Chiral 4-(diphenylphosphanyl)-1-(dialkylamino)butane ligands - Synthesis, applications in asymmetric alkylation and theoretical study

Article abstract of DOI:10.1002/(SICI)1099-0682(200002)2000:2<351::AID-EJIC351>3.0.CO;2-Q

Various 4-(diphenylphosphanyl)-1-(dialkylamino)butane ligands were prepared from commercial (2S,3S)-2,3-O-isopropylidene-d-threitol. These ligands, associated with Pd₂(dba)₃, gave enantioselectivities of up to 75% in the alkylation o

Full text of DOI:10.1002/(SICI)1099-0682(200002)2000:2<351::AID-EJIC351>3.0.CO;2-Q

FULL PAPER
Chiral 4-(Diphenylphosphanyl)-1-(dialkylamino)butane Ligands –
Synthesis, Applications in Asymmetric Alkylation and Theoretical Study
Fabien Robert,[a] Francoise Delbecq,[b] Christelle Nguefack,[a] and Denis Sinou*[a]
¸
Keywords: Aminophosphanes / Palladium / Alkylations / Enantioselectivity / Theoretical study
Various
ligands were prepared from commercial (2S,3S)-2,3-O-
isopropylidene-
Pd2(dba)3, gave enantioselectivities of up to 75% in the
4-(diphenylphosphanyl)-1-(dialkylamino)butane
important role played by two factors in the alkylation
reaction: steric control, leading to the formation of a unique
d-threitol. These ligands, associated with diastereoisomer of the π-allylpalladium complex, and orbital
control, which orients the attack of the nucleophile trans to
alkylation of racemic 1,3-diphenylprop-2-enyl acetate with the phosphorus atom.
dimethyl malonate anion. A theoretical study shows the
Introduction
Results and Discussion
Synthesis of the P,N-Based Ligands
Asymmetric metal catalysis is one of the most powerful
methods for the preparation of enantiomerically pure com-
pounds.[1] An impressive number of chiral ligands have
been prepared and used, among them P,P-, P,N- and N,N-
based ligands. The phosphanyloxazolines, P,N-based li-
gands, induce high enantioselectivities, particularly in the
palladium-catalyzed allylic substitution.[2] In comparison,
P,N ligands with an sp3 nitrogen atom have been used less
in asymmetric catalysis. Hayashi et al. prepared chiral β-
(aminoalkyl)phosphanes[3] and [(aminoalkyl)ferrocenyl]-
phosphanes.[4] Chiral aminophosphanes, in which the chir-
ality is on the substituents of the nitrogen atom through a
pyrrolidine[5Ϫ6] or a naphthyl unit,[7] have also been de-
scribed more recently. These ligands were successfully used
in asymmetric hydrogenation, cross-coupling reactions and
mainly allylic substitution.
We recently reported, in a preliminary communication,
the synthesis of chiral aminophosphanes derived from tar-
taric acid and their use in asymmetric alkylation.[8] We de-
scribe in this paper a full account of the synthesis of these
new chiral ligands and their use as ligands in asymmetric
alkylation, as well as a theoretical study concerning the nu-
cleophilic attack on π-allyl(palladium) complexes associated
with these ligands.
Ligands 3aϪd, bearing at least one aryl group at the ni-
trogen atom, were prepared from commercially readily
available bis(tosylate) 1 (Scheme 1). Reaction of 1 at Ϫ17°C
in tetrahydrofuran with one equiv. of lithium diphenylam-
ide, obtained by treating diphenylamine with n-butyllithium
at Ϫ78°C, gave diphenylamino tosylate 2a in 26% yield.
Subsequent treatment of compound 2a with lithium di-
phenylphosphide in tetrahydrofuran led to the formation of
the diphenylamino phosphane 3a in 40% yield. Steric and
electronic modifications of the substituents at the nitrogen
atom were introduced with methyl(phenyl)-, p-methoxyphe-
nyl(methyl)- and phenyl(α-naphthyl)amine instead of di-
phenylamine; the corresponding amino phosphanes 3b, 3c,
and 3d were obtained in 18%, 7%, and 21% yield, respec-
tively, from compound 1. All attempts to introduce the p-
nitrophenyl(methyl)amino group failed, only the elimin-
ation product was observed.
