 16.[711As anticipated, the
steric difference. This receptor, which is a variant of F in
RS,SRIRR,SS equilibrium ratios are much higher (> 99: 1).[721
Figure 10, should be particularly favorable for the selective
Also, low-temperature NMR spectra of n aldehyde complexes
binding of unsymmetrically substituted cis alkenes such as cis-2of I-Me, show only one dia~tereomer.~’~]
pentene. If quadrant C is less congested than D (variant of receptor D), high selectivities are also possible. When quadrants C
and D are sterically comparable, monosubstituted, trans, and cis
alkenes with identical CHR substituents should give nearly
identical binding selectivities.
Some adducts of I-Me, are not as configurationally stable as
major Re-(C-C)
those
of I.[7 However, the trifluorinated pentamethylcyclopenrotamer
tadienyl ligand, C,Me,CF,, is now readily a~ailable.”~]
This
ligand has an inductive effect similar to that of the cyclopentadienyl ligand, which may enhance configurational stability. However, the steric bulk is identical with that of pentamethylcyclopentadienyl. Thus, the Lewis acid [ ($-C ,Me,CF,)Re(NO)(PPh,)]+ should give alkene binding selectivities comparable to those of I-Me, and possibly greater configurational stability. The electronic properties could be optimized by increasing or decreasing the number of fluorines.
ON
PPh
24. Further Refinements in the Receptor Model
24.1. Receptors with C, Symmetry
Scheme 16. Complexes of &Me, with monosubstituted alkenes; thermodynamic
binding selectivities 1711.
AS would be expected, the propionaldehyde adduct crystallizes as the RS,SR dia~tereomer.[’~’
The orientation of the propionaldehyde ligand is close to that in the cyclopentadienyl
analog, except for an roughly 120” clockwise change in conformation about the =CH-CH, bond. Space-filling representations of the I-Me, moiety in this structure are given at bottom
of Figure 12 and allow the greater bulk of the pentamethylcyclopentadienyl ligand to be easily visualized. There appears to
be a slight secondary effect upon the phosphorus-carbon conformations in the PPh, ligand.
The less stable RR,SS diastereomers of 5b,e-Mejf BF, exhibit some unique characteristics. Low-temperature NMR spectra do not show distinct Re-(C=C) rotamers. However, 13C
NMR spectra of the I-pentene complex (RR,SS)-Sb-Me,’ BF,
indicate that the rotamer with the CHR terminus syn to the
PPh, ligand (ac) is dominant-in contrast to the corresponding
adducts of I (Scheme 6). However, with the styrene complex
(RR,SS)-Se-Me,’ BF;, the opposite rotamer (sc) appears to be
Chiral reagents and catalysts with C, symmetry are frequently employed in enantioselective syntheses. This popular design
element minimizes the number of possible competing diastereomeric transition states, facilitating analyses. With C, receptors, the two pairs of trans quadrants (A and C, B and D) will
have identical steric environments. As a result, M-(X=C) rotamers become degenerate. This greatly reduces the number of
possible isomeric alkene complexes in Scheme 2. In fact, isomers
are no longer possible with symmetrically substituted cis or
geminal alkenes.
Two possible steric profiles are illustrated in G and G in
Figure 13. These are distinguished by the difference in congestion between the pairs of trans quadrants (sum of bar heights 2
vs 8, or 4 vs 6). It is apparent that as the difference becomes
greater, enantioface binding selectivities for monosubstituted
alkenes, unsymmetrically substituted geminal and cis alkenes,
and all types of trans alkenes, will increase.
24.2. Binding Thermodynamics
In Figure 10, the sum of the heights of the four bars is normalized to ten arbitrary units of steric energy. However, other reAngew. Chem. Inf. Ed. Engl. 1997,36,550-583
572
Chiral Recognition in
7c
Complexes
C;! symmetry
REVIEWS
attractive interaction
2”
trans 2vs0
I
this feature can also be arbitrarily introduced in all receptors by
slightly raising the energetic “floor”. Conversely, special repulsive interactions that do not have a steric origin (for instance
dipole-dipole) may sometimes need to be considered.
24.4. Environments within Quadrants, and Chiral Ligands
H
trans 4vs6
variable bindina affinities
environments within auadrants
I
I
- 1
disfavored
I
Figure 13. Extensions and refinements of the receptor model in Figure 10
ceptors might give intrinsically weaker or stronger binding.
Higher bars could represent receptors with poorer thermodynamically weaker binding strengths. For example, the bars in A‘
in Figure 13 are normalized to twenty arbitrary units, which
could indicate a binding energy half of that of A with respect to
a calibrating ligand. Conversely, the bars in A in Figure 13 are
normalized to five arbitrary units, indicating a binding energy
twice that of A.
24.3. Attractive Interactions
Figure 10 was formulated in the context of simple repulsive
steric interactions. However, some receptors may be capable of
special attractive interactions with certain types of X = C substituents. These numerous possibilities would include hydrogen
bonding, dipole-dipole, and charge transfer interactions. The
“small phenyl” effect with I is likely due, at least in part, to
attractive interactions involving the cyclopentadienyl ligand.
However, in this example unfavorable steric interactions are
only ameliorated. In extreme cases, the situation illustrated in
receptor H (Figure 13) might be realized. Here, a strongly attractive interaction is indicated by a negative energy. However,
Angen. Chern. Int. Ed. Engl. 1997, 36. 550-583
For most chiral transition metal Lewis acids, the steric congestion in each quadrant will diminish with increasing distance
from the X=C binding site. This is illustrated in quadrant D of
receptor A ” (Figure 13). However, in some specialized situations, the reverse may occur, as depicted by the “steric wall” in
quadrant A of A “ . Similarly, “pockets” commonly surround
enzyme receptors, allowing substrates of only a limited size
range to be bound. These might be encountered, for example, in
metal complexes with dendrimer-type l i g a n d ~ , [ ~or~ ”macro]
cyclic ligands containing “picket fences”.[75b1
Although the previous two examples are adequate for some
purposes, a more general approach would be to subdivide each
quadrant into subquadrants. This would allow chiral environments within quadrants to be represented. In this context, the
binding equilibria described above generally involve prochird
ligands. However, if receptors are to be designed to selectively
bind one enantiomer of a chiral ligand, such additional levels of
analysis are required. For example, consider a monosubstituted
alkene with an allylic stereocenter (H,C=CHCHR,R,) as discussed in Section 3.3. Assume that the lowest energy conformation about the =CH-C bond always has the allylic hydrogen
pointing “down”, as illustrated in Figure 13. A receptor such as
A“‘ would then selectively bind the enantiomer in which R,, not
R,, abuts the “steric wall”.
In unpublished studies, several complexes of the chiral rhenium Lewis acid I and chirdl monosubstituted alkenes and aldehydes of the type X=CHCHR,R, have been prepared. The
rhenium precursors were enantiomerically pure, whereas the
ligands were racemic. However, in no case was a significant
preference ( > 67 :33) for binding one enantiomer observed. Actually, this is to be expected from the data in Scheme 6. Kinetic
binding selectivities are almost always modest, and coordinated
ligands exchange with free ligands only at very high temperatures, if at all. In order to attain equilibrium with respect to the
CHR,R, stereocenter, exchange with a pool of racemic ligand
would be required.
Finally, it should be noted that receptors that would selectively bind one enantiomer of certain special classes of chiral alkenes are easily visualized. For example, any receptor that can
discriminate between the enantiofaces of trans-2-butene should
be able to discriminate between the enantiomers of trans-cyclooctene. Similarly, any receptor that can discriminate between
the enantiofaces of unsymmetrically substituted cis alkenes
should be able to discriminate between the enantiomers of helically chiral cis a l k e n e ~ . [ ~ ~ ]
24.5. Alternative Constructs
Some of the chiral receptors in Figure 10 have been previously analyzed in alternative but equivalent ways. For example,
513
J. A. Gladysz and B. J. Boone
REVIEWS
Marks et al. have prepared a family of chiral bis(cyc1opentadienyl) lanthanide catalysts for the enantioselective hydrogenation and hydroamination of alkenes.[77]As represented in Figure 14, the larger cyclopentadienyl ligand is pentasubstituted,
and the smaller is disubstituted, with one group chiral. These
likely contain alkene binding sites analogous to that of receptor A. Marks et al. have analyzed the enantioselectivities in
terms of “lateral” and “transverse” shape discrimination, as
defined in Figure 14.
OC’
I
F
H3C
1SPS]
one diastereomer
(assayed after
chromatography)
CH3
>80:20 (Keg)
b, R = CHzCH3 75:25
(crystallizedsample)
OC’
RL
Figure 14. An alternative analysis of receptors [77]
8, R
*
CH3
211821
(one diastereomer
much less stable
kinetically)
20+ B F T [ ~ I I
63~37
63:37
(crystallizedsamples)
= CH3
b,R = CHzCHs
25. Chiral Recognition in Other Alkene, Aldehyde,
and Ketone Complexes
Mo
ON’
A variety of other complexes of chiral transition metal Lewis
acids and prochiral y2 n ligands have been $repared. However,
there is no practical, systematic way to search the literature for
all examples. Furthermore, in many cases binding equilibria
have only been partially characterized. Often, diastereomer ratios are reported after workup and without configurational assignments. All data that the authors have been able to locate
over several years are summarized in Schemes 17A-F. Ratios
that unambiguously constitute equilibrium values are designated Ke,.
25.1. Alkene Complexes of Other d6 [ ($-C,R,)M(L)(L’)]
Lewis Acids
These are the most closely related to alkene complexes of 1,
and will be considered in the order grouped in Scheme 17A.
