Chirality in the absence of rigid stereogenic elements: The design of configurationally stable C3-symmetric propellers

Article abstract of DOI:10.1002/chem.200801505

Residual stereoisomerism is a form of stereoisomerism scarcely considered so far for applicative purposes, though extremely interesting, since the production of stereoisomers does not involve classical rigid stereogenie elements. In three-bladed propeller

Full text of DOI:10.1002/chem.200801505

DOI: 10.1002/chem.200801505
Chirality in the Absence of Rigid Stereogenic Elements: The Design of
Configurationally Stable C -Symmetric Propellers
3
[
a]
[b]
[c]
[c]
Tiziana Benincori, Andrea Marchesi, Tullio Pilati, Alessandro Ponti,*
[
c]
[c, d]
Simona Rizzo, and Francesco Sannicolꢀ*
Dedicated to Sofia Giovanna
Abstract: Residual stereoisomerism is
a form of stereoisomerism scarcely
considered so far for applicative pur-
poses, though extremely interesting,
since the production of stereoisomers
does not involve classical rigid stereo-
genic elements. In three-bladed propel-
ler-shaped molecules, a preferred ster-
eomerization mechanism, related to
the correlated rotation of the rings,
allows the free interconversion of ste-
reoisomers inside separated sets (the
residual stereoisomers) that can inter-
convert through higher energy path-
ways. In light of possible future appli-
cations as chiral ligands for transition
metals in stereoselective processes,
some C -symmetric phosphorus-cen-
stationary phase (CSP HPLC) at a
semipreparative level at room tempera-
ture. Stability was evaluated through
different techniques (circular dichroism
(CD) signal decay, dynamic CSP
HPLC (CSP DHPLC), dynamic NMR
analysis (DNMR)) and the results com-
pared and discussed. Phosphanes were
found much less stable than the corre-
sponding phosphane oxides, for which
preliminary calculations suggest that
the three-ring-flip enantiomerization
3
tered propellers, which could exist as
residual enantiomers, are synthesized
and the possibility of resolving their
racemates into residual antipodes is ex-
plored. While the tris
are configurationally stable at room
temperature, only selected tris-
AHCTUNGTRENN(GUN aryl)phosphane oxides display a con-
ACHTUNGTRENNU(GN aryl)methanes
figurational stability high enough to
allow resolution by HPLC on a chiral
mechanism (M ) would be easier than
0
Keywords:
chirality
·
circular
phosphorus pyramidal inversion. The
parameters affecting the configuration-
al stability of the residual enantiomers
dichroism · NMR spectroscopy ·
phosphane oxides · phosphanes ·
residual enantiomers
of C -symmetric propellers are dis-
3
cussed.
Introduction
The metal complexes of tris
ACHTUNGTRNENUNG( aryl)phosphanes bearing three
[
a] Prof. T. Benincori
Dipartimento di Scienze Chimiche ed Ambientali
Universitꢀ dell’Insubria
identical aryl rings are amongst the most popular mediators
[1]
in homogeneous catalysis. It is common knowledge that
they are propeller-shaped chiral compounds, but chemists
have always accepted their configurational instability as an
unavoidable event due to the very easy helix reversal, and
rarely contemplated the possibility of taking profit of their
inherent chirality.
via Valleggio 11, 22100 Como (Italy)
[
b] Dr. A. Marchesi
Laboratori Alchemia s.r.l.
via S. Faustino 68, 20134 Milano (Italy)
[
c] Dr. T. Pilati, Dr. A. Ponti, Dr. S. Rizzo, Prof. F. Sannicolꢁ
Istituto di Scienze e Tecnologie Molecolari
Consiglio Nazionale delle Ricerche
via Golgi 19, 20133 Milano (Italy)
Attempts to investigate the facial recognition ability of
propeller helicity generally follows the strategy of introduc-
ing chiral substituents on the aryl rings of tris-
Fax : (+39)0250314300
E-mail: alessandro.ponti@istm.cnr.it
ACHTUNGTRENNUNG( aryl)phosphanes in the hope that they could stabilize a pre-
[2]
[
d] Prof. F. Sannicolꢁ
ferred helical configuration. It is, however, difficult to dis-
criminate the effects attributable to the helix from those re-
ferable to the substituents. Some chiral configurationally
stable metal complexes of a tris{3-[1-(diphenylphosphino)-2-
Dipartimento di Chimica Organica e Industriale
C.I.MA.I.NA., Universitꢀ di Milano
via Venezian 21, 20133 Milano (Italy)
Fax : (+39)0250314139
E-mail: francesco.sannicolo@unimi.it
94
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

FULL PAPER
methyl]benzimidazolyl}methane were recently described
within closed subsets, which constitute the residual diaste-
[3]
and applied in a carbon-carbon bond-forming reaction.
We considered that, at least in principle, chiral and config-
urationally stable tris(aryl)phosphanes could be obtained as
reoisomers. The absolute preference for the M mode had
1
1
was demonstrated recently by H NMR spectroscopy in the
A
H
U
G
R
N
U
case of an unlabeled tris
ACHTUNGTRNENUNG( mesytyl)phosphonium cation asso-
[10]
residual enantiomers as long as their structural architecture
could satisfy the conditions required for the existence of re-
sidual stereoisomerism, a phenomenon discovered by Prof.
ciated to chiral C -symmetric counterions.
3
In the case of 1, two residual racemic diastereoisomers
were separated by crystallization. Each racemate is made up
of two enantiomeric non-interconvertible (R,S) sets of eight
diastereoisomers, rapidly interconverting at 878C by the M1
mechanism with stereomerization barriers of about 10–
[4]
K. Mislow in the 1970s. Residual stereoisomers “result
whenever closed subsets of appropriately substituted inter-
converting isomers are generated from a full set under the
operation of a given stereomerization” mechanism. The re-
sidual stereoisomers isolated so far are produced by two
similar phenomena working in very different molecular ar-
chitectures: the dynamic gearing operating in the case of
ꢀ
1
15 kcalmol . The interconversion of diastereomeric residual
racemates started to be active at 1228C, with a quite high
°
ꢀ1
barrier (DG =30.5 kcalmol at 1228C) as demonstrated by
1
H dynamic NMR detected kinetics.
[5]
[6]
bis(trypticyl)methanes and ethers and the correlated ro-
tation of the rings in the case of three-bladed molecular pro-
As two configurationally stable diastereomeric couples of
enantiomers are observable in the case of propeller 1, it is
expected that a couple of configurationally stable enantio-
mers would be obtained by removing the stereogenicity of
the carbon atom on the hub, simply by using at least two
identical aryl rings connected to it. This was demonstrated
eighteen years ago by the successful chromatographic reso-
[7]
[8]
pellers, such as tris
ACHTUNGTRENNUNG( aryl)methanes and tris ACHTUNGTRENUNNG( aryl)amines,
in which three aryl rings devoid of local C axes (the blades)
2
[9]
radiate from a central carbon or nitrogen atom (the hub).
The first classical example belonging to the latter class of
molecules is the maximally labeled (three differently substi-
tuted aromatic rings) compound 1, for which a full set of 32
chiral stereoisomers are predicted. They are generated by
lution of the residual racemate of C -symmetric tris[1- ACHTUNGTRENNUNG( 2-
3
11]
[
methyl)benzimidazolyl)]methane (2). Also in this case the
enantiomerization barrier was found to be considerable
°
ꢀ1
(
DG =28.5 kcalmol at 708C).
The crucial role played by the nature of the atom located
on the hub of the helix was clearly demonstrated by Mislow
[8]
in the case of the maximally labeled tris ACHUTNGRNENUG( aryl)amine 3. Two
achiral diastereomers are expected for this system, since fast
nitrogen pyramidal inversion corresponds to a planar equi-
librium geometry analogous to that shown by tris-
AHCTUNGTRENNUNG( aryl)boranes. The separation of the residual diastereoiso-
mers was achieved by taking advantage of their different,
easily distinguishable, crystalline forms. The configurational
stability of the residual diastereoisomers was found to be
much lower than that exhibited by the carbon-centered pro-
°
ꢀ1
the stereocenter lying on the hub, by the helicity, and by the
three different simple rotors describing the conformations
pellers (DG =17.7 kcalmol at ꢀ218C), probably due to
the fact that the level of the dynamic gearing amongst the
rings, which is responsible for the configurational stability of
the residual stereoisomers, is strongly diminished when pyra-
midality is lost, and the aryl rings diverge in the transition
state required for the nitrogen inversion process.
resulting from the rotation of the aryl rings around the Csp
Caromatic bonds. In other words, ring rotation corresponds to
3
ꢀ
ring edge interchange with respect to the reference plane
2
(
the plane defined by the three neighboring sp carbon
3
atoms bound to the central sp carbon atom). It has been
firmly established that in these systems “the motion of the
rings is coupled” and that any isomerization process in-
The present paper reports the results of the development
of a work of which we have given only a short preliminary
account in the case of tris[3- AHCTUNETGRNNUN(G 2-alkyl)indolyl]phosphane
[7]
[4c]
[12]
volves helicity reversal.
oxides 4a, 4e, and 4 f. The multi-targeted project was fo-
There are four theoretically possible stereomerization
cused on:
mechanisms M (n=0, 1, 2, 3), which are characterized by
n
the number n of rings undergoing edge interchange. The
rings not involved in edge interchanging are said to flip, that
1) Designing, on the basis of previous experiments and
DFT calculations, phosphorus-centered propellers of the
is they pass through the plane containing the Csp
3
ꢀC
Ar P-X type (X=oxygen, lone pair), devoid of any rigid
stereogenic element, which would be promising for gen-
erating stable residual enantiomers.
aromatic
3
bond and perpendicular to the reference plane. The M1
mode, which involves one edge interchange and two ring-
flips, is, without known exceptions, the threshold stereomeri-
zation mechanism. This mechanism does not allow free in-
terconversion within the full set of stereoisomers displaying
the same configuration at the stereogenic carbon, but only
2) Synthesizing the compounds.
3) Structurally characterizing the new compounds, possibly
by XRD analysis.
4) Resolving the residual racemates.
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
95