Scheme 1. a: LiNR1R2, THF, Ϫ17 °C; b: LiPPh2, THF, 25 °C
Reaction of monotosylate 4[9] with sodium azide in DMF
gave the azido alcohol 5 in 95% yield (Scheme 2). Reduction
of compound 5 with molecular hydrogen in the presence of
20% Pd/C led quantitatively to the amino alcohol 6. Con-
densation of this amino alcohol 6 with two equiv. of benzyl
bromide, 2,4,6-trimethylbenzyl bromide, 2-naphtylmethyl
bromide or 2,2Ј-bis(bromomethyl)biphenyl in acetonitrile in
the presence of potassium carbonate gave the bis(alkylam-
ino) alcohols 7aϪd in 58Ϫ85% yield. Tosylation of these
[a]
`
´
´
Laboratoire de Synthese Asymetrique, associe au CNRS,
´
CPE Lyon, Universite Claude Bernard Lyon 1,
43, bd du 11 novembre 1918, F-69622 Villeurbanne Cedex,
France
Fax: (internat.) ϩ 33-4/72448160
E-mail: sinou@univ-lyon1.fr
[b]
Institut de Recherches sur la Catalyse,
´
´
Groupe Theorie et Modelisation,
2, avenue Einstein, F-69626 Villeurbanne Cedex, France
Fax: (internat.) ϩ 33-4/72445399
E-mail: delbecq@catalyse.univ-lyon1.fr
Eur. J. Inorg. Chem. 2000, 351Ϫ358
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ1948/00/0202Ϫ351 $ 17.50ϩ.50/0
351

F. Robert, F. Delbecq, C. Nguefack, D. Sinou
FULL PAPER
Scheme 2. a: NaN3, DMF, 24 h, 95%; b: H2, Pd/C, EtOH, 48 h, 100%; c: RCH2Br (2 equiv.), K2CO3, CH3CN, 30 min; d: LiPPh2,
THF, 25 °C
bis(alkylamino) alcohols 7 led to the bis(alkylamino) tosyl- tion. The substituted product thus obtained was isolated by
ates 8aϪd, whose treatment by lithium diphenylphosphide chromatography. The results are summarized in Table 1.
gave the desired bis(alkylamino) phosphanes 9aϪd in 40%,
45%, 20%, and 80% yield, respectively.
Table 1. Allylic alkylation of (±)-(E)-1,3-diphenylpropenyl acetate
with dimethyl malonate, catalyzed by palladium complexes[a]
Reaction of amino alcohol 6 with two equiv. of tosyl
chloride in pyridine gave the bis(tosylate) 10 in 65% yield Entry Ligand Solvent
(Scheme 3); tosylamino diphenylphosphane 11 was ob-
Conversion ee
(%)[b]
(%)[b]
tained in 73% yield by subsequent treatment of 10 with
1
DIOP
12
3a
3b
3c
THF
THF
THF
THF
THF
THF
THF[c]
95
90
97
94
96
93
47
94
95
4
0
more than two equiv. of lithium diphenylphosphide. Finally,
the bis(diphenyl)amino ligand 12 was obtained as a bypro-
duct in the preparation of compound 2a.
2
0
3
50
68
63
69
30
67
68
8
10
62
75
57
48
39
4
5
6
3d
3d
3d
3d
3d
3d
9a
9b
9c
7
[d]
8
CH2Cl2[c]
9
CH2Cl2
10
11
12
13
14
15
16
Toluene[c]
Toluene/CH2Cl2 (4:1)[d] 12
Scheme 3. a: TsCl, C5H5N, 25 °C, 65%; b: LiPPh2, THF, 25 °C
THF
THF
THF
THF
THF
94
97
99
99
49
9d
11
[a]
[acetate]/[malonate]/[NaH]/[Pd2(dba)3]/[ixgand] ϭ 50:150:150:1:2;
[b]
solvent: THF; 25 °C; 24 h. Ϫ
Determined by HPLC analysis
with chiral stationary column Daicel Chiralcel ODϪH (hexane/2-
propanol 98:2); the configuration of the alkylated product was
(S).[10]
Ϫ
N,O-Bis(trimethylsilyl)acetamide (or BSA) and KOAc
[c]
Asymmetric Alkylation
[d]
were used instead of NaH. Ϫ
N,O-Bis(trimethylsilyl)acetamide
only was used instead of NaH.