First, the propene complex of the neutral rhenium Lewis acid
[ ($-C,Me,)Re(CO)(PMe,)l
(18) has been isolated in diastereomerically pure form in 36 % yield following chromatograthis does not in itself establish a high thermop h ~ . ~Of
’ ~course,
’
dynamic binding selectivity for one propene enantioface. However, this Lewis acid is isoelectronic and isosteric with the
nitrosyl-substituted cation I-Me, (Figure 1I), and a very high
binding selectivity is most probable. Although the relative configurations of the rhenium and carbon stereocenters in 18 werc
not determined, they are very likely analogous to those of
(RS,SR)-5-Me: X - (n = 0, 5; Schemes 7 and 16).
The cationic iron propene complex 19a’ X - can be viewed as
isoelectronic with the rhenium propene complex 5a’ BF,. The
frontier orbitals and steric properties of the iron fragment
should be similar to those of I. Furthermore, metal-ligand
bonds in this first-row metal complex will be 6-9% shorter,
which should enhance chiral recognition. The configurational
diastereomcrs of 19a+ TfO- interconvert below room temperature, and the equilibrium ratio can be bound as > 80: 20.[791The
574
I
k‘SnPh3
F
BFi
I
__.________
x-~9.w
8, R
Q
disfavored
lateral
A+
I \PPh3
I \CO
e
R
221831
8 , R = CH3
b, R = C6H5
(NMR data for two
isomers; ratios not given)
23lW]
R = IPC6H13 57:4
(crystallized sample;
cisltrans 2-butene
complexes also prepared)
-
KBq
8 , R = CH3
89:11
b, R CH(CH3)z
23:77
C, R = C6H5
>95:5
d, R = CHzOCH3
83:17
e, R = C(=O)CH2CH3 4357
1, R = C(=O)OCH3
55:45
(also data for some
prophos adducts)
Scheme 17A. Alkene complexes of other d6 [(q6-C,H,)M(L)(L)] chiral transition
metal Lewis acids with prochiral alkenes and aldehydes. Kc, = equilibrium constant.
analogous 1-butene complex 19b’ BF, has been isolated as a
75: 25 mixture of diastereomers.[801Some data are also available
for the related iron complexes 20+ BF, and 21.r81, A closer
investigation of these compounds may establish high thermodynamic binding selectivities. However, in no case have the relative
configurations of the iron and carbon stereocenters yet been
established.
The neutral chromium and molybdenum alkene complexes
22[831and 23[s41have been synthesized. Since nitrosyl ligands
are stronger 7c acceptors than carbonyl ligands, both metal fragHowments will have d orbital HOMOSsimilar to that of I.[28a1
ever, due to the identical sizes of the nitrosyl and carbonyl
ligands (and the abscnce of a bulky phosphane ligand), the steric
properties of the four quadrants should not be as differentiated.
Based upon the limited data available, monosubstituted alkenes
appear to give low binding selectivities. Nonetheless, highly
diastereoselective syntheses involving nucleophilic additions to
cationic 7c ally1 precursors have been reported.[84b1
Anprw. Chenz. I n f . Ed. Engl. 1997, 36, 550-583
REVIEWS
Chiral Recognition in K Complexes
A series of cationic chiraphos ruthenium alkene complexes,
24+ PF;, and prophos analogs with one less methyl group, have
been reported.[s51In contrast to the preceding examples, complexes 24+ PF, have ligand-based chirality. The diastereomers
interconvert rapidly in solution, and relative configurations
have been assigned by NMR. Styrene exhibits a slightly higher
enantioface binding selectivity than propene (>95:5 vs 89: 1I),
in contrast to the situation with I. Curiously, isopropylethylene
gives an opposite binding selectivity (23 :37). This might reflect
some nonideal behavior of the ruthenium receptor such as
changes in the chiraphos ligand or Ru-(C=C) conformations.
With regard to the latter, the energies of the two highest occupied d orbitals are likely more closely spaced than with the other
Lewis acids in Scheme 17A. Alternatively, the phenyl groups of
the chiraphos ligand may create some type of pocket or "steric
wall" that for one diastereomer readily admits the substituents
on the C = C units of propene and styrene, but not that of isopropylethylene. Crystallographic data that would help address
these possibilities are not yet available.
25.2. Alkene Complexes of Other Non-Platinum Chiral
Lewis Acids
These are grouped in Scheme 17 B. The only adducts of early
transition metal chiral Lewis acids that appear to have been
isolated to date are the four-coordinate niobium and tantalum
propene complexes 25 and 26f86,871
and the zirconium(I1) complexes 47 (Scheme 17 F). ElectronicalIy, these differ considerably from alkene complexes of I or I-Me, (e.g. d Zvs d6). However, the types of spectator ligands 25 and 26 can be viewed as
similar; the nitrosyl ligand is replaced by a larger NAr moiety.
In solution, all four possible configurational diastereomers and
M -(C=C) rotamers can be detected. Binding selectivities are
modest. With 25, the diastereomer ratio can be bound as
60:40-80:20 (74:26 and 58:42 rotamer ratios).[861Nonetheless. only one isomer crystallizes. The relative configurations of
the metal and carbon stereocenters and the M-(C=C) conformation are identical to those in VII (Scheme 6), the most stable
isomer for monosubstituted alkene complexes of I (angle of
niobium-carbon-carbon plane with Nb-P bond, 7.1").
Several complexes of five-coordinate, cationic molybdenum
with monosubstituted alkenes, 27+ PF,, and analogs with the
"pom-pom" ligand (MeO),PCH,CH,P(OMe), have been prepared.[881These have metal-centered chirality, and the configurational diastereomers equilibrate in solution at room temperature. In all cases, binding selectivities are very high (>99:1).
Several crystal structures have been determined, and the relative
configurations and Mo-(C=C) conformations depicted in
XXXVII (more stable) and XXXVIII (less stable) established.
With these fragments, it would seem particularly worthwhile to
map the relative steric properties of the quadrants through additional binding studies. The available data indicate that the two
least congested quadrants have a geminal relationship (both syn
to the cyclopentadienyl ligand), and the receptor is likely of type
B (Figure 10).
Recently, the cationic arene ruthenium styrene complexes
28+ SbF, were isolated.r89]Although the metal fragment lacks
a cyclopentadienyl ligand, there are some similarities with the
Angeu
Chenr. Inr Ed Engl 1997, 36, 550-583
mdu+
Ph
HTP:Fi
28* S b F i
a, X
b, x
XXXIX
XL
(RRSS)
(RSSR)
--
K,
Ph
89:Il
OMe
72:28
8.8%
29.6%
47.6
"C
KBS
BFi
Me
93:07
91:09
89:lI
Scheme 17 B Alkene complexes of other chiral non-platinum Lewis acids.
Lewis acids in Section 25.1. In solution, the configurational
diastereomers rapidly equilibrate below room temperature. Depending upon the phosphorus donor ligand, styrene enantioface
binding selectivities can be as high as 89: 11. A crystal structure
and extensive NMR analyses establish the configurations and
Ru-(C=C) conformations shown in XXXIX (more stable) and
XL (less stable). Only very small quantities of the alternative
Ru-(C=C) rotamers are present. Hence, the two least congested quadrants of this receptor have a cis relationship.
Square-planar adducts of chiral rhodium Lewis acids and
chelating trisubstituted alkenes, such as 29' BF; ,I9'' are intermediates in a large family of highly enantioselective catalytic
hydrogenation reactions. These have ligand-based chirality.
Binding selectivities are high and depend slightly upon temperature. Selectivities are even higher with complexes having other
chelating diphosphanes. However, as emphasized in other sections of this review, the least stable isomers are more reactive.[901
575
REVIEWS
J. A. Gladysz and B. J. Boone
25.3. Alkene Complexes of
Chiral Platinum Lewis Acids
Equilibrium data are available for
alkene complexes of several types of chiral d* platinum(I1) Lewis acids, as summarized in Scheme 17 C.r'o.91 - 9 9 1 All of
these have ligand-based chirality, and in
most cases configurations have been assigned. Note that the configurations of
any stereogenic nitrogen donor atoms,
which would rapidly invert in the free
ligands, become fixed upon coordination. In the absence of special features
(discussed below), four-coordinate
square-planar adducts uniformly give
poor alkene enantioface binding selecK,,, 53:47
K,,, 5347
tivities. However, with five-coordinate
trigonal-bipyramidal adducts, binding
selectivities can be much higher.
In the square-planar complexes 30,['01
the alkene is trans to a chiral primary
amine. The stereocenter is one atom removed from platinum, and alkene enantioface binding selectivities are very low.
With 31,[921the alkene is cis to a sulfoxide ligand stereocenter, and binding
selectivities are slightly higher. Trends
seem to parallel those noted with
24' PF, : styrene gives a higher binding
selectivity than propene (75:25 vs
56:44), and isopropylethylene exhibits
I
an opposite binding selectivity (34: 66).
With the chelating N-methylproline
Ph
CN
derivatives 34,[941the alkene is trans to a
37tsel
tertiary nitrogen stereocenter. Binding
M=Pd,Pl
selectivities remain low. Adducts of
Kq 55:& (rrms)
styrene, p-chlorostyrene, and p-methoxystyrene have been characterized. Curiously, electron-withdrawing and electron-donating para substituents both
slightly lower binding selectivities
(59:41 and 62:38 vs 65:35).