A. Ponti, F. Sannicolꢁ et al.
modular series of compounds 4, in which the effects of the
different size of the group located in the 2-position of the
indole ring could give interesting information on the effi-
ciency of the correlated ring rotation and, in turn, on the
configurational stability of the residual enantiomers. In this
light, we prepared phosphane oxides 4a, 4e, and 4 f, in
which the correlated rotation of the ring should be progres-
sively enhanced.
We also considered worth synthesizing the tris[1- ACHTUNGTRENNUNG( 2-meth-
yl)benzimidazolyl]phosphane oxide (5), which, unlike 2, is a
phosphorus-centered system with identical blades. It could
give interesting quantitative information on the role of the
atom located on the propeller axis, since it controls both the
hub–blade distance and the pyramidality of the systems. In
this light, in order to perform a direct comparison under
identical experimental conditions, we prepared and rein-
5
6
) Evaluating the enantiomerization barriers of the residual
enantiomers by different techniques (circular dichroism
vestigated known tris[1- AHCTUNGTERUNNN(G 2-metyl)benzimidazolyl]methane
(2).
(
CD), CSP dynamic HPLC (CSP DHPLC), and
1
H NMR spectroscopy)
) Discussing their dependence upon some specific structur-
al features.
Results and Discussion
Computational study of compound 4: We resorted to com-
putational electronic structure methods to check a few
points that we considered crucial for the successful develop-
ment of the research project. Some results at the B3LYP/3-
In particular, we planned to consider the effects produced
on configurational stability by four structural parameters:
[12]
1
) The hub–aryl bond length: the length of the PꢀC
21G* level have already been published.
There we
aromatic
[15]
bond is much greater (1.78 ꢃ) than that exhibited by the
showed that the Psss/Msss are the major isomers and that
Csp
the C
3
ꢀC
bond (1.54 ꢃ) in 1 and, even more so, by
bond (1.45 ꢃ) in 2. Thus, the correlated
aromatic
the M enantiomerization barrier significantly depends on
aromatic
0
3
ꢀN
the size of substituent R to the extent that 4 f is configura-
tionally stable enough to be resolved into residual enantio-
mers at room temperature, whereas 4a is not. We extended
the above investigation by performing both HF and DFT-
B3LYP calculations using a much larger basis set (cc-
pVDZ). The solvent effect has also been studied within the
sp
rotation of the rings is expected to be less effective in
phosphorus- than in carbon-centered propellers.
2
3
) The aryl-hub-aryl angles characterizing the propeller: the
wider the angles, the lower the engagement of the rings,
and the lower the expected configurational stability.
) The bulkiness of the substituents located in the ortho po-
sition on the aromatic rings: the bulkier the groups, the
stronger the inter-ring dynamic gearing, and then the
higher the configurational stability
[16]
framework of the polarized continuum model (PCM).
The most significant energetic results are shown in Table 1.
We begin considering in vacuo results. Both HF and
B3LYP results indicate that the sss isomers are the most
stable and largely the major ones. The ssa conformers repre-
sent a few percent of the isomers, whereas the aas and the
aaa diastereoisomers are virtually absent. Since 4a, 4e, and
4 f essentially comprise the sss and ssa stereoisomers, we cal-
4
) The barrier for the pyramidal inversion of the phospho-
rus atom in tris
higher than in tris
antee configurational stability at room temperature.
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
[13]
culated the activation barriers for the M enantiomerization
0
We started the research by paying attention to the behav-
Msss!Psss and for the M diastereomerization Psss!Mssa
1
ior of phosphane oxides, since we wanted to preliminarily
check whether an efficient dynamic gearing could be ach-
ieved in non-inverting phosphorus-centered propellers. Fur-
thermore, efficient methodologies are available for the reso-
lution of racemic phosphane oxides, while the resolution of
racemic phosphanes is more difficult (particularly at an ana-
lytical level) so that enantiopure phosphanes are generally
prepared by reduction of the corresponding enantiopure
(or Msss!Pssa) processes. (The M and M stereomeriza-
2
3
tion paths are virtually impossible, because of the huge
steric hindrance produced when more than one indolyl
moiety undergoes edge interchange.) In all cases the M bar-
1
rier is much lower than the M barrier, so that the M dia-
0
1
stereomerization within each residual enantiomer is much
faster than the M interconversion between residual enantio-
0
mers.
At
room
temperature
(thermal
energy
[14]
ꢀ1
phosphane oxides.
ꢁ0.6 kcalmol ) the M processes can be considered fast
1
[
17]
We have chosen the 1,2-disubstituted-3-indolyl group as
the aryl ortho-differentiated aromatic ring, since its synthetic
versatility offers the possibility of acceding to an ample
from the standpoint of configurational stability.
In con-
trast, the barriers for the M processes are large enough at
0
room temperature that the individual residual enantiomers
96
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

Chiral Compounds
FULL PAPER
Table 1. Calculated energetics of tris[3-(1-methyl-2-R)indolyl]phosphane
[
a]
oxides 4a (R=Me), 4e (R=Et), and 4 f (R=iPr).
°
[b]
°
[c]
DE
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(ssaꢀsss)
P
A
H
U
G
R
N
U
G
A
H
U
G
R
N
U
(ssa)
DE (M
0
)
1
DE (M )
ꢀ
1
ꢀ1
ꢀ1
[kcalmol
]
[kcalmol
]
[kcalmol ]
2
2.6
1
.7
97.0:3.0
96.6:3.4
72.4:27.6
25.2
19.9
19.9
6.9
7.0
4.0
4
4
4
a
e
f
.2
3
2.8
1
.7
99.4:0.6
97.4:2.6
68.5:31.5
26.0
20.2
19.1
10.6
7.7
5.3
.1
4
.1
99.7:0.3
98.4:1.6
75.1:24.9
32.1
24.6
23.6
11.3
8.0
5.9
3.1
.3
1
[
a] The energy difference between the two most stable sss and ssa iso-
mers DE
(ssaꢀsss) and the corresponding population ratio (P) ratio at
room temperature are reported along with the activation energies for se-
lected M and M isomerization processes. Italic: HF/cc-pVDZ, Bold:
ACHTUNGTRENNUNG
0
1
B3LYP/cc-pVDZ, Roman: PCM-B3LYP/cc-pVDZ. [b] For the Msss !
Psss process. [c] For the Psss!Mssa or Msss!Pssa processes; for the re-
verse processes the DE
from the reported value.
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(ssaꢀsss) energy difference should be subtracted
Figure 1. Selected B3LYP/cc-pVDZ optimised structures of tris[3-(1-
methyl-2-isopropyl)indolyl]phosphane oxide (4 f). a) Msss isomer; b)
transition state of the M isomerization from Msss to Pssa; c, d) transi-
1
can be observable by chromatographic methods at this tem-
perature. Moreover, these results suggest that isolation of
configurationally stable residual enantiomers could be feasi-
ble. It is also worth noting that 4a and 4e have similar M0
barriers despite ethyl being significantly bigger than methyl,
whereas when R is iPr the barrier is ꢁ4.5 kcalmol larger.
This can be explained by inspecting the molecular structural
changes occurring upon reaching the transition state, as de-
tailed below.
tion state of the M isomerization from Msss to Psss: c) view from above
the oxygen atom, d) view from below the phosphorus atom with arrows
representing the eigenvector related to the stereomerization mode.
0
matic moieties due to the increased size of the R substitu-
ent, which pushes the portions of the aromatic rings close to
the phosphorus apart and forces the lower portions to come
into contact. This can be illustrated by the variation of sev-
ꢀ
1
eral geometric parameters in the M transition state, for ex-
0
Some B3LYP/cc-pVDZ selected structures of 4 f are re-
ported in Figure 1. The main structural features of the sss
isomers of 4a, 4e, and 4 f do not significantly depend on the
R substituent: the CꢀP bond length is 1.82 ꢃ, the C-P-C
ample, the angle P-C -C is 134.2, 134.3, and 132.98 in 4a,
4e, and 4 f, respectively.
3
3a
To investigate the possibility of resolving the residual
racemates by chiral reverse-phase HPLC analysis, we also
performed PCM calculations to estimate the solvent effect
of an acetonitrile/water 1:3 mixture on the energetics of
these phosphane oxides. The results of PCM-B3LYP calcula-
tions using averaged solvent parameters are marginally dif-
ferent from those obtained by averaging the results for pure
acetonitrile and pure water.
angle is 1088, and the aromatic moieties make an angle of
about 358 with the P=O bond. At the corresponding M0
transition state, the angle between the aromatic moieties is
evidently close to 08, the C-P-C angle is 1128, and the CꢀP
bond length is 1.86 ꢃ. As before, these length and angles
negligibly depend on the substituent R. Therefore, for sys-
tems differing by an ortho substituent only, such as 4a, 4e,
and 4 f, the unequal heights of the activation barriers do not
depend on the deformation about the central phosphorus
atom. The nature and components of the eigenvector related
to the reactive mode with negative curvature (Figure 1d) in-
Inclusion of the solvent effects due to the acetonitrile/
water 1:3 mixture at the PCM-B3LYP/cc-pVDZ level re-
duces the differences between the minimum energy struc-
ꢀ1
tures by 1–2 kcalmol . Even if the energetic change is
rather small in absolute, this change has a large effect on
the stereoisomer population, because it is a significant frac-
tion of the energy difference. As a consequence, the molar
fraction of the ssa stereoisomers becomes significant (25-
32%), whereas the saa and aaa isomers are in the small per-
centage range. This behavior indicates that the isomer
energy differences are based on electrostatic interactions in
addition to steric repulsion interactions. Inclusion of the sol-
vents effects leads to a decrease in the stereomerization acti-
dicate that the crucial region of the M transition state is
0
that below the P=O moiety in which the aromatic rings
come to a close contact. Indeed, the hydrogen atoms in the
4-position of the indole moiety are as close as 1.79 ꢃ in 4a
and 4e, and 1.77 ꢃ in 4 f, that is, less than twice the hydro-
[18]
gen van der Waals radius (1.0ꢂ0.1 ꢃ). This causes a small
deviation from planarity of the indole rings at the transition
state of 4a and 4e, about 18 and 38 respectively, but is as
large as 178 at the transition state of 4 f. Thus, the substitu-
ent-induced difference in the activation barriers can be
traced back to a larger deviation from planarity of the aro-
ꢀ
1
vation barriers amounting to 0–1 kcalmol for the M pro-
0
ꢀ
1
cesses and to 2–3 kcalmol for the M processes. The barri-
1
er height change due to solvent effects is again rather small,
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
97