The effectiveness of the above ligands in asymmetric al-
lylic alkylation was tested in the reaction of racemic (E)-
1,3-diphenylprop-2-enyl acetate with dimethyl malonate; an
in situ catalyst prepared from Pd2(dba)3 and the chiral li-
gand in THF was used (Scheme 4).
It should first be noted that the use of ligands with C2
symmetry, such as DIOP or ligand 12, led to the formation
of the racemic alkylated product (Entries 1Ϫ2). Ligands
3aϪd gave enantioselectivities up to 69% in THF as the
solvent (Entries 3Ϫ6). Dissymmetrization at the nitrogen
atom increased the enantioselectivity, although an enanti-
oselectivity of 50% was observed with 3a as the ligand; val-
ues of 68% and 69% were obtained in the presence of 3b or
3d, respectively. The presence of a methoxy group on the
phenyl ring lowered the enantioselectivity (63% for 3c). To
Scheme 4. a: CH2(CO2Me)2, base, cat. Pd2(dba)3, ligand
As a typical procedure, tris(dibenzylideneacetone)dipalla- increase the enantioselectivity, the alkylation was performed
dium and the amino phosphane were mixed in a solvent at under various conditions, with 3d as the ligand. While the
room temperature under argon; rac-(E)-1,3-diphenylprop- use of BSA as the base in the presence of KOAc gave both
2-enyl acetate was added, then the sodium salt of dimethyl lower yield and enantioselectivity (Entry 7), performing the
malonate was added to the π-allylpalladium complex solu- reaction in CH2Cl2 gave almost the same e.e. (68%) in the
352
Eur. J. Inorg. Chem. 2000, 351Ϫ358

Chiral 4-(Diphenylphosphanyl)-1-(dialkylamino)butane Ligands
FULL PAPER
presence or absence of KOAc (Entries 8Ϫ9). The use of and Wang (BPW91). For palladium and phosphorus, the
toluene (Entry 10) or a mixture of toluene/CH2Cl2 (Entry pseudopotentials of Hay and Wadt are used with the corre-
11) as solvent again gave low yield and enantioselectivity. sponding double ζ basis set of Dunning/Huzinaga (D95). A
The use of ligands 9 in this alkylation reaction led to high d polarization function is added to the palladium, nitrogen
chemical yields, with enantioselectivities depending on the and carbon atoms. The ligands on the phosphorus center
nature of the aromatic ring. The highest enantioselectivity are modelled by PH3 and NH3 to save computational time.
(75%) was obtained with ligand 9b, which has 2,4,6-trimeth- Since the allyl moiety is symmetrical with one phenyl group
ylbenzyl substituents on the nitrogen atom. Finally, ligand on each terminal carbon atom, we chose to model it in the
11 gave low yield (only 49%) and quite modest enantioselec- first instance by the unsubstituted allyl group C3H5. By do-
tivity (39%). To obtain more insight into the reasons behind ing this, we obtained the electronic effects without taking
the high enantioselectivities when these ligands are used, we into account the steric effects.
undertook a theoretical study.