In contrast to 34, the chelating nitrogen in the alanine derivatives 32 and
33Cg31 is not a stereocenter. Styrene
enantioface binding selectivities become
3Bw
K,,,
40ml
Keq
41
K,
even lower (53 :47). With the sarcosine
a, WR'= trans CHdCH3 6535
&(:I6
a, WR'= WCH3
8020
0, WR'= WCH3
derivatives 35 and 36,[97'nitrogen sterb, WR'= WC(=O)H
>95:<5
b, WR'= H/cH&H3
91:09
b, WR'= bansCYCl
88:12
C, WR'= WCN
w:33
c, WR'= WC&
5545
eocenters-now
secondary-are
red. WR'= trans CWCN >95:<5
d, WR'= bans CHdCH3 92:08
stored. The alkene and nitrogen are cis
Scheme 17 C. Alkene complexes of chiral platinum Lewis acids.
in the latter, and allylic alcohols now
give high binding selectivities ( > 98: 2stereocenter, 35 and 36 cannot be directly obtained in non86: 14). When the allylic alcohol and nitrogen are trans as in 35,
racemic form. However, proline analogs of 36 give only slightly
binding selectivities plummet (52: 48- 50: 50). This suggests
diminished binding selectivities.
some type of special attractive interaction involving one quadThe trigonal-bipyramidal complexes 4Olg9]feature a C,-symrant of the metal fragment in 36 (see H in Figure 13). Indeed, the
metric receptor with a chiral secondary diamine. The configuracrystal structure of 36b shows a close contact of the NH proton
tions of the nitrogen stereocenters are controlled by carbon
and the alcohol oxygen.['OO1Since sarcosine lacks a carbon
&>
576
Angew. Chem. Int. Ed. Engl. 1997, 36, 550-583
Chiral Recognition in 7c Complexes
REVIEWS
stereocenters in the chelate backbone. Propene, I-butene, and trans2-butene give high enantioface binding selectivities. As would be expect1. acetone -45 “C
ed from the quadrant diagrams for
+
A
2. crystallization
C,-symmetric receptors (G, G’ in
Figure 13), the equilibrium constant
for trans-2-butene (which must bind
in a matched or mismatched manner
with respect to hoth CHCH, termini)
1.O equiv
1.0 equiv
4.0 equiv
95% yield
is greater than that of propene (92:s
973 SSFUSSS
vs 84: 16). Curiously, styrene gives a
neariy as selective:
much lower selectivity (55 :45). ReH,C-CHCH( R)OH,
(€)-CH&H-CHCH(R)OH
lated complexes in which a methyl
much poorer:
ligand has replaced one axial chloH&=CHCH(CH,)CH&H,
rine (41) have been
Scheme 17 D. Reactions of a chir.a1 copper Lewis acid and racemic chiral alkenes [ 1031.
The binding selectivity for trans-2butene is slightly lower (85:15).
Complexes 39 feature a C,-symmetric receptor with two tercertainly reflects a high thermodynamic binding selectivity
tiary amino groups.[951Chirality is now derived from carbon
rather than any type of kinetic effect. Structural data are not yet
stereocenters on the N-alkyl groups and not stereocenters in the
available, but with the SSdiamine the R alcohol is preferentially
chelate backbone. Propene shows a slightly lower binding selecbound. Chiral alkenes without allylic OH-substituted stereotivity than with 40 (80:20 vs 84:16). Fumaronitrile, with
centers give much lower binding selectivities, suggesting that a
two trans CN groups, gives a higher diastereomer ratio than
special attractive interaction involving the hydroxyl group plays
acrylonitrile (>95:5 vs 67:33). In all of these C,-symmetric
a key r ~ l e . [ ’ ~ ~ I
systems, the more stable diastereomers are believed to have the
A further conceptual step would entail the selective reaction
CHR substituents trans to the bulkier nitrogen substituents (cis
or “kinetic resolution” of one enantiomer of a chiral alkene. Of
to NH or NCH,). Complex 38[961can be regarded as an sp2
course, such processes do not have to involve a metal alkene
nitrogen donor analog of 39, but with only one chiral N-alkyl
complex. However, only this category is considered here. A
group. While it is not surprising that the binding selectivity is
relatively early example would be the (binap)ruthenium(It)-catalow, it would be interesting to study a C,-symmetric variant.
lyzed hydrogenation of chiral secondary allylic alcohols.[1041
The three-coordinate complex 37[’*] also contains a chelate ligEnantiomers typically exhibit 11- to 75-fold rate differences.
and with two sp2 nitrogen donors, but the carbon stereocenters
However, this versatile catalyst system remains difficult to interare further removed from the binding site. Accordingly,
pret. More recently, a chiral zirconium catalyst has been found
fumaronitrile gives a much lower enantioface binding selectivity
to effect highly selective carbomagnesations of the C=C moi(55:45) than in 39.
eties of chiral2-alkyl d i h y d r ~ p y r a n s . ” The
~ ~ ] enantiomers normally react at much different rates. A related chiral titanium
catalyst effects highly selective hydrogenations of the N=C moi25.4. Complexes of Chiral Alkenes
eties of one enantiomer of substituted dihydropyrroles.[106]In
both systems, there is excellent evidence and precedent for the
Although chiral alkenes are outside the focus of this review,
structures of the intermediate Lewis acids and X=C adducts.
several points deserve attention in passing. First, chiral platThese lead to convincing rationales for the observed selectiviinum Lewis acids have played a historically important role in the
ties. Related complexes and transformations are highlighted in
resolution of racemic chiral alkenes, such as trans-cycloSection 26 on reactivity.
octene.” O ’ , ‘ O 2 ] However, they have been used to complex both
enantiomers. The diastereomers were then physically separated
and the enantiomerically pure alkenes then detached. Hence,
25.5. Aldehyde and Ketone Complexes of Other Chiral
Transition Metal Lewis Acids
thermodynamic binding selectivities were not factors. Approaches to designing receptors that can selectively bind one
enantiomer of a chiral alkene are discussed in Section 24.3.
To our knowledge, binding selectivities have been reported
Towards this end, a chiral d ” copper(1) Lewis acid has refor only one other K adduct of an aldehyde and a chiral transicently been reported that can efficiently discriminate betion metal Lewis acid, the rhenium benzaldehyde complex
tween the enantiomers of chiral secondary allylic alcohols.[’031
43’ TfO- (Scheme 17 E)
Careful NMR measurements inAs illustrated in Scheme 17D, equimolar quantities of
dicate a > 97: 3 mixture of configurational diastereomers. The
[Cu(NCMe),]’ BF; and the same C,-symmetric chelating direlative configurations and Re-(O=C) conformation depicted
amine utilized in 40 (Scheme 17C) are combined. Then an excess
in Scheme 17 E are likely. This receptor, which has relatively few
of racemic 3-buten-2-01 is added. Crystallization gives the
degrees of freedom, is well suited for binding studies, steric
analyses of quadrants, and structural modifications and/or optialkene complex 42’ BF, in 95% yield and 97% diastereomeric
mization, in analogy to the studies on I. It seems probable that
purity. Although additional tests would be desirable, this almost
Angew. Chem. In1 Ed. Engl 1997. 36, 550-583
577
REVIEWS
J. A. Gladysz and B. J. Boone
~
43' TfO-"oq
Keq,9713
Scheme 17 E. Benzaldeiiyde
complex of a chiral rhenium
Lewis acid.
the least and most congested
quadrants have a trans relationship.
We are not presently aware of
any 7t complexes of prochiral ketones and chiral transition metal
Lewis acids.[681However, a chiral
titanium catalyst depicted in
Scheme 19 probably gives high
binding selectivities with aryl ketones.
25.6. Addendum
Additional complexes of chiral transition metal Lewis acids
and prochiral alkenes that have come to our attention since this
review was accepted for publication are summarized in
Scheme 17 F. These include three examples of d" [($-C,R,)M(L)](L' Lewis acids (44' BAr,, 45' BAr,, 46fX-)."27"2*1
The zirconium complexes 47 are especially noteworthy.[' 291
These can be crystallized at low temperature, and N M R spectra
show only one isomer. Given the variety of highly enantioselective transformations catalyzed by closedly related Lewis acids
(cf. Scheme 19), high binding selectivities with many types of
alkenes can be anticipiated.
The C,-symmetric copper Lewis acid in 52' PF; gives very
high binding selectivities with a variety of aromatic alkenes." 341
The crystal structure of the styrene complex 52a' PF; shows
that the phenyl substituent occupies a cleft where it is roughly
coplanar to one mesityl group and perpendicular with the other.
This is consistent with attractive interactions discussed in Sections 8, 11,15, and 20, and such tandem effects have great
promise as general design elements in chiral recognition. This
interpretation is further supported by the much lower binding
selectivity of vinylcyclohexane (55g' PF;) and aromatic alkenes with ortlzo substituents that disrupt the edge-face interaction (55d,e+PF;). The binding selectivities are too high [or
electronic effects analogous to those in Scheme 14 to be measured.
P
EtO
44' BAr4-[lz7I
X = OCH3, CH3
(only rotarners detected)
46+ x-IW
45+ BAr4-[1271
K,,
(no isomers reported)
80:ZO
(rotamers also detected)
N C C C N
R'-
471'291
R = CH3. CH2CH3
R' = H. CH3
Keo likely high
48IW
R = CH3, Ph
(NMR data for two
isomers; ratios not given)
Me
4g[1311
50" 321
(no isomers reported)
(no isomers reported)
H*
pFs
KJI1331
K,, 6238
(kinetic ratio 9O:iO)
26. Binding, Reactivity, and Product Selectivity
Scheme 1 7 F Addendum to Schemes 17A-17E.