A. Ponti, F. Sannicolꢁ et al.
but now also in a relative sense, especially for M barriers
0
(
<5%). For M0 stereomerizations, steric repulsion is the
major contribution to the activation barrier, as discussed
above.
Thus, the presence of the polar acetonitrile/water 1:3 mix-
ture largely affects the isomer populations, makes the dia-
stereomerization within the residual enantiomers sets faster,
and leaves approximately unchanged the enantiomerization
rate between residual enantiomers.
Figure 2. ORTEP presentations of a single enantiomer of: 4c (left), 4h
In conclusion, the picture arising from the computational
results is that 1) P/Msss are by far the major isomers pres-
(
middle), and 5 (right).
ent, 2) M diastereomerization within each residual enantio-
1
mer is fast at room temperature, and 3) residual enantio-
Table 2. Structural parameters of three-bladed molecular propellers from
XRD data.
mers should undergo rather slow M enantiomerization at
0
room temperature.
Substrate
Hub–Aryl
bond length [ꢃ]
Aryl-Hub-Aryl
angle [8]
[
a]
[a]
Synthesis of phosphane oxides 4 and 5: Phosphorus oxybro-
mide reacts with 1,2-disubstituted indoles in dry pyridine, di-
rectly affording tris(3-indolyl)phosphane oxides 4a and 4c–
4
4
4e
4
4
5
2
3
a
c
1.781
1.795
1.777
1.779
1.773
1.674
1.448
1.441
107.1
106.8
106.7
106.9
107.2
105.4
111.9
117.3
f
h
4
f. By using 2-methylbenzimidazole as a substrate, 5 is ob-
tained in good yields. In the synthesis of 4e, we observed
that significant amounts of the C -symmetric phosphane
1
oxide 4h, isolated by chromatography, were formed if the
starting 2-ethyl-1-methylindole contained some 1-methylin-
dole. Compound 4h could be obtained in good yields by
using a 2:1 mixture of 2-ethyl-1-methylindole and 1-methyl-
indole.
The reaction of 4c with boron tribromide in methylene di-
chloride afforded the tris[3-(2-bromomethyl-1-methyl)indo-
lyl]phosphane oxide (4g), which was converted, in a crude
state, into 4b in quantitative yields by reaction with lithium
aluminium tetradeuteride in THF.
[
a] Mean value over the three blades
conformationally free unconstrained tris ACHUTNGRNEGUN( aryl)phosphane
[19]
oxides, while it is definitely narrower in the case of 5. Fur-
thermore, the PꢀN bond length in 5 (1.67 ꢃ) is much short-
er than in the indole-based phosphane oxides 4 (1.78 ꢃ),
even though it is longer than the Csp
3
ꢀN bond in 2 (1.45 ꢃ).
The combination of these parameters suggests that they
should synergistically cooperate to raise the configurational
stability of the residual enantiomers, which should increase
along the sequence 4a<5<2, as a consequence of the paral-
lel increase of the inter-ring gearing.
Characterization of phosphane oxides 4 and 5: All of the
tris(3-indolyl)phosphane oxides 4 and the benzimidazole de-
rivative 5 were characterized on the basis of their analytical
1
and spectral data, in particular H NMR and MS spectra. A
single enantiomeric couple (presumably Psss/Msss) was in-
Resolution of the residual racemates of phosphane oxides 4,
5, and of tris[1- ACHTUNGTERNNUN(G 2-methyl)benzimidazolyl]methane (2): Pre-
liminary experimental evidence for the existence of tris[3-
1
variably detected by H NMR spectroscopy down to ꢀ908C
in all cases.
As for the single crystal XRD structures, the data are now
available for all compounds, except for 4b, the trisdeuterat-
ed form of 4a, and 4g, which was isolated only in a crude
state as an intermediate for the synthesis of 4b. The XRD
structures of 4c, 4h, and 5 are shown in Figure 2; those of
(1,2-dimethyl)indolyl]phosphane oxide 4a as a couple of re-
sidual enantiomers was given by H NMR analysis. Progres-
1
sive addition of europium tris[3-(eptafluoropropyl-hydroxy-
methylene)-(+)-camphorate] to a solution of 4a in CDCl3
produced clear splitting of the sharp signal attributable to
the methyl group in the 1-position of the indole rings. The
same effect was observed by treating the solution of 4a in
[12]
4
a, 4e, and 4 f were previously published.
In all cases both enantiomeric forms are present in the
crystalline structure (the crystal space groups contain an in-
version centre). The Mssa and Pssa are the main conformers
populating the crystalline form of 4c and 4h, while the Psss/
Msss are the sole isomers present in the case of 5, in analogy
with the structures found for 4a, 4e, and 4 f.
The hub–aryl bond length and the aryl-hub-aryl angle,
which are relevant structural parameters directly influencing
the gearing level of the blades, are summarized in Table 2.
It is interesting to note that the C-P-C angle value in 4c
and 4h falls in the range of the classical values found for
CDCl with an equimolar amount of (+)-Naproxen. Surpris-
3
ingly, in both cases, the N-Me singlet was the only signal to
undergo splitting under the effect of these chiral additives.
The experimental proof for the existence of 4a as a couple
of residual enantiomers was achieved by CD and UV-detect-
ed chiral stationary phase (CSP) HPLC (ChromTech, Chiral
AGP 100ꢄ4.0 mm, 5 mm; eluant: acetonitrile/K HPO4
2
ꢀ
1
buffer (pH 4.5, 1 mgmL ) 3:1). A plateau interconnecting
the peaks in the UV-detected chromatogram at room tem-
perature suggested that the enantiomerization process is
98
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

Chiral Compounds
FULL PAPER
active at this temperature, even though at a quite slow
rate.
We were able to analytically resolve the residual racemate
of phosphane oxide 4 f, even though the difference in the re-
tention times of the antipodes was not as high as in the case
of 4e (Figure 3b, black curve). Also in this case, we ob-
tained some milligrams of almost enantiopure residual anti-
podes of 4 f by semipreparative CSP HPLC: the first eluted
enantiomer was found to be dextrorotatory (589 nm,
CH Cl ) and was isolated in a 99.9:0.1 e.r. (Figure 3b, red
[12]
A very similar behavior was shown by the tris[3-(2-me-
thoxymethyl-1-methyl)indolyl]phosphane oxide (4c). A pla-
teau between the peaks of the enantiomers was clearly visi-
ble in the chromatogram at room temperature, indicating
that slow interconversion of the antipodes was occurring
during elution (see below).
2
2
CSP HPLC experiments performed on 4e demonstrated
that complete analytical resolution was achieved In this
case, the difference in the retention times of the residual an-
tipodes was about four minutes (Figure 3a, black curve).
curve), while the e.r. of the second eluted levorotatory enan-
tiomer was 98.5:1.5 (Figure 3b, green curve).
The CSP HPLC resolution of the residual racemate of 5
was found more difficult than those previously performed in
the case of the indole-based phosphane oxides, due to the
small difference in the retention times of the antipodes (Fig-
ure 3c, black curve). Only the more retained levorotatory
antipode could be isolated in a good enantiomeric purity
level (e.r ꢁ98.7%, Figure 3c, green curve).
As anticipated, we also tried to resolve the already known
tris[1- ACHUTNGRENNUG( 2-methyl)benzimidazolyl]methane (2), since we con-
sidered essential that the experimental conditions employed
for evaluating the configurational stability would be as simi-
lar as possible for all the substrates. The CSP HPLC resolu-
tion of racemic 2, prepared according to the procedure de-
[11]
scribed in the literature,
was rather troublesome, due
again to the small difference in the retention times of the
antipodes. Attempts to reproduce the chromatographic
methodology described in literature (repeated medium-pres-
sure chromatographic processes on microcrystalline cellu-
lose triacetate with 95% ethanol as eluant) failed, since the
eluted products showed an unsatisfactory enantiomeric
purity. In addition, some degradation of the starting material
was observed. A nearly enantiopure sample of both enantio-
mers was obtained by employing a mixture of MeCN and
sodium acetate buffer (pH 4.5) in a 15:85 ratio. A more re-
tained achiral impurity, probably resulting from the loss of
one benzimidazole unit, was present in the resolved enantio-
mers. The first eluted dextrorotatory enantiomer was ob-
tained with an e.r of 94.5:5.5 (Figure 3d, red curve), while
an e.r. of 93.5:6.5 was found for the second eluted antipode
(Figure 3d, green curve).
Figure 3. CSP HPLC chromatograms and enantiomeric purity control of
the residual enantiomers obtained from semipreparative HPLC resolu-
tion on chiral stationary phase: black line: residual racemate, red line:
first eluted residual enantiomer; green line: second eluted residual enan-
tiomer. a) 4e, b) 4 f, c) 5, d) 2. An achiral impurity, formed during the
chromatographic resolution process together with 2-methylbenzimida-
zole, is visible at about 9 min. elution time.
The resolution process was scaled to semipreparative level
and solutions of nearly enantiopure residual enantiomers in
the eluant were isolated. The first eluted dextrorotatory en-
antiomer was obtained with an enantiomeric ratio (e.r.)
larger than 97.5:2.5 (Figure 3a, red curve), while the more
retained levorotatory isomer displayed an e.r. of about
Evaluation of the enantiomerization barriers of residual en-
antiomers: For the quantitative evaluation of the thermody-
namic parameters involved in the enantiomerization process
of the residual enantiomers, three techniques were consid-
ered: CD decay kinetics, dynamic CSP HPLC, and dynamic
95.0:5.0 (Figure 3a, green curve).
The removal of one of the three ortho ethyl groups from
e produced the total loss of configurational stability at
1
4
H NMR spectroscopy.
room temperature. Attempts to resolve 4h into antipodes
by CSP HPLC, under the experimental conditions which so
efficiently resolved all the indole-based phosphane oxides
investigated in this paper, were unsuccessful.
This result emphasizes the importance of the bulkiness of
the ortho substituents in developing an efficient correlated
rotation of the rings. As expected, an increase in the size of
the ortho alkyl group from ethyl to isopropyl favourably af-
fected the configurational stability of the residual enantio-
mers.
For all techniques, the measurement at a single tempera-
ture provides one rate constant (we assume that enantiomer-
ization is a first-order process). By measuring the tempera-
ture dependence of the rate constant, the activation enthal-
py and entropy for the enantiomerization process are ob-
tained. To improve the statistical reliability, this convention-
al two-step procedure has been substituted by a one-step
procedure featuring the simultaneous fitting of all variable-
temperature data to an appropriate model and resulting in
the desired thermodynamic parameters. It should be noted
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
99