Results
Theoretical Study
The optimized geometry is shown in Figure 1. The most
Some theoretical studies on η3-allyl complexes of pal-
important information is that the PdϪCtrans-P bond (2.23
ladium have already been reported. In the past, semi-em-
pirical methods have been used to study the nucleophilic
attack on such complexes: Extended Hückel type (EHT)
calculations were carried out to probe the effect of ligand
asymmetry,[11] to distinguish between the central and the
terminal carbon atoms,[12] to calculate the effect of the sub-
stituents on the regioselectivity,[13] and CNDO type calcu-
lations were carried out to determine the ligand effects.[14]
The conclusions of these studies were that the nucleophilic
attack on η3-allylpalladium complexes is orbital-controlled
rather than charge-controlled. More recently, studies using
ab initio methods, both at the HartreeϪFock level and with
the density functional theory (DFT), have begun to ap-
pear.[15Ϫ20]
In the present work we have used both types of methods,
EHT and DFT. To determine the site of the nucleophilic
attack, we have used the same method as in our previous
work.[13] Since the reaction is frontier orbital controlled, we
have focused first on the shape of the complex LUMO
(Lowest Unoccupied Molecular Orbital) which interacts
with the nucleophile HOMO (Highest Occupied Molecular
Orbital). The larger the coefficient on one carbon atom in
this orbital, the easier is the attack of the nucleophile on
that carbon atom. Then we have put some nucleophiles at
the same distance to both terminal carbon atoms and com-
pared the overlap populations. These calculations are based
on the EHT method. However, it is known from experi-
ments that the presence of ligands of different natures on
the palladium center induces a dissymmetry in the com-
plex.[21Ϫ23] The EHT method does not allow the optimiz-
ation of geometries. This is why we used a DFT method
for these optimizations, to have reliable geometries for the
EHT calculations.
˚
˚
A) is longer than the PdϪCtrans-N one (2.16 A). This result
can be explained by the trans effect of the phosphorus
atom, which is greater than that of the nitrogen atom.[28]
Experimentally, the same difference was observed,[15] al-
though the nitrogen ligand was not of the same nature (sp2
instead of sp3). The C1ϪC2 and C2ϪC3 bond lengths are
˚
not equal either (1.417 and 1.432 A, respectively). This me-
ans that the C2ϪC3 bond has a weaker double bond charac-
1
3
˚
ter. The carbon atoms C and C are 0.14Ϫ0.15 A above
the PϪPdϪN plane, and the allyl plane makes an angle of
111° with the PdϪC1ϪC3 plane. Similar geometries were
found experimentally.[22,23,29] The charge on carbon atoms
C1 and C3 (0.55 eϪ) is the same. As found in our previous
EHT study, the LUMO is mainly the antibonding combi-
nation of the dxy orbital of the metal center and the n-allyl
orbital and the next unoccupied orbital is composed almost
exclusively of the π*-allyl orbital.
Figure 1. Optimized geometry for palladium complex Pd(NH3)-
(PH3)(η3-allyl)
After this geometry optimization, we performed EHT
calculations, following the method of our previous work.[13]
We choose the same standard distances but we kept the
difference between PdϪC1 and PdϪC3 bonds fixed (0.08
˚
˚
A). The following geometry was used: PdϪP ϭ 2.3 A,
1
3
˚
3
˚
˚
PdϪN ϭ 2.15 A, PdϪC ϭ 2.20 A, PdϪC ϭ 2.12 A,
1
2
2
˚
C ϪC ϭ C ϪC ϭ 1.43 A, PϪPdϪN ϭ 100°. For the sake
Computational Methods
of simplicity, the allyl plane was taken to be perpendicular
The parameters used for the EHT calculations are the to the PϪPdϪN plane and to the bisector of the PϪPdϪN
usual ones[24][25] and the program is ICON version 8.0 from angle. The hydrogen atoms on the carbon atoms C1 and C3
the group of R. Hoffmann at Cornell University.[26] The ab are bent away from the metal center by 15°. The projection
initio calculations are performed with Gaussian 94[27] with on the xy plane is shown in Figure 2. The complex LUMO
the DFT formalism. The chosen function is that of Perdew is also shown.
Eur. J. Inorg. Chem. 2000, 351Ϫ358
353

F. Robert, F. Delbecq, C. Nguefack, D. Sinou
FULL PAPER
respectively. The best conformation of the nucleophile was
searched for by rotating around the approach axis. The re-
sults are shown in Table 2. For all nucleophiles, the overlap
population is larger with carbon atom C1 than with C3; this
confirms that a nucleophile will preferentially attack at the
carbon atom trans to the phosphorus atom, as long as only
electronic interactions are taken into account. It is obvious
that steric constraints, if for instance the carbon atoms are
substituted, can reverse this trend.