Does the extensive body of binding data analyzed above lead to
any insight regarding reactivity or optimization of product selectivity? Indeed, for many Rh'-catalyzed enantioselective hydrogenations, the less stable configurational diastereomer of the
intermediatc alkene complex (see 29' BF,, Scheme 17B) i s
known to be more reactive.[", '08] However, transition state
properties arc commonly correlated to and interpreted in the
context of reactant and product properties. Thus, for a given
transformation, the enantioselectivities obtained with various
types of alkenes could be compared to the enantioface binding
selectivities of different chiral metal receptors (I, I-Me,, I-Me,,
etc.). A correlation could be used as evidence for the steric
environment of the transition state.
Consider first the more easily visualized realm of stoichiometric transformations and an attack of an external reagent
578
upon the 7t ligand that fixes a product stereocenter. When the
equilibration of configurational diastereomers is slow relative to
the rate of the attack, diastereomer ratios will directly translate
into product enantiomer
Enantioselectivities will be
high 1) when kinetic binding selectivities are high, o r 2) when
thermodynamic binding selectivities are high and equilibration
can be effected prior to reagent addition. Interestingly, the potential importance of high kinetic binding selectivities in enantioselective transformations does not appear to be widely
appreciated. In this case, the activating Lewis acid, ligand, and
attacking agent can be combined in a simple sequential operation, whereas with case 2 an intervening step is required (for
example Scheme 7 for alkene complexes of I).
A n ~ e i i ,Chm~.In/. Ed. Engl 1997, 36.
5 5 0 - 583
Chiral Recognition in x Complexes
REVIEWS
When the equilibration of configurational diastereomers is
rapid relative to the rate of attack, the Curtin-Hammett limit
will apply,[31and product enantioselectivities will be determined
by the differences in energies of the competing transition states.
In some cases, there will be a correlation with binding selectivities, and in other cases (such as the Rh'-catalyzed hydrogenations) none.
In these contexts, the additions of nucleophiles to complexes
of I with monosubstituted alkenes and aldehydes are instructive.
As shown in Scheme 18, the former are attacked at the substituted carbons and on the C = C faces opposite the rhenium to give
neutral alkyl complexes with C, stereocenters (XLI, XLII) .I1
lo]
The enantiomer and diastereomer ratios of the products are
identical with those of the reactants. Thus, additions are enantiospecific, diastereospecific, and regiospecific.['] In contrast,
similar additions to aldehyde complexes give alkoxide complexes (XLV, XLVI) in which the diastereomer ratios do not match
those of the reactants.[63a.7 3 . ' ' 'I R ate studies and other
' 9
Ar
P h
ON
'"-1
&
Ppm
ON
ddlbn lo C=O
taceantitopm,
1
CN-
Scheme 18. Relationship between binding and reactivity for nucleophilic additions
to alkene and aldehyde complexes of I [110-1121.
Angeu. Clwnr. Int. Ed. Engl. 1997,36,550-583
data show that the aldehyde ligands first undergo x to 0 isomerization (XLIII, XLIV) .[' "] Furthermore, two regimes can be
identified: one at higher nucleophile concentrations in which
isomerization is rate determining ("saturation kinetics"), and
another at lower nucleophile concentrations in which addition is
rate determining. Interestingly, in the former limit it still remains possible for the diastereomer ratio of the x complexes to
correlate (by a "torquoselective"[' ' 31 isomerization to an E or Z
(r complex) with the diastereomer ratio of the alkoxide complexes.
Consider next the rapidly growing body of enantioselective reactions catalyzedbychiralmetall~cenes.~~~
1 3 a , 7 7 - 1 0 5 . L o 6 * 1 1 4 -'"]
One example is the asymmetric hydrogenation of alkenes by the
lanthanide catalysts shown in Figure 14.[771Another is the
asymmetric hydrosilylation of aryl ketones by the titanium catalyst shown in Scheme 19.11141
These reactions commonly involve a 7~ complex of the substrate. A subsequent migration of
a hydride or alkyl moiety from the catalyst then fixes a product
stereocenter. Suppose that for the more stable configurational
diastereomer of the x complex, the receptor favors the M(X=C) rotamer with the reactive terminus syn to the migrating
group. In this case, it would be difficult to imagine a situation
in which the most stable isomer were not the most reactive. In
Scheme 19 the alternative Ti-(O=C) rotamers would certainly
be less stable, and therefore correlations of enantioface binding
selectivities, reactivities, and product configurations would be
expected." '*I
Binding data for complexes of chiral bent metallocene Lewis
acids and prochiral alkenes, aldehydes, and ketones are currently extremely scarce." 19,
Equilibrium measurements would
be of direct use in optimizing this growing area of catalytic
chemistry, which now also includes enantioselective imine hyd r ~ g e n a t i o n s , " ~ ~31. alkene carbornagne~ations[~~",and
carboaluminations,[' 61 and diene cyclopolymerizations.15a~
As an illustration of an attractive system for study, niobocene
alkene hydride complexes of the formula [($-C,R,),Nb(H,C=CHR)(H)] have been carefully characterized.['2o1 It
should be possible to prepare similar alkene complexes of chiral
niobocene Lewis acids that would be close structural models for
the key titanium and zirconium intermediates in the preceding
transformations. Regardless, we suggest at this time that the
titanium Lewis acid in Scheme 19 probably belongs to receptor
type F (Figure 10, mirror image), with a cis relationship between
a prohibitively congested quadrant and the next most congested
quadrant.
Of course, 7~ complexes of chiral transition metal Lewis acids
and alkenes, aldehydes, and ketones play key roles in a variety
of other catalytic enantioselective reactions. However, from
available data it is only possible to speculate on transformations
in which binding selectivities will correlate with product configurations. One recently reported possibility would be the palladium-catalyzed alkylation of alkenes containing two allylic geminal alkoxycarbonyl groups."
The key intermediates might be
modeled by alkene complexes of chiral bis(phosphane) platinum(o) Lewis acids. Chiral palladium(I1) catalysts that combine
monosubstituted and disubstituted alkenes and CO to give alternating copolymers of high enantiomeric purities have also
been described." 3b, 'I' These mechanistically well-defined systems are particularly suited for additional study.
579
J. A. Gladysz and B. J. Boone
REVIEWS
side view
0
top view
side view
-
r H
l*
H
f
I
O-
> -" I
minor
II
RP
'/h
H
1
H
l*
\
active Lewis
acid catalyst
1
Scheme 19. Probable relationship between binding selectivity and enantioselect~vltyobtained with titanium Lewis acid that catalyzes ketone hydrosilylation [114]. The angle
between the cyclopentadienyl planes is exaggerated for clarity
Other obvious areas for inquiry would include metal-catalyzed enantioselective cyclopropanations[' 231 and epoxidati on^["^.
of alkenes. However, in many of these reactions
metal alkene complexes are not intermediates. In cases in which
the first new bond formed fixes all product stereocenters, correlations with the binding selectivities of a specific chiral receptor
might nonetheless be expected. This could, as noted above, be
used as a probe of the steric properties of the reactive site.
However, in some cases product stereocenters are fixed in subsequent steps." 241 Here, no meaningful correlations would be expected.
27. Conclusion
All stereoisomers that can result when achiral alkenes, aldehydes, and ketones bind in an '1 7~ manner to chiral transition
metal Lewis acids have been identified (Scheme 2 ) . Further
analyses show that there are only six types of chiral metal receptors, as classified by the relative steric properties of four quadrants. The corresponding quadrant diagrams in Figure 10 allow
many types of binding selectivities to be easily predicted or
rationalized. They also provide starting points for more refined
bonding models of broader utility or improved quantitative capability (Figure 13).
In particular, these constructs enable ready identification of
receptors that will be superior for binding one enantioface of a
prochiral alkene, aldehyde, or ketone. Alternatively, rational
ways to modify receptors to achieve improved binding selectivities become apparent. Receptors for which the steric properties
are unknown may be "mapped" by measurements of various
binding equilibria, as exemplified for the chiral rhenium Lewis
acid I. Although I has not been structurally optimized, a literature survey (Figure 17) shows that it presently offers the highest
binding selectivities for the broadest spectrum of alkenes. Other
580
metal fragments certainly hold promise, and even higher selectivities should eventually be realized.
Electronic effects can also be incorporated into these bonding
models. In appropriate situations, they can be pronounced.
In the case of complexes of I with aromatic aldehydes, a
clear-cut underlying structural basis has been established.
Finally, many of these concepts can be extended to transition
states. In some cases a correlation between binding selectivities
and product configurations can be expected, but in other
cases not.
This review points the way to less empirical and more quantitative approaches to chiral recognition, which is required in
some context in all enantioselective syntheses. An important
future direction will be to extend this analysis to o complexes,
which are responsible for enantioselection in many types of
chiral Lewis acid promoted transformations. Here, E /Z and
rotameric equilibria about the o bond will play much the same
roles as the configurational diastereomers and M-(X=C) rotamers discussed above. Complementary types of steric and
electronic" 261 effects will be operative.
We thank the U . S . National Institutes of Health for support of
this research, Dr. T-S Peng for assistance with an early version of
the manuscript, and the Alexander von Humboldt Foundationfor
a Senior Scientist Research Award that enabled the completion of
this review at the Fachbereich Chemie der Philipps-Universitat in
Marburg, Germany.