A. Ponti, F. Sannicolꢁ et al.
that the CD signal is proportional to the enantiomeric
purity, so its exponential decay directly provides the desired
rate constant. Conversely, to extract the rate constant from
the CSP HPLC or NMR profiles one needs to formulate a
model and to fit the latter to the experimental data. This in-
direct method may affect the accuracy of the thermodynam-
ic parameters so obtained.
The most appropriate technique was selected on the basis
of the configurational stability of the compound to be stud-
ied. When the residual enantiomers were stable enough at
room temperature to be obtained in an acceptable purity
level by semipreparative CSP HPLC, the measurement of
CD decay kinetics at different temperatures was chosen as
the most direct and reliable method. The CD measurements
were carried out for 2, 4e, 4 f, and 5 in MeCN/water mix-
tures similar to those used as eluant for the semipreparative
resolution process and for the CSP DHPLC experiments
Figure 5. Variable-temperature dynamic CSP HPLC of 4c (eluant MeCN/
water 2:8, flow 0.7 ccmin ); open circles: experimental; solid line: best-
fit.
ꢀ
1
(
see Experimental Section). The experimental and best-fit
CD decays of 2 and 4e are shown in Figure 4.
When the residual enantiomers were found too labile to
survive at room temperature for at least for a week, the
CSP DHPLC method was considered as the second choice
system for acceding to reliable barrier values. This technique
has been applied to 4a and 4c (Figure 5); it was also ap-
plied to 4e for the sake of comparison, even though com-
plete coalescence could not be attained. The most strict limi-
tation of this technique is the rather narrow temperature
range sustainable by the chiral stationary phase, which un-
dergoes degradation at about 658C. Another point which
deserves consideration, differentiating this technique from
CD and NMR based methods, is the presence of a chiral
enantiopure species, the stationary phase, which could, in
principle, influence the path of the enantiomerization pro-
cess.
[12]
The DNMR technique was applied in the case of sub-
strates displaying methylenic hydrogen atoms, like 4b, 4c,
and 4d, or methyl groups, like 4 f, which are diastereotopic
and magnetically non-equivalent in the minimum energy
states and in the TSs associated with the M mechanism,
1
whereas they are enantiotopic, that is, isochronous, in the
C - or C -symmetric TSs associated with the M process.
3v
s
0
Variable-temperature experiments could supply indica-
tions on the enantiomerization energetics. However, the sol-
vent (generally [D ]DMSO) and the concentration em-
6
1
ployed in H NMR experiments are different from those
used in CSP DHPLC and CD based kinetic work for solu-
bility reasons. This constraint makes questionable the com-
parison of the data supplied by the DNMR spectroscopy
with those obtained from the two other techniques. Further-
more, in some cases the signal pattern and its variations
with temperature were not interpretable in full detail.
The thermodynamic parameters derived from the analysis
of the CD decay and from fitting the CSP HPLC profiles
are reported in Table 3.
In a single case, namely 4e, it was possible to apply all of
the three techniques, thus allowing us to assess their reliabil-
°
ity and consistency. The activation enthalpy DH value is
the same for both techniques within the estimated error,
°
whereas the activation entropy DS is largely different not-
withstanding solvent mixture and temperature range were
very similar.
This discrepancy can be better understood by enlarging
the view to all of the results reported in Table 3. It can be
Figure 4. Variable-temperature CD decays; thin solid line: experimental;
thick dashed line: best-fit. Top: 2 (solvent MeCN:water/15:85). Bottom:
4
e (solvent MeCN:water/28:72). The CD curves have been vertically
shifted for the sake of clarity. Note the different timescale in the two
panels.
100
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

Chiral Compounds
FULL PAPER
Table 3. Experimental activation parameters for M
0
enantiomerization of C
3
-symmetric trisCAHTUNGTRENNUNG
oxides 4a, 4c, 4e, 4 f, 5, and 2 obtained by CD and CSP DHPLC techniques.
the hub to the blades: the
shorter the bond the stronger
°
°
°
System
Method
T
DH
DS
DG (258C)
ꢀ1
ꢀ1
ꢀ1
ꢀ1
[
8C]
[kcalmol
]
[calK mol
]
[kcalmol
]
the engagement and, by conse-
quence, the higher the enantio-
merization barrier. We also
note that our results yield an
activation free energy for 2 at
2
CD
70–85
23–70
15-65
25–59
40–70
50–82
50–80
27ꢂ4
18ꢂ1
ꢀ4ꢂ3
ꢀ20ꢂ4
28ꢂ1
25ꢂ2
[
a]
4
4
4
4
4
5
a
c
e
e
f
CSP DHPLC
CSP DHPLC
CSP DHPLC
CD
CD
CD
19.4ꢂ0.4
22.2ꢂ0.1
22.0ꢂ0.5
23ꢂ2
ꢀ21.6ꢂ0.3
ꢀ23.2ꢂ0.1
ꢀ6ꢂ2
26.0ꢂ0.4
29.1ꢂ0.1
23.8ꢂ0.5
25ꢂ2
ꢀ
1
ꢀ6ꢂ5
708C of 28 kcalmol , in agree-
ment with that reported in liter-
ature for the racemization pro-
14ꢂ1
ꢀ38ꢂ3
25ꢂ1
[
a] From reference [12].
cess carried out in dimethoxy-
ethane (28.4 kcalmol at 708C).
°
ꢀ1
[11]
seen that all DS values are negative, but the values from
ꢀ
1
ꢀ1
CSP DHPLC experiments are about ꢀ20 calK mol ,
Finally, we admit that the activation parameters of 5 have
°
whereas those from CD decay experiments are only about
a peculiar behavior that we are not able to explain. DH is
ꢀ
1
ꢀ1
ꢀ
5 calK mol , except for 5. It is then reasonable to
deduce that the activation entropy DS contains a sort of
technique-related systematic error. Hence, the activation en-
thalpy DH values from CD decay and CSP HPLC experi-
lower than expected on the basis of the geometrical parame-
°
°
ters, while DS is more negative than expected on the basis
°
of the employed technique, and, conversely, DG at 298 K is
°
just as expected on the basis of the relationship to the struc-
ments are consistent with each other, whereas the activation
entropy DS (and then the activation free energy DG )
values are consistent only within a single technique.
turally related compound 2.
We also carried out variable-temperature H NMR experi-
°
°
1
ments even though, as anticipated, these experiments can at
most represent a consistency check of the enantiomerization
barriers obtained from CD decay and CSP HPLC experi-
ments. To investigate the enantiomerization dynamics of 4a
As for dynamic NMR spectra, extensive fitting rounds of
the NMR lineshapes showed that, owing to the limited af-
fordable temperature range and the presence of tempera-
ture-dependent fitting parameters (chemical shifts), the
fitted activation parameters are not accurate and statically
reliable enough for the quantitative determination of the ac-
tivation barriers. In the present case, dynamic NMR spec-
troscpopy can provide qualitative information on the config-
urational stability and an auxiliary consistency check of the
activation barrier values obtained by CD decay or CSP
HPLC. If the temperature-dependent NMR lineshapes can
be satisfactorily fitted by fixing the activation parameters to
those from CD decay or CSP HPLC and leaving the chemi-
cal shifts and J couplings only as free parameters, then the
NMR data are consistent with the previously obtained acti-
vation parameters.
1
2
by H NMR (or H NMR) spectroscopy in addition to the
CSP DHPLC technique, we prepared the tris[3-(2-monodeu-
teromethyl-1-methyl)indolyl]phosphane oxide (4b). We
found, however, that the diastereotopic hydrogen atoms of
the methylene groups of 4b were magnetically isochronous
1
in the H NMR spectra recorded in different solvents even
at very low temperatures.
The methylene protons of the methoxymethylene groups
of 4c gave rise to the AB system typical of two coupled dia-
stereotopic protons at room temperature ([D ]DMSO)
(Figure 6). The activation parameters were fixed to the
values from CSP HPLC, and H NMR lineshapes were sub-
jected to a constrained best-fit procedure. As shown in the
right panel of Figure 6, the lineshape can be accurately
6
1
°
The activation enthalpy DH , in the case of the indolyl
phosphane oxides, grows in the sequence 4a<4c<4e<4 f,
showing that the activation barrier in a series of homologue
compounds in which the ortho substituent is an alkyl group
is mainly affected by the steric hindrance. It is worth noting
fitted to the model and the best-fit spin parameters (w =
A
4.64–4.67 d, w =4.66–4.68 d, J =12.1 Hz) are compatible
B
AB
with indolic methoxymethylene protons.
The decoupled signal of the methylene protons in the
1
that the CH OMe substituent is effectively less hindering
H NMR spectra of 4e ([D ]DMSO) recorded at different
2
6
°
than ethyl (Table 3). The largest DH is shown by 2 despite
temperatures (from 25 to 1508C) does not undergo com-
[12]
the ortho substituent being methyl, probably because the
aryl–hub distance is the shortest of the set (CꢀN bond).
plete coalescence to a singlet even at high temperature.
Again, we fixed the activation parameters to those obtained
°
1
Compound 5, which is similar, has the smallest DH and we
by CSP HPLC or by CD decay and subjected the H NMR
1
will comment later about this somehow puzzling result.
lineshapes to a constrained best-fit procedure. The H NMR
°
On discussing the activation free energy DG at 258C
spectra of 4e prove to be consistent with the activation pa-
rameters derived from both CD decay and CSP HPLC and
the goodness-of-fit measure and the best-fit spin parameters
°
(
Table 3), it is worth recalling that differences in DG are
only meaningful within a single experimental technique. As
°
before, the DG increases as the ortho substituent becomes
(w =2.98–3.02 d, w =2.90–2.97 d, JAB =14.0 Hz, JAX =
A B
more hindering (4a<4c<4e by CSP HPLC; 4e<4 f by CD
decay). Considering the CD results only, the activation bar-
rier increases in the order 4e<4 f~5<2, demonstrating the
great influence exerted on configurational stability by the
7.4 Hz) are only marginally different.
Reduction of phosphane oxide 4e to tris[3-(2-ethyl-1-meth-
yl)indolyl]phosphane (6): Since the chromatographic separa-
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
101