Figure 2. Projection of the model allylϪPd complex on the xy
plane and shape of its LUMO
We observe that the px coefficient is larger on the carbon
atom C1 than on C3 (0.53 and 0.48, respectively). If the
reaction is orbital-controlled, an incoming nucleophile will
prefer to attack the carbon atom with the largest contri-
bution in the LUMO. Hence it will preferentially attack C1,
trans to the phosphorus atom. Why is the LUMO dissym-
metrical? The molecular orbital interaction diagram be-
tween the PϪPdϪN and the allyl fragments shows that the
LUMO, which is mainly antibonding between dxy and the
n-allyl orbital, also contains a small antibonding contri-
bution of the π-allyl orbital and a small bonding contri-
bution of the π*-allyl orbital, according to Equation 1.
Table 2. Overlap populations between a nucleophile and the termi-
nal carbon atoms C1 and C3; energy in eV of the nucleophile
HOMO
Nucleophile
HOMO
C1
C3
NH3
Ϫ13.71
Ϫ12.95
Ϫ12.35
0.143
0.157
0.180
0.138
0.152
0.175
NH(Me)2
ϪHC(CO2Me)2
The interaction between the nucleophile and the allylic
complex can be separated into terms of two-electron and
four-electron interactions, by a simple perturbational treat-
ment.[13][30] We consider the function f(x), which is the sum
of the stabilizing interactions of the nucleophile HOMO
with the vacant orbitals of the complex and the destabiliz-
ing interactions between the nucleophile HOMO and the
occupied orbitals of the complex (Equation 2). The more
positive f(x) is, the more the interaction is stabilizing.
LUMO ϭ 0.80 (PdϪPϪN LUMO) Ϫ 0.69 (n) Ϫ0.04 (π)
(1)
ϩ 0.08 (π*)
In fact, the LUMO of the PϪPdϪN fragment is dissym-
metric because the two ligands are different. This allows
the mixing of the three allyl orbitals which are of different
symmetry and they do not mix when the metal fragment is
symmetrical (Figure 3).
f(x) ϭ 0.75Σ(x ϩ Ei) Ci2 ϩ Σ(0.875 Ej Ϫ 0.125x)2 C 2/(Ej Ϫ x) (2)
j
In Equation 2, x is the energy of the nucleophile HOMO,
Ci and Cj are the coefficients of the px orbital of the at-
tacked carbon atom in the occupied and vacant complex
orbitals, respectively, and Ei and Ej are the energies of the
corresponding complex orbitals; x was varied and f(x) was
calculated for C1 [f1(x)] and C3 [f3(x)]. The difference
F(x) ϭ f1(x) Ϫ f3(x) is drawn in Figure 4. If F(x) < 0, the
nucleophile prefers carbon atom C3, if F(x) > 0, the nucle-
ophile prefers C1. The attack preference depends on the
HOMO of the nucleophile. For a high-lying HOMO (see
Figure 3. Shape of the LUMO of the PϪPdϪN fragment and of
the n,π and π* orbitals of the allyl fragment
However, we saw previously that other metal orbitals Table 2), such as that of the malonate anion, carbon atom
must be taken into account in the interaction with a nucle- C1 trans to the phosphorus atom is preferentially attacked;
ophile: for instance, the next LUMO and some occupied this agrees with the experimental results. The attractive two-
orbitals which have contributions to px of C1 and px of C3. electron interactions always favor the attack on Ctrans-P and,
Therefore, we tested the approach of some nucleophiles by conversely, the repulsive four-electron interactions favor the
putting them at the same distance to carbon atoms C1 and attack on Ccis-P. Therefore, the site of attack depends on
C3, and comparing the overlap population that represents the balance between these two terms. The attractive term
the strength of the formed bond. To avoid too many steric increases greatly with the HOMO energy. The repulsive
repulsions, the allyl part is taken as being planar, this means term, being proportional to the sum of the interacting or-
that the hydrogen atoms are no longer bent away from the bital energies (Equation 2), depends little on the HOMO
palladium center but are in the allyl plane (this looks like position.