Received: February 5, 1996
Revised: May I , 1996 (A1441EI
German version: Angeu'. Chem. 1997. 109. 566-602
[l] See, in addition to references cited below: R. Noyori, Asymmetric Cutalysis in
Orgunic Synthesis, Wiley, New York, 1994.
[2] Reviews of chiral Lewis acids: a) K. Narasaka, Synthesfs 1991, 1, b) K.
Maruoka, H. Yamamoto in Catulytic Asymmetric Synthesis (Ed.: I. Ojima),
VCH, New York, 1993, Chapter 9.
[ 3 ] J. I. Seeman. Chem. Rev. 1983. 83, 83.
Angew. Chem. Int. Ed. Engl. 1997,36, 550-583
Chiral Recognition in
K
Complexes
[4] An example of chiral recognition involving prochiral species is the coupling of
planar carbon radicals R R R C ' to give an unequal mixture of dl (racemic)
and meso diastereomers.
[5] lsotactic and syndotactic polypropylene are achiral, but oligomers and some
closely related polymers are chiral. Lead references: a) H.-H. Brintzinger. D.
Fischer, R. Mulhaupt, B. Rieger, R. M. Waymouth, Angew. Chem. 1995,107,
1255; Anguir. Chem. I n / Ed. Engl. 1995.34, 1143; b) G. Guerra, L. Cavallo.
G. Moscardi, M. Vacatello, P. Corradini, J. Am. Chem. Soc. 1994, 116,2988;
c) G. W. Coates. R. M. Waymouth, ihid. 1993, 115, 91; d) M. A. Giardello,
M. S Eisen. C 1.Stern. T. J. Marks, ibid. 1995, 117, 12114.
[6] H. Brunner. Adv Orgunornet. Chem. 1980, 18, 151.
[7] There is extensive literature on the quanitification and analysis of Lewis
acidlbase binding strengths. Recent lead papers: a) I. D. Brown, A. Skowron,
J. Am. Chem. Soc. 1990. 112, 3401. b) P. Laszlo, M. Teston, ihid. 1990, 112.
8750.
[S] For current definitions, see E. L. Eliel, S. H. Wilen, StereochernisfryojOrganIC Compounds. Wiley. New York, 1994, Glossary, pp. 1191- 1210. For additional analyses of C=C, C=O, and C=N faces in unsaturated organic
compounds, see S. A. Kaloustain, M. K. Kaloustain, J . Chem. Erluc. 1975,
52. 56.
[9] In principle, such diastereomers could also interconvert by exchange of the
=CHR face bound to the metal, leaving the =CH, moiety unchanged. The
introduction of a C I S or trans deuterium label can differentiate these pathways.
[lo] G. Paiaro, A. Panunzi, J . Am Chem. SOC.1964, 86, 5148.
[ l l ] For R:S conventions at metal stereocenters, see a) C. Lecomte, Y Dusausoy.
J. Protas. J. Tirouflet, A. Dormand, J. Organomet. Chem. 1974, 73.67; b) K.
Stanley. M. C. Baird, J . Am. Chern. Soc. 1975, 97, 6598; c) T. E. Sloan, Top.
Sterm(.hun~.1981. 12. 1
[12] In synclinal (s(.j M-(C=C) rotamers. the substituents with the highest R/S
priorities on the metal ($-C,H, for compounds in this review) 1111 and the
C=C centroid (=CHR > =CH,) define (60+30)' torsion angles. In anticlinal (acj. synperiplanar (sp),and antiperiplanar (up)M-(C=C) rotamers,
these substituents define (120+30)', (0*30)", and (180+30)" torsion angles,
respectively. See Table 2.2 in ref. [XI.
[13] a) For examples involving cis cycloalkenes and a chiral zirconium Lewis acid,
see M. T. Didiuk. C. W. Johannes, J. P. Morken, A. M. Hoveyda, J. Am.
Chem. SOL 1995. f I 7 , 7097; b) For an example involving cis-2-butene and a
chirdi palladium Lewis acid, see 2. Jiang, A. Sen, ibid. 1995, 117, 4455.
[14] a) M Herberhold, Metal n-Complexes. Vol. ZI, Elsevier, New York, 1974.
Part 2; b) D M P. Mingos in ComprehensiveOrganometallicChemistry, Vol.
3. (Eds: G. Wilkinson, F. G. A. Stone. E. W. Abel), Pergamon, Oxford, 1982,
pp. 41 - 15.
[lS] Quadrant diagrams frequently appeared in the early literature on enantioselective hydrogenations and hydroformylations of alkenes with chiral Rh' catalysts [lbd,b]. The authors have sought to identify the earliest use of a quadrant diagram for the analysis of any reaction leading from a prochiral alkene
to a chiral product. Since such a search cannot be conducted systematically,
examples from readers are welcome. More recently, some highly stylized version of quadrant diagrams have evolved [16c].
[16] a) W. S Knowles. Acr. Chem. Res. 1983, f6, 106; b) G. Consiglio, P. Pino,
Top. Curr. (%em 1982. 105.77; c) K. B. Sharpless, W. Amberg, M. Beller, H.
Chen. J. Hartung, Y. Kawanami, D. Liibben, E. Manoury, Y Ogino. T. Shibata, T. Ukita. J . Org. Chem. 1991, 56. 4585; d) H. C. Kolb, M. S. Van
Nieuwenhze. K B. Sharpless, Chem. Rev. 1994, 94, 2483.
[17] a ) D. White, N. J. Coville. Adv. Organomet. Chem. 1994,36, 95; b) for calculations of steric sizes of alkene ligands by a molecular mechanics model, see
D. P. White. T. L. Brown, Inorg. Chem. 1995,34, 2718.
[18] Ref. [8], pp. 690-700.
[19[ a ) 0 Eisenstein. R. Hoffmann, J Am. Chem. Soc. 1981,103,4308;b) T. C. T.
Chang. B M. Foxman, M. Rosenblum, C. Stockman, ibid. 1981, 103, 7361;
c) A. D. Cameron. V. H. Smith, Jr., M. C. Baird, J . Chem. Soc. Dalton Trans.
1988. 1037.
1201 Specifically.the intersection of the perpendicular from the metal to the C =C
bond is first identified (point X). The displacement of point X from the
midpoint of the C=C bond is then divided by half the C=C bond length and
expressed as a percentage.
1211 S. D. Ittel. J. A. Ibers, Adv. Organornet. Chem. 1976, 14, 33.
(221 Molecular orbital analyses predict that diamagnetic d6. sixteen-valence-electron (($-C,H,)M(L)(L')] species should have pyramidal, not planar, ground
states: P. Hofmann, Angew. Chem. 1977,89,551;Angew. Chem. h t . Ed. Engl.
1977. 16, 536; see also P. Hamon, L. Toupet. J.-R. Hamon, C. Lapinte.
Organomr/rrllics 1996, IS, 10.
1231 To our knowledge. the first adducts of I were reported in 1979: W. Tam, W. K.
Wong. J. A. Gladysz. J . Am. Chem. Soc. 1979, 101, 1589.
[24] Third-row transition metal sixteen-valence-electron species appear to be
much less stable relative to their eighteen-valence-electron counterparts than
first- or second-row sixteen-valence-electron species: C P. Casey, T. L.
Underiner. P. C. Vosejpka, J. A. Gavney, Jr., P. Kiprof. J. Am. Chem. Soc.
1992, 114. 10826, and references therein.
[251 The vector that connects rhenium and the cyclopentadienyl centroids in 1-111
actually extends behind the plane of the paper. In contrast, the rheniumAngen Chem Inr Ed Engl 1997, 36, 550-583
REVIEWS
nitrogen and rhenium-phosphorus bonds are in the plane of the paper.
Although this somewhat diminishes the steric effect of the cyclopentadienyl
hgand, crystal structures of Lewis base adducts of the types 11 and I11 always
show one or two X-Re-(Cp(carbon)) angles of less than 90 Thus, the cyclopentadienyl ligand extends into the plane of the paper.
[26] G. L. Crocco, K. E Lee, J. A. Gladysz, Orgonomernllics 1990. 9, 2819.
[27] For other relevant data, see a) S. G Davies, I. M. Dordor-Hedgecock. K H.
Sutton, M. Whittaker, J . Am. Chem. Soc. 1987, 109. 571 1 ; b) S. C. Mackie.
M. C. Baird, Organometallics 1992, 11, 3712.
[28] a) B. E. R. Schilling, R. Hoffmann, J. W. Faller. J Am. Chem. SOC.1979, I U I ,
592; b) W. A. Kiel, G.-Y. Lin, A. G. Constable. F.B. McCormick, C. E.
Strouse, 0. Eisenstein, J A. Gladysz, J . Am. Chem. SOC.1982, fU4, 4865;
c) P. T. Czech, J. A. Gladysz, R. F. Fenske, 0rganometuNic.s 1989, 8, 1810.
1291 a) D. L. Lichtenberger, A. R. Rai-Chaudhuri, M. J. Seidel. J. A. Gladysz,
S. K. Agbossou, A. Igau, C. H. Winter. 0rganometnllic.s 1991, f U , 1355,
b) D. L. Lichtenberger, N. E. Gruhn,A. Rai-Chaudhuri, S . K. Renshaw, J A.
Gladysz, J. Seyler, A. Igau, unpublished results.
[30] For example, with x alkyne ligands, one substituent on the C r C bond is direct
into a cleft in the bulky PPh, ligand: J. J. Kowalczyk, A. M Arif, J. A
Gladysz, Organometallics 1991, f0, 1079.