A. Ponti, F. Sannicolꢁ et al.
py, the sole technique applicable on a racemate. The signals
of the methylene protons appear as a sharp quartet, which is
progressively converted into a complex broad multiplet by
lowering the temperature. These observations indicate that,
in the case of phosphane 6, the enantiomerization process is
fast even at room temperature.
There are at least two processes that can be envisaged as
being responsible for the enantiomerization of the residual
antipodes of phosphanes: helix reversal, produced by the M0
mechanism or pyramidal inversion at the phosphorus centre.
The latter mechanism causes enantiomerization, because it
converts, for example, the Msss isomer into the Maaa
isomer, which is one of the stereoisomers constituting the
residual antipode.
This hypothesis looks credible on the basis of the consid-
eration that dramatic changes in geometry on passing from
phosphane oxides to phosphanes are not expected and, by
consequence, also the level of ring engagement and the con-
figurational stability should not be much different in phos-
phane and phosphane oxide. The extensive theoretical and
experimental work reported in literature on the inversion
1
6
Figure 6. H-DNMR experiments on 4c ([D ]DMSO). Left panel: experi-
mental spectra of the methylene protons of the methoxymethylene
groups. Right panel: enlarged view of the external minor peaks; line: ex-
perimental spectra, dots: constrained best-fit spectra.
[21]
tion of residual enantiomeric phosphanes is a problem that
is still waiting for a satisfactory solution, a point deserving
investigation was the reductive conversion of the most
stable phosphane oxides, for example, 4e and 4 f, into phos-
phanes and the evaluation of their configurational stability
in comparison to the parent compounds.
The reduction of phosphane oxide 4e to tris[3-(2-ethyl-1-
methyl)indolyl]phosphane (6) was performed with an excess
of trichlorosilane and triethylamine (TEA) in toluene under
barrier of trivalent phosphorus derivatives suggests, how-
ꢀ
1
ever, that energies much higher than 20 kcalmol are in-
volved in the inversion processes of phosphanes, even
though no specific data is available for tris ACHUTNGRNENGU( aryl)phosphanes.
To have a reliable indication on the more favorable enan-
tiomerization path we resorted to a supplementary computa-
tional investigation of phosphane 6. We computed the acti-
vation barriers for the two processes at the DFT-B3LYP/cc-
pVDZ level, as performed for the corresponding phosphane
oxide 4e. The barrier for the inversion from Msss to Maaa
[20]
an argon atmosphere, according to a known procedure.
ꢀ
1
Lithium aluminum hydride, even if activated with trimethyl-
silyl chloride, showed a much lower reducing activity than
trichlorosilane. Phosphane 6 was obtained in 40% isolated
yield and was characterized on the basis of MS and NMR
spectral data. It was found to be an extremely oxidizable
compound and was easily converted into phosphane oxide
was calculated to be 25.1 kcalmol and that for the M iso-
0
merization from Msss to Psss (and vice versa) amounted to
ꢀ
1
11.7 kcalmol . The computational data thus suggest that
the M process is much faster than inversion and that the
0
enantiomerization of 6 occurs through the M mechanism,
0
just as in the indolic phosphane oxides 4. The calculations
also indicate that the mechanical inter-ring gearing at the
transition state is strongly diminished on passing from phos-
phane oxide 4e to phosphane 6. Indeed, the aromatic moiet-
ies of 6 can be considered planar, deviating by less than 0.58
from planarity, and the distance between the close-contact
hydrogen atoms in the 4-position of the indole moieties is
1.90 ꢃ in 6, which is significantly longer than the 1.79 ꢃ in
4e. It seems that these differences are due to the presence
of the oxygen atom in 4e that pushes the portions of the ar-
omatic rings close to the phosphorus apart, thus causing the
nonplanar deformation of the aromatic rings.
4
e by brief exposure to air even in the solid state.
The lowest reaction temperature necessary to observe an
acceptable reaction rate for the reduction was found to be
about 808C, which was too high to reduce the enantiopure
oxide to the enantiopure phospane without contemporary
racemization. As anticipated, phosphane oxide 4e was
found to be configurationally quite stable at 258C, while
helix inversion was rather fast at 708C. Thus, the reduction
of an enantiopure sample of 4e at 708C unavoidably afford-
ed racemic phosphane 6.
We found phosphane oxide 4 f unreactive in the presence
of lithium aluminum hydride and trichlorosilane, even when
heated to reflux in mesitylene. The reason for this reactivity
drop was attributed to the high hindrance exerted by the
three isopropyl groups surrounding the central phosphorus
atom.
Conclusions
This work reports some results that serve as a guide for de-
signing configurationally stable C -symmetric propeller-
3
Configurational stability of tris[3-(2-ethyl-1-methyl)indolyl]
shaped residual enantiomers. Several members of a series of
molecular propellers with a non-invertible atom as the hub
have been obtained as configurationally stable residual anti-
phosphane (6): The configurational stability of phosphane 6
1
was qualitatively assessed by H NMR dynamic spectrosco-
102
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

Chiral Compounds
FULL PAPER
podes and their enantiomerization energy barriers have
been evaluated by different techniques.
A new steric and electronic design of the blades is re-
quired to accede to C -symmetric stable tris-
3
Even though a reliable comparison of the barriers was
found possible only if the technique employed for their as-
sessment is the same, the results of the research clearly dem-
onstrate that three structural parameters are, in the main,
responsible for the efficacy of the correlated rotation of the
blades and consequently for the configurational stability of
the residual antipodes: 1) the size of the substituents located
in the ortho position on the aromatic rings, 2) the length of
the bond connecting the central group, that is, the propeller
hub, and the aromatic rings, and 3) the aryl-hub-aryl angle
describing the pyramidality of the propeller.
ACHTUNGTRENNUNG( aryl)phosphanes that can be used as mediators in catalysis:
1) the phosphorus center should not be overcrowded by too
bulky ortho substituents, which would preclude complexa-
tion with a metal; and 2) the influence of the electronic
availability at the phosphorus atom on configurational sta-
bility must be investigated and electronically tailored blades
made available.
Experimental Section
The importance of the bulkiness of the ortho substituent
is demonstrated by the experimental racemization energy
barriers DG at 258C, which, according to the CSP HPLC
D
Organic synthesis: [a] were recorded with a Jasco P-1030 polarimeter.
CD spectra and dichroism decay experiments were recorded with a Jasco
P-810 spectropolarimeter provided with a temperature-controlled probe
Peltier Jasco PFD 425S. CSP HPLC analyses and semipreparative separa-
tions were performed on a Waters 600E instrument equipped with a UV
detector Waters 486 and recorded at 220 nm. Analytical column: Chrom-
Tech CHIRAL-AGP column (100ꢄ4.0 mm, 5 mm); loop: 20 mL; substrate
°
experiments, progressively increase along the sequence 4a
°
ꢀ1
°
(
2
R=Me, DG =25 kcalmol ), 4c (R=CH OMe, DG =
2
°
ꢀ
1
ꢀ1
6 kcalmol ), and 4e (R=Et, DG =29 kcalmol ). The
ꢀ
1
concentration: 1 mgmL
;
Semipreparative column: CromTech
series is extendable to 4 f (R=iPr) if the CD decay data of
CHIRAL-AGP column (150ꢄ10.0 mm, 5 mm); loop: 500 mL. NMR spec-
tra were recorded on Bruker AV 400 and Bruker AC 300 spectrometers.
Mass spectra were recorded on Bruker Daltonics high resolution FT-ICR
(Fourier Transform Ion Ciclotron Resonance) model APEXTM II (4.7
Tesla Magnex cryomagnet supplied with ESI source) and Thermofiningan
LCQ Advance (APCI). Purification by column chromatography was per-
formed by using Merck silica gel 60 (230–400 mesh for flash-chromatog-
raphy and 70–230 mesh for gravimetric chromatography). 1,2-Dimethyl-
indole and 1-ethyl-2-methylindole were bought from Sigma Aldrich and
were used as received. 1-Methyl-2-methoxymethylindole was synthesized
by alkylation of the 2-hydroxy-1-methylindole, prepared in turn by reac-
tion of the lithium derivative of 1-methylindole with formaldehyde as de-
scribed below.
°
ꢀ1
°
ꢀ1
4
e (DG =23.8 kcalmol ) and 4 f (DG =25 kcalmol ) are
compared (Table 3). The resolution of the residual race-
mates into antipodes could be achieved at an analytical
level in the case of 4a and 4c, which are configurationally
labile at room temperature, and on a semipreparative scale
in the case of 4e and 4 f.
The relevance of the length of the bond connecting the
central group and the aromatic rings is demonstrated by
comparing the configurational stability of substrates exhibit-
ing closely similar geometry of the blades, bearing identical
ortho substituents (Me groups), but differing in this parame-
2
-Hydroxymethyl-1-methylindole: A 1.6m solution of nBuLi in hexane
ter. In the series of 2 (Csp
Csp
bond length=1.67 ꢃ) and 4a (PꢀC bond length=
.78 ꢃ), configurational stability dramatically decreases
3
ꢀN bond length=1.45 ꢃ), 5 (Pꢀ
sp2
(32 mL, 50 mmol) was added to a solution of N-methylindole (5.0 g,
38 mmol) in THF (40 mL) and the reaction mixture was heated to reflux
for 3 h under a nitrogen atmosphere. After cooling to 08C, gaseous form-
aldehyde, produced by pyrolysis of paraformaldehyde (5.0 g), was bub-
bled through the reaction mixture, which was then stirred at room tem-
perature for 1 h. The solvent was evaporated under reduced pressure,
2
sp2
1
°
ꢀ1
°
from 2 (DG =28 kcalmol , CD decay) to 5 (DG =
ꢀ
1
2
5 kcalmol , CD decay), which is stable at 258C, to 4a,
which undergoes fast racemization at room temperature.
More complex is the quantitative evaluation of the effects
of the aryl-hub-aryl angle. Even though not fully compara-
ble, since obtained on structurally different substrates and
under different experimental conditions, the dramatically
then CH
was separated and dried over Na
dryness to give a residue that was purified by crystallization from diiso-
propyl ether (4.9 g, 80%). H NMR (300 MHz, CDCl
2
Cl
2
(20 mL) and water (20 mL) were added. The organic phase
SO . The solvent was evaporated to
2
4
1
3
3
): d=7.60 (d, J-
3
A
H
U
G
R
N
U
G
A
H
U
G
R
N
U
G
1H), 6.48 (s, 1H), 4.83 (s, 2H), 3.83 ppm (s, 3H).
°
different stereomerization barriers for 3 (117.38, DG =
1-Methyl-2-methoxymethylindole:
(2.7 g, 17 mmol) was added to a 60% suspension of NaH (0.9 g,
38 mmol) in MeCN (50 mL). After stirring for 1 h at 508C, CH I (1.6 mL,
3
2-Hydroxymethyl-1-methylindole
ꢀ
1
1
1
7.7 kcalmol at ꢀ218C, determined by H NMR experi-
[8]
°
ments)
3
and 4a–4 f (about 106–1078, DG =20–
2
6 mmol) was added and the reaction mixture heated at 308C for 2 h. A
second portion of CH I (1.6 mL, 26 mmol) was added and the reaction
mixture stirred for 12 h. The solvent was evaporated to dryness and the
residue was subjected to column chromatography (silica gel, CH Cl ) to
give 1-methyl-2-methoxymethylindole (1.14 g, 39%). M.p. 658C;
ꢀ
1
0 kcalmol
at 258C) demonstrate that this parameter
3
strongly influences the correlated rotation of the rings.
The nature of the atom constituting the propeller hub
probably exerts the most important effect on the stability of
residual enantiomeric propellers, since it controls all the
three parameters discussed above.
2
2
1
3
4
H NMR (400 MHz, CDCl
3
): d=7.60 (dt,
J
A
H
U
G
R
N
U
G
J ACHTUNGTRENNUNG( H,H)=
3
4
0
.85 Hz, 1H), 7.33 (dd, J ACHTUNGTREUNN(GN H,H)=8.0 Hz, J ACHTUTGNENRNUG
(
An unexpected difficulty in designing stable residual en-
antiomeric phosphanes, similarly to amines, was found to be
related to the lack of an apical group on the central atom
above the reference plane. Such an apical group seems to
(s, 1H).
Tris[3-(2-methoxymethyl-1-methyl)indolyl]phosphane oxide (4c): A solu-
tion of 2-methoxymethyl-1-methylindole (1.1 g, 6.3 mmol) in pyridine
(
2 mL) was added to a mixture of POBr
2 mL) under a nitrogen atmosphere. After stirring overnight at 908C,
Cl was added and the organic layer was washed with a 1N HCl solu-
tion, dried (Na SO ), and concentrated under reduced pressure. Purifica-
3
(0.56 g, 2 mmol) and pyridine
(
play a very important role on the M mechanism barrier;
0
CH
2
2
phosphorus pyramidal inversion seems a less probable race-
mization path.
2
4
tion of the residue by chromatography (silica gel, CH
2
Cl
2
/MeOH 9.5:0.5)
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
103