[31] a) F. Agbossou, E. J. O'Connor. C. M. Garner, N. Quiros Mendez. J. M
Fernandez. A. T. Patton, J. A. Ramsden, J. A. Gladysz, Inorg Synth. 1992.29,
211; b) Improved procedure for introducing PPh,: Y Zhou. M. A. Dewey.
J. A. Gladysz, Organometallics 1993, f2,3918.
1321 J. M. Fernandez, 1 A. Gladysz, 0rgunon.retallics1989, 8 , 207.
[33] J. J. Kowalczyk. S. K. Agbossou. J. A. Gladysz, J . Orgunomet. Chem. 1990,
397, 333.
[34] M. A. Dewey, Y. Zhou, Y. Liu, J. A. Gladysz, Organometullics1993,12, 3924.
1351 G. S. Bodner. T.-S. Peng, A. M. Arif, J. A. Gladysz, Organometallics1990, 9,
1191.
[36] Y. Wang. J A. Gladysz, Chem. Ber. 1995, 128, 213
[37] T:S. Peng, A. M. Arif, J. A. Gladysz, Helv Chim. Acta 1992, 75, 442.
[38] a) The configuration at rhenium is specified first [I 11 and followed by those
of the =CHR stereocenters [lo]. In complexes with two =CHR stereocenters,
theconfiguration of thecarbon with the higher Cahn-Ingold-Prelog priority
(e.g. =CHCHCH, > = M C H , ) is given first. If the priorities are equal, the
configuration of the =CHR group anti to the PPH, ligand is given first. h)
The configuration of a given carbon is not affected by a 1 8 0 rotation about
the Re-(C=C) axis. However, when the preceding convention is applied to
the complexes in Scheme 8 with symmetrical cis alkenes. there i s an apparent
inversion. c) In many schemes only one enantiomer of a racemate is depicted.
These always correspond to the first enantiomer specified in the diastereomer
designation (for example the RS enantiomer for the RS,SR diastereomer)
[39] a) In many cases replicate determinations have been conducted and standard
deviations are much less. b) Unless noted otherwise, detection limits are 1%.
Hence a ratio of > 99: 1 implies that the second isomer cannot be detected. In
such cases, either an authentic sample enriched in the second isomer is available. or two isomers of a closely related compound can be detected easily
under similar conditions.
[40] C. Roger, T.-S. Peng, J. A. Gladysz, J . Organomet Chem. 1992, 439, 163.
[41] T.-S. Peng, J. A. Gladysz, J . Am. Chem. Soc. 1992, 114, 4174.
[42] No significant counter anion effects have been detected to date for the equilibria in Scheme 7. For example, 5e+ PF, and 5h' PF; give RS,SR/RR.SS
ratios of 92:8 and > 99: 1, respectively [36]. Similar data are available for K
aldehyde complexes of 1.
[43] E. L. Eliel, M. Manoharan, J. Org. Chem 1981. 46, 1959.
1441 This may be viewed equivalently as an effect involving two approximately
perpendicular unsaturated systems: a) H. Brunner, Angew. Chem. 1983, 95,
921; Angen. Chem. Int. Ed. Engl 1983,22. 897, see Sections 6-8; h) review:
M. Nishio, Y Umezawa, M. Hirota, Y. Takeuchi, Tetrahedron 1995,Jf. 8665;
c) I. Dance, M. Scudder, J Chem. SOC.Dalton Trans. 1996, 3755; d j K. D.
Schladetzky, T. S. Haque, S. H. Gellman, J. Org. Chem. 1995.60.4108;e) H.
Brunner, R. Oeschey, B. Nuber. Organometallics 1996. IS. 3616, and references therein
[45] J. Pu. TLS. Peng, C. L. Mayne, A. M. Arif. J A. Gladysz, Organometullics
1993, 12, 2686
[46] J. J. Kowalczyk, A. M. Arif, J. A. Gladysz, Chem. Ber. 1991, 124, 729.
1471 Some of these data were reported in the discussion section of ref. [45]; others
were determined by B. J. Boone and are given here
[48] T:S. Peng, J. Pu, J. A. Gladysz, Organornetallics 1994. f3. 929.
[49] T.4. Peng, Y Wang, A. M. Arif, J. A. Gladysz, Orgunomefullics 1993, IZ.
4535.
[50] Y Wang, F. Agbossou, D. M. Dalton, Y Liu, A . M . .4rif. J. A. Gladysz,
Organomerallics 1993. 12, 2699.
[Sl] a) Since these reactions proceed by associative substitution (Scheme 5) [34],
we speculate that less congested cs (as opposed to n) O=C adducts are generated first. The greater nucleophilicity of the O=C group would then be explained by the smaller size (no substituents on oxygen), the higher energy
oxygen-based nonbonding HOMO, and the greater bond dipole. Nonconjugated analogs such as 5-hexen-2-one behave similarly: E. J. Fairfax. M. Sc
Thesis, University of Utah, 1993 b) For similar results involving the achiral
581
J. A. Gladysz and B. J. Boone
osmium Lewis acis [OS(NH,),]~+.see W D. Harman, W. P. Schaefer, H.
Taube, J Am. Chem Soc. 1990, 112, 2682.
[52] For an independent synthesis of this compound (>98:2 RS,SR/RR,SS), as
well as of allyl alcohol, 3-buten-1-01,and related adducts of I, see S . Legoupy,
C. Crevisy, J.-C Guillemin, R. Gree, Tetrahedron Lett. 1996, 37, 1225.
[53] G. A. Stark, A. M. Arif. J. A. Gladysz, Organomerallics 1994, 13, 4523.
[54] J. Pu, T.-S. Peng, A. M. Arif, J. A. Gladysz, Organometallics 1992, 11,
3232.
[55] M. Sanau, T.-S. Peng, A. M. Arif, J. A. Gladysz, J. Organomet. Chem. 1995,
503, 235.
[56] Lead references: a) J. W. Faller, B. V. Johnson, J . Organomet. Chem. 1975,96,
99; b) E. Bye, B. Schweizer,J. D. Dunitz, J . Am. Chem. SOC.1982,104,5893;
c) H. Brunner, B. Hammer, C Kruger, K. Angemund, I. Bernal,
Organometalhcs 1985,4, 1063; d) S. G. Davies, A. E. Derome, J. P McNally,
J . Am. Chem. Soc. 1991, 113, 2854; e) J. Polowin, S. C. Mackie, M. C. Baird,
Organometallics 1992, 11, 3724; f) S. Garner, A. G. Orpen, J . Chem. SOC.
Dalton Trans. 1993,533, g) H. Brunner, R. Oeschey. B. Nuber, Angew. Chem.
1994, 106, 941; Angew. Chem. Inr. Ed. Engl. 1994, 33, 866
[57] N. Quiros Mendez, J. W. Seyler, A. M. Arif, J. A. Gladysz, J. Am. Chem. SOC.
1993, 115,2323.
[58] N . Quiros Mendez, C. L. Mayne, J. A. Gladysz, Angew. Chem. 1990, 102,
1509; Angew. Chem. Int. Ed. End. 1990,29, 1475.
I591 a) B. J. Boone, D. P. Klein, N. Quiros Mendez. J. W. Seyler, A. M. Arif, J. A.
Gladysz, J. Chem. Soc. Chem. Commun. 1995, 279, b) B. J. Boone, D. P.
Klein, J. W. Seyler, N. Quiros Mkndez, A.M. Arif, J. A. Gladysz, 1 Am.
Chem. SOC.1996.118,2411
[60] a) The crystal structures of three closely related platiumum(rr) complexes with
nitrostyrene, styrene, and p-dimethylaminostyrene have been determined.
The Pt-CHAr bond lengths increase monotonically from 2.216(11) to
2.236110)to 2.262(16) A, while the Pt-CH, bond lengthsvary from 2.174(13)
to 2.180(12) to 2 137117) A[60b]. Similarly, the Pd-CHAr bonds in a series
of p-chlorostyrene, styrene and p-methoxystyrene complexes lengthen
monotonically from 2.235(6) to 2.255(6) to 2.264(5) A, while the Pd-CH,
bond lengths vary from 2.187(6) to 2.182(5) to 2.18215) A[6Oc]. b) S. C.
Nyburg, K. Simpson, W. Wong-Ng, J . Chem. Soc Dalton Trans. 1976, 1865;
c) K. Miki, 0. Shiotani, Y. Kai, N. Kasai, H. Kanatani. H. Kurosawa,
Organomerallics 1983, 2, 585.
[61] S. C. Peacock, L. A. Domeier, F. C. A. Gaeta, R. C. Helgeson, J. M. Timko,
1978, 100,8190.
D. J. Cram, J. Am. Chem. SOC.
1621 a) S. Chang, R. M. Heid, E. N. Jacobsen, Tetrahedron Lett. 1994,35,669, and
references therein; b) T. V. Rajanbabu, A. L. Casalnuovo, J Am. Chem. SOC.
1996,118,6325, and references therein; c) J. M. Hawkins, S. Loren, M. Nambu, ibid. 1994, 116, 1657; d) A. Schnyder, L. Hintermann, A. Togni, Angew.
Chem. 1995, 107,996; Angew. Chem. Int. Ed. Engl 1995,34,931.
[63] a) C. M. Garner, N. Quiros Mendez, J. J. Kowalczyk, J. M. Fernandez. K.
Emerson, R. D. Larsen, J. A. Gladysz, J . Am Chem. Soc. 1990,112, 5146; b)
D. P. Klein, N Quiros Mendez, 3. W. Seyler, A.M. Arif, 1. A. Gladysz, J .
Organomet. Chem. 1993, 450, 157; c) A trifluoroacetdldehyde complex of I
has also been isolated. However, due to its extremely low solubility, the
RS,SR/RR,SSratio (99: 1 in CD,NO,) cannot be measured accurately under
the conditions used for the acetaldehyde complex 16a' BF,: G. A. Stark,
J. A. Gladysz, Inorg. Chem. 1996, 35, 5509.
[64] Y. Wang, A. M. Arif, J. A. Gladysz, Organometallics 1994, 13, 2164.
[65] a) W. E. Buhro, S. Georgiou, J. M. Fernandez, A. T. Patton. C. E. Strouse,
J. A. Gladysz, Organomrtalhcs 1986, 5, 956; b) W E. Buhro, M. C. Etter, S.
Georgiou, J. A. Gladysz, F. B. McCormick, ibid. 1987, 6, 1150; c) F. B. McCormick, ibid. 1984, 3, 1924.
[66] W. A. Schenk, N. Burzlaff, H. Burzlaff, 2. Nalurforsch. B 1994, 49, 1633.
[67] D. P. Klein, D. M. Dalton, N. Quiros Mendez, A. M. Arif. J. A. Gladysz, J.
Organomet. Chem. 1991, 412, C7.
[68] a) For a titanium li ketone complex, see J. E. Hill, P. E. Fanwick. I. P. Rothwell, Organometalhcs 1992, 11, 1771; b) For zirconium li ketone complexes,
see F. R. Askham, K. M. Carroll, S. J. Alexander, A. L. Rheingold. B. S.
Haggerty, ibid. 1993, 12, 4810, and references therein; c) L. Giannini, E.
Solari, A. Zanotti-Gerosa, C. Floriani, A. Chiesi-Villa, C. Rizzoli. Angex.
Chem. 1996,108,79;Angeu. Chem. Int Ed. Engl. 1996,35,85;d) N. Peulecke,
A. Ohff, A. Tillack, W. Baumann, R. Kempe. V. V. Burlakov, U. Rosenthal,
Organometallics 1996, 15, 1340; e) For references to nickel and palladium II
ketone complexes, see Y.-H. Huang, J A. Gladysz, J. Chem. Educ 1988, 65,
298.
[69] One of the simpler approaches can be illustrated with cis-2-pentene. The
methyl and ethyl substituents are assigned (arbitrary) values. These act as
multipliers for the bar heights of the occupied quadrants. The sum is taken as
the steric energy With methyl = 1 and ethyl = 2, receptor A gives energies of
6, 7, 8, and 9 units for the four possible isomers.
[70] a) D. L. Lichtenberger, G. E. Kellog, Acc. Chem. Res. 1987, 20, 379; b) C.
Elschenbroich, A. Salzer, Organometallics,2nd ed.. VCH, New York, 1992, p.
47; c) J. R. Sowa, Jr., R. J. Angelici, J. Am. Chem. SOC.1991, 113, 2537.
[71] T.-S. Peng, C. H. Winter, J. A. Gladysz, Inorg. Chem. 1994, 33, 2534.
[72] Interestingly, the kinetic binding selectivitiesobtained for alkenes with I-Me,
are not higher than those with the less congested analog 1. In fact. in
582
the case of the I-pentene complex 5-Me:PF;
, the less stable R R S S
diastereomer forms faster [71].
[73] F. Agbossou, J. A. Ramsden, Y-H. Huang, A M. Arif, J. A. Gladysz,
Organometallics 1992, 11, 693.
[74] P. G. Gassman. J. W. Mickelson, J. R. Sowa, Jr., J . Am. Chem. SOC.1992,114,
6942.
[75] a) J. Issberner, R. Moors, F. Vogtle, Angew. Chem. 1994, 106, 2507; Angew.
Chem. Int Ed. Engl. 1994,33,2413; b) J. P. Collman, X. Zhang, K. Wong, J. I.
Brauman J . Am. Chem. SOC.1994, 116, 6245; c) E. Rose, A. Kossanyi, M.
Quelquejeu, M Soleilhavoup, F. Duwavran, N. Bernard, A. Lecas, J. Am.
Chem. Soc. 1996, 118, 1567.
[76] These generalizations are not necessarily limited to chiral trans or cis alkenes
with C , symmetry. Consider, for example, a racemic trans cycloalkene with a
remote gemindl dimethyl group in the backbone. Each configurational
diastereomer would now have two M-(C=C) rotamers. However, both rotamers of one diastereomer would likely be more stable than both rotamers
of the other. Alternatively, a C,-symmetric chiral receptor could be employed,
in which case each diastereomer would again have only one M-(C=C)
rotamer.
1771 a) M. A. Giardello, V. P. Conticello, L. Brard, M. Sabat, A. L. Rheingold,
C. L. Stern, T J. Marks, J. Am Chem. SOC.1994, lf6, 10212; b) M.A.
Giardello, V. P. Conticello, L. Brard, M. R. Gagne, T. J. Marks, ibid 1994,
116,10241 ;c) C. M. Haar, C. L. Stern, T. J. Marks, Organomefallics1996,15,
1765; d) For a similar recent analysis, see L. Bell, R J. Whitby, R. V. H. Jones,
M. C. H. Standen, Tetrahedron Lett. 1996, 37, 7139.
1781 a) J.-M. Zhuang, D. Sutton, Orgunometallics 1991, 10, 1516; b) For some
related alkene complexes, see Y.-X. He, R. J. Batchelor, F. W. B. Einstein, D.
Sutton, J . Organomet. Chem. 1996, 509, 37.
[79] a) M. Brookhart, J. R. Tucker, G. R. Husk, J . Am. Chem. Sac. 1983,105,258;
b) See also K. R. Aris, J. M. Brown, K. A. Taylor, J . Chem. SOC.
Dalton Trans.
1974,2222.
[SO] D. L. Reger, C. J. Coleman, P. J. McElligott, J . Organomet. Chem. 1979,17f,
73.
[81] D. L. Reger, C. J. Coleman, Inorg. Chem. 1979, 18, 3155.
[82] J. W. Faller, B. V. Johnson, C. D. Schaeffer, Jr., J . A m Chem. SOC.1976, 98,
1395.
[83] M Herberhold, H. Alt, C. G. Kreiter, Liebigs Ann. Chem. 1976, 300.
1841 a) K. H. Dotz, C. Lyon, J. Rott, J. Organomet. Chem. 1988,345,117;b) Many
adducts of [(q5-C,H,)Mo(NO)(C0)] and chiral alkenes have been reported. The following papers describe at least one complex of a prochiral alkene
formed by reaction of a IT allyl complex. NMR spectra showed only one
configurational diastereomer unless otherwise noted: c) transCH,CH=CHCH,C(CH,),CHO: R. D. Adams, D. F. Chodosh, J. W. Faller,
A. M. Rosan. J Am. Chem. Soc. 1979, 101, 2570; d) qz-1,3-cyclooctadiene:
J. W. Faller, K. H. Chao, H. H. Murray, Organometalhcs 1984, 3, 1231; e)
H,C=C(R)CH,X (80.20 diastereomer mixture for R = CH,, X = SPh).
W. E. VanArsdale, R. E. K. Winter, J K. Kochi, J Organornet. Chem. 1985,
296.31 ; f f H,C=C(CH,)CH,CH(CN),, H,C=CHCH,C(CO,CH,),CH,Ph:
Organometallics 1986, 5 , 645; g) H,C=CRCH,M(CO),: H.-J. Miiller, U.
Nagel, W. Beck. ihrd 1987,6,193; h) cyclopentenone. R. H. Yu, J. S. McCallum, L. S. Liebeskind, ibrd. 1994, 13, 1476.
[85] a) G. Consiglio, P. Pregosin, F. Morandmi, J. Organomet. Chem. 1986, 308,
345; b) G. Consigho, F. Morandini, bid. 1986, 310, C66; c) G Consiglio. E
Morandini, Chem. Rev. 1987, 87, 761
[86] a) A. D. Poole, V. C. Gibson, W. Clegg, J . Chem. Soc. Chem. Commun. 1992,
237; b) A. D. Poole, Ph.D Thesis, University of Durham, UK, 1992.
1871 V. C. Gibson, A. D Poole, J. Chem. Soc. Chem. Commun. 1995,2261.
[88] S E. Kegley, K. A Walter, D.T. Bergstrom, D. K. MacFarland, B. G.
Young, A. L. Rheingold, Organometalhcs 1993, 12, 2339.
I891 J. W. Faller, K. J. Chase, Orgonometallics 1995, 14, 1592
[90] C. R. Landis, J. Halpern, J . Am. Chem. Soc 1987, 109, 1746.
[91] Some of these studies have been reviewed: K. Saito, K. Kashiwabara, J.
Organornet. Chem. 1987,330, 291.
[92] a) H. Boucher, B. Bosnich. J . Am. Chem. SOC.1977, 99, 6253; b) R. G . Ball,
N. C. Payne. Inorg. Chem. 1977, 16, 1871, and references therein.
1931 S. Shinoda. Y. Sudo, Y. Yamaguchi. T. Iwayanagi, Y Saito, J. Organornet.
Chem. 1976, 121,93
[94] S . Shinoda, Y. Yamaguchi, Y. Saito, Inorg. Chem. 1979, 18, 673.
[95] a) A. de Renzi, B. di Blasio, A. Saporito, M. Scalone, A. Vitagliano, Inorg.
Chem. 1980,19,960; b) V. G. Albano, F. Demartin. A. De Renzi, G. Morelli,
Inorg. Chim. Acta 1988, 149. 253.