A. Ponti, F. Sannicolꢁ et al.
1
31
yielded 4c (0.65 g, 20%). M.p. 1828C; H NMR (300 MHz, CDCl
3
): d=
(s); P NMR(300 MHz, CDCl
for C30
3
): d=5.76 ppm (s); MS (APCI): m/z calcd
OP: 716; found: 717.8 [M+1] .
3
3
+
7
.34 (d, J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=8.46 Hz, 1H), 7.14 (t, J
A
H
N
T
E
N
N
H
27Br
3
N
3
3
3
2
J ACHTUNGTRENNUNG( H,H)=8.09 Hz, 1H), 6.77 (t, J ACHTUNGTRENNUNG
Tris[3-(2-deuteromethyl-1-methyl)indolyl]phosphane oxide (4b): A solu-
tion of 4g (0.10 g, 0.14 mmol) in THF (10 mL) was added to a suspension
of LiAlD
0 min at 258C. A 32% NaOH solution (0.1 mL) was added and the inor-
ganic salts were removed by filtration. The filtrate was concentrated
under reduced pressure and the residue was purified by chromatography
2
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=12.8 Hz, 1H), 4.81 (d,
JACHTUNGTRENNUNG
1
3
CH ), 3.02 ppm (s, 3H; -O-CH
3
3
); C NMR (300 MHz, CDCl
J AHCTUNGTRENNGNU( C,P)=11.7 Hz), 128.54 (d, J ACHTNUGTRENNUNG( C,P)=
1.7 Hz), 122.87 (s), 121.87 (s), 121.31 (s), 109.77 (s), 108.62 (d, JACHTGNUTRENNUNG
3
): d=143.58
4
(0.010 g) in THF (1 mL), and the reaction mixture stirred for
2
3
2
(
d,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(C,P)=18.7 Hz), 138.22 (d,
3
1
1
1
3
1
CDCl
3
36 3
H N
(
silica gel, CH
H NMR(CDCl
2
Cl
2
/MeOH 9.5:0.5) to yield 4b (quantitative yield);
+
1
3
3
, 300 MHz): d=7.29 (d, J ACHTUNGRNETNUN(G H,H)=7.49 Hz, 1H), 7.08 (t,
ꢀ
1
eluant MeCN/H
Tris[3-(1-ethyl-2-methyl)indolyl]phosphane oxide (4d):
POBr (1.6 g, 5.5 mmol) and pyridine (2 mL) was added to a solution of
-ethyl-2-methylindole (3.7 g, 23 mmol) in pyridine (2 mL) under nitro-
2
O 20:80, flow rate 0.7 mLmin
.
3
3
3
J
A
H
N
T
E
N
N
J AHCTUNGRETNNUN(G H,H)=7.76 Hz, 1H), 6.65 (d, J-
4
A
mixture of
A
H
U
G
R
N
N
JACHTUNGTRENNUNG
3
1
3
2H);
C H D N PO: 482; found: 483.1 [M+1] .
+
1
3
0
27
3
3
gen. The reaction mixture was heated at 608C under stirring for 5 h, then
diluted with dichloromethane and the resulting solution exhaustively ex-
tracted with a solution of 1N HCl. The organic layer was dried over
Analytical HPLC resolution of tris[3-(2-ethyl-1-methyl)indolyl]phos-
phane oxide (4e): (+)-4e, 97.5:2.5 e.r, retention time 3.6 min; (ꢀ)-4e,
9
0
5:5 e.r., retention time 6.8 min; eluant MeCN/H
.6 mLmin ; semipreparative HPLC: (+)-4e, retention time 4.0 min;
2
O 30:70; flow rate
Na
2
SO
4
and evaporated to dryness to give a residue, which was subjected
Cl /MeOH 9.5:0.5). The first product
ꢀ1
to chromatography (silica gel, CH
2
2
(
4
ꢀ)-4e, retention time 6.8 min; eluant MeCN/H
2
O 28:72; flow rate
eluted was some unreacted 1-ethyl-2-methylindole; the second product
eluted was the tris[3-(1-ethyl-2-methyl)indolyl]phosphane oxide (4d)
ꢀ1
mLmin
(2-methyl)benzimidazolyl]phosphane oxide (5):
(2.2 g, 7.6 mmol) in pyridine (15 mL) was rapidly added to a mix-
.
1
3
Tris[1-
POBr
3
A
H
U
G
R
N
N
A solution of
(
2.6 g, 30%). M.p. 2228C; H NMR (300 MHz, CDCl
3
): d=7.31 (d, J-
3
4
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=8.09 Hz, 1H), 7.06 (td, J
A
H
U
G
R
N
N
ACHTUNGTREN(NUGN H,H)=1.1 Hz, 1H),
3
4
3
ture of 2-methylbenzimidazole (3.1 g, 23.5 mmol) in pyridine (30 mL)
under a nitrogen atmosphere. After 5 h stirring at 808C, most of the sol-
6
8
3
.70 (td,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.53 Hz,
J
A
T
N
T
E
N
N
JACHTUNGTRENNUNG
3
4
.09 Hz, 1H), 4.19 (q,
H), 1.37 ppm (t,
J ACHUTNGRENUNG( H,H)=7.17 Hz, 2H), 2.57 (d, J AHCTUNGTRENNUNG
3
13
vent was removed under reduced pressure. CHCl
the residue and the organic layer were washed with a 5% HCl solution,
dried (MgSO ), and concentrated under reduced pressure to yield pure 5
(2.0 g, 60%) as a white solid. M.p. 190–1928C; H NMR (300 MHz): d=
3
(25 mL) was added to
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
3
2
3
d=145.17 (d,
J ACHTUNGTRENNUNG( C,P)=18.2 Hz), 136.55 (d, J AHCTUNGTRENNUNG
2
4
J
ACHTUNGTRENNUNG
1
1
3
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(C,P)=132.8 Hz),
38.21
(s),
15.46
(s),
12.47 ppm
(s);
3
3
3
1
7.8 (d, J
ACTHUNGTRENNGNU( H,H)=7.89 Hz, 1H), 7.32 (t, J CAHTUNGTRENNNUG
P NMR(300 MHz, CDCl
OP: 521; found: 522.2 [M+1] .
Bis[3-(1-methyl-2-ethyl)indolyl]-[3-(1-methyl)indolyl]phosphane
4h): A solution of 2-ethyl-1-methylindole (8.8 g, 56 mmol) and 1-methyl-
indole (2.9 g, 22 mmol) in pyridine (6 mL) was added to a mixture of
POBr (6.0 g, 21 mmol) and pyridine (6 mL) under a nitrogen atmos-
phere. After stirring 12 h at 908C, CH Cl was added and the organic
layer was washed with a solution of 1N HCl, dried (Na SO ), and con-
centrated under reduced pressure. The residue was purified by column
chromatography (silica gel, CH Cl /MeOH 9.5:0.5). The first fraction
3
): d=7.98 ppm (s); MS (APCI): m/z calcd for
3
+
A
H
U
G
R
N
N
33 36 3
C H N
1
3
C NMR (400 MHz, CDCl
s), 125.74 (s), 121.10 (s), 111.98 (s), 17.49 ppm (s); P NMR(300 MHz,
CDCl PO: 440;
): d=ꢀ15 ppm (s); MS (ESI): m/z calcd for C24
found: 441.16 [M+H] ; analytical HPLC: (+)-5 retention time 17.6 min;
3
oxide
31
(
(
3
21 6
H N
+
3
(
ꢀ)-5 retention time 25.8 min; eluant MeCN/H
2
O 9:91; flow rate
2
2
ꢀ1
0
5
4
.8 mLmin ; semipreparative HPLC: (+)-5, retention time 9.4 min; (ꢀ)-
2
4
,
retention time 11.0 min; eluant MeCN/H
2
O
15:85; flow rate
ꢀ1
mLmin
.
2
2
Tris[3-(2-ethyl-1-methyl)indolyl]phosphane (6): All the manipulations
were performed under argon by using standard Schlenck techniques.
HSiCl (0.9 mL) was added to a solution of 6 (0.20 g, 0.38 mmol) and
3
TEA (1.0 mL) in degassed toluene (5 mL) and the resulting suspension
stirred at 808C under argon for 5 h. A 10% NaOH solution (1 mL) was
added at 08C to the reaction mixture, and the organic layer was separat-
ed, dried, and concentrated under reduced pressure to give a residue,
eluted gave some unreacted 2-ethyl-1-methylindole (5.71 g, 35.9 mmol).
The second fraction yielded the tris[3-(1-methyl-2-ethyl)indolyl]phos-
phane oxide (4e) (0.60 g, 1.15 mmol). The last fraction eluted gave the
bis[3-(2-ethyl-1-methyl)indolyl]-[3-(1-methyl)indolyl]phosphane
oxide
1
(
7
4h) (1.40 g, 2.83 mmol). M.p. 1718C ; H NMR (300 MHz, CDCl
3
): d=
3
3
.81 (d, J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.97 Hz, 1H), 7.28 (m, 4H), 7.07 (m, 4H), 6.89 (d, J-
3
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.92 Hz, 2H), 6.78 (t,
J
A
H
U
G
R
N
N
(H,H)=7.51 Hz, 2H), 3.72 (s, 9H), 3.29
3
2
3
which was subjected to chromatography (silica gel, CH
2
Cl
2
/AcOEt 8:2).
(
sextet,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.34 Hz,
J AHCTUNGTERNNUN(G H,H)=15.87 Hz, 2H), 3.13 (sextet, J-
1
2
3
The first fraction eluted yielded 6 (0.080 g, 40%); H NMR (300 MHz):
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.34 Hz,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=15.87 Hz, 2H), 1.07 ppm (t,
J AHCTUNGTRENNUN(G H,H)=
3
1
3
2
d=7.23 (d,
J AHCTUNGTERNNNU(G H,H)=8.2 Hz, 1H), 6.98 (m, 2H), 6.68 (m, 1H), 3.68 (s,
7
1
1
.44 Hz, 6H); C NMR (300 MHz, CDCl ): d=150.46 (d,
3
JACHTUNGTRENNUNG
3
3
2
3
3H), 2.97 (q, J ACHTNUGTRENNUN(G H,H)=7.1 Hz, 2H), 0.96 ppm (t, J ACHTUNTGRENNUNG
C NMR (300 MHz, CDCl ): d=147.04 (d, JACHNTUGTRENUNGN
3
8.82 Hz), 138.05 (d,
J ACHTUNGTRENNUNG( C,P)=10.17 Hz), 137.32 (d, J ACHTUNGTRENNUNG
1
3
2
2
2
2
35.32 (d, J ACHTUNGTRENNUNG( C,P)=21.37 Hz), 129.58 (d, J ACHTUGNTRENNUN(G C,P)=9.66 Hz), 128.66 (d, J-
(
ACHTUNGTRENNUNG( C,P)=12.21 Hz), 122.40 (s), 122.33 (s), 120.99 (s), 120.74 (s), 120.63 (s),
3
31
1
29.61 (s), 19.08 (d,
J
A
H
U
G
R
N
U
G
P NMR
1
20.28 (s), 110.32 (d, J ACHTUNGTRENNUNG( C,P)=128.69 Hz), 109.34 (s), 108.84 (s), 104.23
1
(300 MHz, CDCl ): d=ꢀ79.77 ppm (s); MS (EI): m/z calcd for
(
d,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(C,P)=132.75 Hz), 33.03 (s), 29.43 (s), 19.11 (s), 13.77 ppm (s);
P NMR(300 MHz, CDCl ): d=7.24 ppm (s); MS (APCI): m/z calcd for
PO: 494; found: 495.2 [M+1] .
Tris[3-(2-bromomethyl-1-methyl)indolyl]phosphane oxide (4g): A solu-
tion of BBr (1m, 0.53 mmol) in CH Cl (0.5 mL) was added to a solution
of 4c (0.10 g, 0.18 mmol) in CH Cl . The reaction mixture was stirred for
h and then diluted with water. The organic layer was separated, dried
Na SO ), and concentrated under reduced pressure to give 4g, which,
due to its instability, was employed in a crude state without further purifi-
3
+
3
1
C H N P: 505; found: 505 [M] .
33 36
3
3
+
[11]
C
31
H
32
N
3
Tris[1-
A
H
U
G
R
N
U
G
Analytical CSP HPLC:
(
7
1
+)-2, 89% ee, retention time 5.8 min; (ꢀ)-2, 87% ee, retention time
.5 min; eluant: MeCN/aqueous sodium acetate buffer 10 mm, pH 4.50–
3
2
2
ꢀ
1
0/90; flow rate 0.8 mLmin ; semipreparative CSP HPLC: (+)-2, reten-
2
2
tion time 7.2 min; (ꢀ)-2, retention time 8.8 min; eluant MeCN/aqueous
3
(
ꢀ1
sodium acetate buffer 15:85, 10 mm, pH 4.50; flow rate 4 mLmin . The
2
4
pH of the solutions of the two residual enantiomers, collected in separat-
ed vials, were adjusted to neutrality with NaHCO , then CHCl and NaCl
3 3
were added and the organic layer separated, dried on Na SO and the
2 4
solvent removed under reduced pressure, giving yellowish residues of 2
1
3
cation (0.12 g; 95%); H NMR (200 MHz, CDCl
3
): d=7.36 (d, J
A
H
U
G
R
N
U
G
3
3
8
.33 Hz, 1H), 7.20 (t,
J
A
H
U
G
E
N
N
(H,H)=7.6 Hz, 1H), 6.97 (d,
J
A
H
U
G
R
N
U
G
3
1
H), 6.82 (t,
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(H,H)=7.6 Hz, 1H), 4.97 (s, 2H), 3.84 ppm (s, 3H);
C NMR (300 MHz, CDCl
(C,P)=11.7 Hz), 128.36 (d,
21.90 (s), 110.05 (s), 107.13 (d, J
1
3
3
2
(95% purity, HPLC).
3
): d=142.09 (d, J
A
H
U
G
R
N
U
G
2
J
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
J
A
H
U
G
R
N
U
G
Crystallographic data: CCDC-695918 (4h), 695919 (4c) and 695920 (5)
contain the supplementary crystallographic data for this paper. These
1
1
A
H
U
T
N
U
G
104
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 94 – 105