[96] H. van der Poel, G. van Koten. Inorg. Chem. 1981,20,2950
[97] a) L. E. Erickson, G. S. Jones, J. L. Blanchard, K. J. Ahmed, Znorg. Chem.
1991,30, 3147; b) L. E. Erickson, P. Hayes, J. J. Hooper, K. F. Morris, S. A.
Newbrougb, M Van Os, P. Slangen, Inorg. Chem. 1997, 36, 284; c) K. F.
Morris, L. E. Erickson, B. V Panajotova, D. W. Jiang, F. Ding ibid. 1997.36.
601.
[98] R. van Asselt, C. J. Elsevier, W. J. J. Smeets, A. L. Spek, Inorg. Chem. 1994,33,
1521.
199) M. E. Cucciolito, M. A. Jama, F Giordano, A. Vitaghano, V. De Felice,
Organometallics 1995. 14, 1152.
Angew. Chem. Int. Ed. E n ~ l 1997,36.
.
550-583
Chiral Recognition in n Complexes
[loo]Hydrogen bonding involving coordinated NRR'H groups has abundant
precedent. See ref. [102a]and a) M. A. Dewey, D. A. Knight, D. P. Klein,
A.M. Arif. J. A. Gladysz, Inorg:. Chem. 1991, 30. 4995; h) D. Braga, F.
Grepioni, P. Sahatino, G. R. Desiraju, OrganometaNics 1994, 13, 3532; c)
L.E. Erickson, unpublished results from NMR experiments with alkeneplatinum adducts with other types of hydrogen bond acceptors.
[I011 a) A. C. Cope, C. R. Ganellin, H. W Johnson, Jr., T. V. van Auken, H. J. S.
Winkler, J . Am. Chem. Soc. 1963,85,3276;
h) A.C. Cope, J. K. Hecht, H. W.
Johnson, Jr., H. Keller, H. J. S. Winkler, ibid. 1966,88, 761.
[lo21 For other adducts of chiral platinum Lewis acids and chiral alkenes, see a) G.
Uccello-Barretta, R. Lazzaroni, C. Bcrtucci, P. Salvadori, Organometallics
1987, 6, 550: b) R. Fulwood, D. Parker, G. Ferguson, B. Kaltner, J .
Organomet. Chem. 1991,119,269:
c) M .Goldman, Z. Kustanovich, S. Weinstein, A. Tishhee, E. Gil-Av, .L Am. Chem. Soc. 1982,104,1093;d) Examples
can he found in refs. [97],[99],and [100c].
[lo31 M. E. Cucciolito, F.Ruffo, A. Vitagliano, M. Funicello, Tetrahedron Lell.
1994,35, 169.
[I041 M. Kitamura, I. Kasahara, K. Manahe, R. Noyori, H. Takaya, J . Org. Chem.
1988,53, 708.Additional examples and analyses are given in ref. [I].
[105]a) J. P. Morken, M. T. Didiuk, M. S. Visser, A. H. Hoveyda, J. Am. Chem.
Sue. 1994,f16,3123;h) M. S. Visser, J. P. A. Harrity, A. H. Hoveyda, ibid.
1996,ii8,3779;c) M. S.Visser, N . M. Heron, M. T. Didiuk, J. F. Sagal, A. H.
Hoveyda, ibirf. 1996,ii8, 4291.
[lo61 A. Viso, N. E. Lee, S. L. Buchwald, J . Am. Chem. Soc. 1994,li6,9373.
[I071 S.D. Orth, J. Barrera, M. Sahat, W. D. Harman, h o g . Chem. 1994, 33,
3026.
[I081 Recent lead references: a) J. S. Giovannetti, C. M. Kelly. C. R. Landis, J. Am.
Chem. Soc. 1993,f15,4040;h) M.J. Burk, J. E. Feaster, W. A. Nugent, R. L.
Harlow, ibid. 1993,ii5,10125;c) J. Alhrecht, U. Nagel, Angew. Chem. 1996,
108,444;Angew. Cheni. Int. Ed. Engl. 1996,35,407.
[lo91 This assumes that coreactants attack each diastereomer in an identical manner (for example at the C=C face syn or anti to the metal), as in a
stereospecific [8]reaction.
[I101 a) T . 3 . Peng, A. M. Arif, J. A. Gladysz, J . Chem. Soc. Dalton Trans. 1995,
1857;h) For these additions, the reactants and products have identical relative
configurations at rhenium. However, the R/S designation changes [II,381.
[ll11 D.M. Dalton, C. M. Garner, J. M. Fernandez, J. A. Gladysz. J. Org. Chem.
1991,56,6823.
[112]D.P. Klein, J. A. Gladysz, J . Am. Chem. Soc. 1992,lid, 8710.
. 1993,115,
[I131 B. E. Thomas, IV, J. D. Evanseck, K. N. Houk, J. Am. C h ~ mSoc.
4165.
[114]M. B. Carter, B. Schi~tt,A. Gutierrer, S. L. Buchwald, J . Am. Chem. SOC.
1994,il6,1 1 667.
Angew. Chem. Inl. Ed. Engl. 1997,36,550-583
REVIEWS
[I151 C. A. Willoughhy, S. L. Buchwald, J. Am. Chem. SOC.1994,116, 11 703,and
earlier work cited therein.
[I161 D. Y. Kondakov, E. Negishi, J . Am. Chem. Soc. 1996.ii8, 1577.
[117]Review: A. H. Hoveyda, J. P. Morken, Angew. Chem. 1996, 108, 1378;
Angew. Chem. Inf. Ed. Engl. 1996,35,1263.
[I181 Strictly speaking, the exact enantiomer ratios would (assuming that the
diastereomeric x X=C complexes equilibrate rapidly relative to the rate of
migration) remain under Curtin-Hammett control.
[119]Chelating alkene adducts have been characterized: Z. Wn, R. F.Jordan, J. L.
Petersen, J . Am. Chem. Soi~.1995,117,5867.
[I201 B. J. Burger, B. D. Santarsiero, M. S. Trimmer, J. E. Bercaw, J . Am. Chem.
Soc. 1988,ii0,3134.
[I211 a) B. M. Trost, C. B. Lee, J. M. Weiss, J. Am. Chem. Soc. 1995,117,7247:h)
B. M. Trost, Acc. Chem. Rcs. 1996,29, 355.
[122]a) M. Brookhart, M. 1. Wagner, J . Am. Chem. Soc. 1996, 1i8, 7219, and
references therein; h) Z. Jiang, M. T. Boyer, A. Sen, ibid. 1995,117.7037.
[I231 Some recent lead references: a) M. P. Doyle, W R. Winchester, J. A. A.
Hoorn, V. Lynch, S. H. Simonsen, R. Ghosh, J . Am. Chem. Soc. 1993,ii5,
9968;h) M.P. Doyle, Aldrichimica Actu, 1996,29, 3; c) H. Nishiyama, Y
Itoh, Y. Sngawara, H. Matsumoto, K. Aoki. K. Itoh. Bull. Chem. Soc. Jpn.
1995,68, 1247:d) A. B. Charette, C. Brochu, .I. Am. Chem. SUC.1995,f f 7 ,
1 1 367.
[124]Recent lead references for mechanistic aspects of the Jacohsen epoxidation:
a) W. Zhang, N. H. Lee, E. N. Jacohsen, J . Am. Chem. Sac. 1994,ii6,425;
h) S. Chang, J. M. Galvin, E. N . Jacobsen, ibid. 1994, 116. 6937; c) M.
Palucki, P. J. Pospisil, W. Zhang, E. N. Jacohscn, ibid. 1994,ii6,9333.
[125]Recent lead reference for mechanistic aspects of the Sharpless epoxidation:
Y-D. Wu, D. K. W. Lai, J . Am. Chem. SOC.1995,ii7,1 1 327.
[126]E. J. Corey, C. J. Helal, Tetrahedron Lett. 1995,36,9153.
[127]E.Haputman, S.Saho-Etienne, P. S.White, M. Brookhart, J. M. Garner,P. J.
Fagan, J. C. Calahrese, J. Am. Chem. SOC.1994,ii6,8038.
[I281 H. F.Luecke, B. A. Arndtscn, P. Burger, R. G. Bergman, .
I
Am. Chem. SOC.
1996,li8,2517.
[129]H. G. Alt, R. Zenk, J. Orgunomet. Chem. 1996,522,177.
[130]A. Mayr, A. M. Dorries, G. A. McDermott, S. J. Geih, A. L. Rheingold,
J. Am. Cheni. Soc. 1985,107,7775.
[I311 Y Kim, J. Gallucci, A. Wojcicki, OrganomrtaNic~1992,11, 1963.
[I321 G.Schmid, B. Kilanowski, R. Boese, D. Blaser, Chem. Ber. 1993,126,899.
[133]B. Jedlicka, R.E. Rulke, W. Weissensteiner,R. l:ernander-Galin, F. A. Jalon,
B. R. Manzano, J. Rodriguez-de la Fuente, N. Veldman, H. Kooijmdn, A. L.
Spek, J. Orgunomet. Chem. 1996,5i6,97.
[134]R. W. Quan, Z. Li, E. N. Jacohsen, J. Am. Chem. Soc. 1996,ii8,8156.
583
Документ
Категория
Без категории
Просмотров
2
Размер файла
4 113 Кб
Теги
development, acid, enantioface, mode, general, selectivities, recognition, transitional, complexes, alkenes, chiral, aldehyde, metali, ketone, lewis, binding
1/--страниц
Пожаловаться на содержимое документа