Chiral Compounds
FULL PAPER
data can be obtained free of charge from The Cambridge Crystallograph-
ic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
cal atropisomers, exhibiting a stereogenic axis located on the inter-
section of the plane of the labeled aromatic ring with the plane of
the double bond: S. E. Biali, Z. Rappoport, A. Mannschreck, N.
Pustet, Angew. Chem. 1989, 101, 202; Angew. Chem. Int. Ed. Engl.
1
989, 28, 199; E. Rochlin, Z. Rappoport, F. Kastner, N. Pustet, A.
Acknowledgements
Mannschreck, J. Org. Chem. 1999, 64, 8840.
10] B. Laleu, G. Bernardinelli, R. Chauvin, J. Lacour, J. Org. Chem.
2006, 71, 7412.
[11] C. Foces-Foces, F. H. Cano, M. Martinez-Ripoll, R. Faure, C. Rous-
sel, R. M. Claramunt, C. Lopez, D. Sanz, J. Elguero, Tetrahedron:
Asymmetry 1990, 1, 61.
[12] T. Benincori, G. Celentano, T. Pilati, A. Ponti, S. Rizzo, F. Sannicolꢁ,
Angew. Chem. 2006, 118, 6339; Angew. Chem. Int. Ed. 2006, 45,
6193.,
[
This work was supported by MIUR: FIRB-RBNE033 KMA “Composti
molecolari e materiali ibridi nanostrutturati con proprietꢀ ottiche riso-
nanti e non risonanti per dispositivi fotonici”. T.B. thanks the Diparti-
mento di Chimica Organica e Industriale of the University of Milan for
hospitality. We thank Prof. F. Bonomi and Prof. S. Iametti (DISMA, Uni-
versity of Milan) for assistance and helpful suggestions in CD spectra re-
cording experiments.
[
[
13] R. Laser, J. F. Blount, K. Mislow, J. Am. Chem. Soc. 1980, 102, 2777.
14] J. Albert, R. Bosque, J. M. Cadena, S. Delgado, J. Granell, G.
Muller, J. I. Ordinas, M. F. Bardia, X. Solans, Chem. Eur. J. 2002, 8,
[
1] “Chiral monodentate phosphorus ligands”: C. Bruneau, J. L.
Renaud in Phosphorus Ligands in Asymmetric Catalysis (Ed.: A.
Bçrner), Wiley-VCH, Weinheim 2008, pp. 5–35, and references
herein.
2] M. R. Whitnall, K. K. Hii, M. Thornton-Pett, T. P. Kee, J. Organo-
met. Chem. 1997, 529, 35; M. T. Powell, A. M. Porte, K. Burgess,
M. T. Powell, Chem. Commun. 1998, 2161; A. M. Porte, J. Reibens-
pies, K. Burgess, Tetrahedron 2001, 57, 5027; P. Wyatt, H. Eley, J.
Charmant, B. J. Daniel, A. Kantacha, Eur. J. Org. Chem. 2003, 4216;
F. Sannicolꢁ, T. Benincori, S. Rizzo, Triarilfosfine chirali a simmetria
2
279; L. Horner, H. Winkler, A. Rapp, A. Mentrup, H. Hoffmann,
P. Beck, Tetrahedron Lett. 1961, 2, 161; S. Otsuka, A. Nakamura, T.
Kano, K. Tani, J. Am. Chem. Soc. 1971, 93, 4301; W. S. Knowles,
M. J. Sabacky, B. D. Vineyard, J. Chem. Soc. Chem. Commun. 1972,
[
1
1; K. Tani, L. D. Brown, J. Ahmed, J. A. Ibers, M. Yokota, A. Na-
kamura, S. Otsuka, J. Am. Chem. Soc. 1977, 99, 7876; B. Bodganov-
ic, B. Henc, A. Lçsler, B. Meister, H. Pauling, G. Wilke, Angew.
Chem. 1973, 85, 1013; Angew. Chem. Int. Ed. Engl. 1973, 12, 954; T.
Yoshikuni, J. C. Bailar Jr. , Inorg. Chem. 1982, 21, 2129; J. Albert,
J. M. Cadena, J. Granell, G. Muller, J. I. Ordinas, D. Panyella, C.
Puerta, C. SaÇudo, P. Valerga, Organometallics 1999, 18, 3511; V. V.
Dunina, L. G. Kuz’mina, M. Y. Rubina, Y. K. Grishin, Y. Veit, E. I.
Kazakova, Tetrahedron: Asymmetry 1999, 10, 1483; P. Antognazza,
T. Benincori, S. Mazzoli, F. Sannicolo, T. Pilati, Phosphorus, Sulfur
Silicon Relat. Elem. 1999, 144, 405; J. Albert, R. Bosque, J. M.
Cadena, J. Granell, G. Muller, J. I. Ordinas, Tetrahedron: Asymmetry
3
C come leganti di metalli per la preparazione di catalizzatori per re-
azioni stereoselettive, Italian Patent MI2005A 002029; ꢅ. Pintꢆr, G.
Haberhauer, I. Hyla-Kryspin, S. Gromme, Chem. Commun. 2007,
3
711.
[
[
3] M. Ciclosi, J. Lloret, F. Estevan, P. Lahuerta, M. Sanau’, J. P. Prieto,
Angew. Chem. 2006, 118, 6893; Angew. Chem. Int. Ed. 2006, 45,
6
741.
2
000, 11, 3335; T. Benincori, E. Cesarotti, O. Piccolo, F. Sannicolꢁ, J.
4] P. Finocchiaro, D. Gust, K. Mislow, J. Am. Chem. Soc. 1973, 95,
Org. Chem. 2000, 65, 2043; T. Benincori, R. Cirilli, R. Ferretti; F. La
Torre, O. Piccolo, L. Zanitti, Chromatographia 2002, 55, 25; F. La
Torre, O. Piccolo, L. Zanitti, Chromatographia 2002, 55, 25.
8
9
1
172; P. Finocchiaro, D. Gust, K. Mislow, J. Am. Chem. Soc. 1974,
6, 3198; P. Finocchiaro, D. Gust, K. Mislow, J. Am. Chem. Soc.
974, 96, 3205.
[
15] Symbol a indicates that the prioritary edge (Cahn–Ingold—Prelog
[
5] W. D. Hounshell, C. A. Johnson, A. Guenzi, F. Cozzi, K. Mislow,
Proc. Natl. Acad. Sci. USA 1980, 77, 6961; F. Cozzi, A. Guenzi,
C. A. Johnson, K. Mislow, W. D. Hounshell, J. F. Blount, J. Am.
Chem. Soc. 1981, 103, 957; C. A. Johnson, A. Guenzi, K. Mislow, J.
Am. Chem. Soc. 1981, 103, 6240; C. A. Johnson, A. Guenzi, R. B.
Nachbar, Jr., J. F. Blount, O. Wennerstrçm, K. Mislow, J. Am. Chem.
Soc. 1982, 104, 5163; H.-B. Bꢇrgi, W. D. Hounshell, R. B. Nach-
bar, Jr., K. Mislow, J. Am. Chem. Soc. 1983, 105, 1427; A. Guenzi,
C. A. Johnson, F. Cozzi, K. Mislow, J. Am. Chem. Soc. 1983, 105,
(
CIP) rules) of a ring is anti to the PꢀO bond (or to the lone pair)
and located below the reference plane (the plane defined by the
three carbon atoms connected to the P centre); s indicates a syn re-
lationship. Alternative descriptors for propellers have been suggest-
ed; however, they require the introduction of symbols not familiar
to the stereochemical language; R. A. Baber, A. G. Orpen, P. G.
Pringle, M. J. Wilkinson, R. L. Wingad, J. Chem. Soc. Dalton Trans.
2
005, 659, and reference therein.
[
[
16] J. Tomasi, M. Persico, Chem. Rev. 1994, 94, 2027.
17] For instance, one may recall that the rotation barrier in ethane is
1
438; H. Iwamura, K. Mislow, Acc. Chem. Res. 1988, 21, 175.
[
6] Y. Kawada, H. Iwamura, J. Org. Chem. 1980, 45, 2547; Y. Kawada,
H. Iwamura, J. Am. Chem. Soc. 1981, 103, 958; Y. Kawada, H. Iwa-
mura, Tetrahedron Lett. 1981, 22, 1533; Y. Kawada, H. Iwamura, J.
Am. Chem. Soc. 1983, 105, 1449; Y. Kawada, H. Iwamura, Y. Oka-
moto, H. Yuki, Tetrahedron Lett. 1983, 24, 7910; N. Koga, Y.
Kawada, H. Iwamura, J. Am. Chem. Soc. 1983, 105, 5498; Y.
Kawada, Y. Okamoto, H. Iwamura Tetrahedron Lett. 1983, 24, 5359;
H. Iwamura, T. Ito, K. Toriumi, Y. Kawada, E. Osawa, T. Fujiyoshi,
C. Jaime, J. Am. Chem. Soc. 1984, 106, 4712; N. Koga, H. Iwamura,
J. Am. Chem. Soc. 1986, 108, 1426; H. Iwamura, J. Mol. Struct. 1985,
ꢀ
1
ꢁ
3 kcalmol ; N. Moazzen-Ahmadi, H. P. Gush, M. Halpern, H. Ja-
gannath, A. Leung, I. Ozier, J. Chem. Phys. 1988, 88, 563.
18] P. K. Mandal, E. Arunan, J. Chem. Phys. 2001, 114, 3880.
19] D. M. P. Mingos, T. E. Mꢇller, J. Organomet. Chem. 1995, 500, 251;
T. E. Mꢇller, J. C. Green, D. M. P. Mingos, C. M. McPartlin, C. Whit-
tingham, D. J. Williams T. M. Woodroffe, J. Organomet. Chem. 1998,
[
[
5
51, 313; S. Sasaki, K. Sutoh, F. Murakami, M. Yoshifuji, J. Am.
Chem. Soc. 2002, 124, 14830; S. Sasaki, R. Chowdhury, M. Yoshifuji,
Tetrahedron Lett. 2004, 45, 9193.
[
[
20] K. Naumann, G. Zon, K. Mislow, J. Am. Chem. Soc. 1969, 91, 7012.
21] H. A. Bent, Chem. Rev. 1961, 61, 275; R. D. Baechler, W. B. Farn-
ham, K. Mislow, J. Am. Chem. Soc. 1969, 91, 5686; R. D. Baechler,
K. Mislow, J. Am. Chem. Soc. 1970, 92, 3290; A. Rauk, L. C. Allen,
K. Mislow, Angew. Chem. 1970, 82, 453; Angew. Chem. Int. Ed.
Engl. 1970, 9, 400; R. D. Baechler, K. Mislow, J. Am. Chem. Soc.
1
26, 401; N. Koga, Y. Kawada, H. Iwamura Tetrahedron 1986, 42,
1
679; Y. Kawada, H. Yamazaki, G. Koga, S. Murata, H. Iwamura, J.
Org. Chem. 1986, 51, 1472; N. Koga, H. Iwamura, Chem. Lett. 1986,
47; Y. Kawada, J. Ishikawa, H. Yamazaki, G. Koga, S. Murata, H.
2
Iwamura, Tetrahedron Lett. 1987, 28, 445.
7] K. Mislow, Acc. Chem. Res. 1976, 9, 26.
8] R. Glaser, J. F. Blount, K. Mislow, J. Am. Chem. Soc. 1980, 102,
[
[
1
971, 93, 773; R. D. Baechler, J. D. Andose, J. Stackhouse, K.
2
777.
9] The first resolution of the racemate of residual enantiomeric propel-
lers was claimed for some asymmetric tris(aryl)ethylenes; however,
this cannot be considered as residual, since these systems are classi-
Mislow, J. Am. Chem. Soc. 1972, 94, 8060; E. Magnusson, Aust. J.
Chem. 1985, 38, 23.
[
AHCTUNGTRENNUNG
Received: October 31, 2008
Chem. Eur. J. 2009, 15, 94 – 105
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
105