the transition state of an SN2 reaction). We chose NH3,
We therefore explained why the electronic interactions
NHMe2 and the malonate anion ϪCH(CO2Me)2 as nucleo- favor attack on carbon atom Ctrans-P. We can predict that
philes, the latter of which is the nucleophile experimentally the use of a nucleophile with a higher HOMO than malon-
used in the present work. The amines and the malonate ate, such as ϪC(CH3)(CO2Me)2, for example, would en-
˚
˚
anion are put at 2 A and 2.1 A away from the carbon atoms, hance the regioselectivity. How can the enantioselectivity be
354
Eur. J. Inorg. Chem. 2000, 351Ϫ358

Chiral 4-(Diphenylphosphanyl)-1-(dialkylamino)butane Ligands
FULL PAPER
for both [Pd/3a] and [Pd/3b]. For [Pd/3b] there are two exo
and two endo isomers depending on the relative position
(pseudo-axial or pseudo-equatorial) of the C6H5 and CH3
substituents on the nitrogen atom. For [Pd/3a] the exo form
is more stable by 30 kcal molϪ1. For [Pd/3b], the exo form
is more stable by 18 kcal molϪ1 when the methyl group is
pseudo-axial and the endo form is more stable by 67 kcal -
molϪ1 when the methyl group is pseudo-equatorial. How-
ever, the structure with the pseudo-equatorial methyl group
is far less stable than that with the pseudo-axial one
(47 kcal molϪ1). Therefore, of the four diastereoisomers of
[Pd/3b], the most stable also has the exo geometry. The exo
form of [Pd/3a] and [Pd/3b], attacked trans to the phos-
phorus atom, gives the (S) enantiomer, which is the one
experimentally obtained. Despite the fact that the EHT
method only gives qualitative results, the energy differences
are large enough to be reliable. Hence, the enantioselectivity
is explained by the existence of only one diastereoisomer of
the complexes [Pd/3a] and [Pd/3b], regioselectively attacked
trans to the phosphorus atom.
Figure 4. Plot of f(x) versus energy of the nucleophile HOMO
explained? For the π-allylpalladium complexes, two dia-
stereoisomers are possible, exo and endo (Figure 5).
Conclusion
In conclusion, the new chiral amino phosphane ligands,
easily obtained from tartaric acid, give enantioselectivities
of up to 75% in the palladium-catalyzed alkylation of ra-
cemic 1,3-diphenylprop-2-enyl acetate with malonate anion.
The calculations summarized in this paper allow the origin
of the enantioselectivity in this asymmetric allylic alkylation
to be understood. Two factors play an important role. The
first one is steric control, which leads to the formation of a
unique diastereoisomer (exo) of the η3-allylpalladium com-
plex. The second one is orbital control, which orients the
nucleophile attack trans to the phosphorus atom; this is due
to the dissymmetrical ligand. This is not the case for DIOP
or ligand 12. Our calculations also allow some predictions:
To enhance the enantioselectivity, the HOMO of the nucle-
ophile must be as high as possible. This means that the
anion of CH(CH3)(CO2Me)2 or of cyclopentanedione
should be more efficient than the malonate anion
Figure 5. Most stable diastereoisomers of the model (PH3)(NH3)-
(1,3-diphenylallyl)Pd complex
The phenyl rings may be positioned syn or anti, therefore
eight diastereoisomers are possible. Using the EHT
method, we calculated all of them and optimized the angle
of the phenyl rings with the allyl plane. When the phenyl
group is syn, it prefers to be in the allyl plane, when it is
anti, it prefers to be perpendicular to this plane. This is
in agreement with previous calculations.[5] The most stable
isomers (by 13 kcal molϪ1) are those depicted in Figure 5
(exo/syn/syn and endo/syn/syn). With our model ligands
PH3 and NH3, they are quasi isoenergetic, which cannot
explain the enantioselectivity, since, by attack trans to the
phosphorus atom, they each lead to one of the enantiomers
(S) and (R). We therefore modeled the real catalyst by sub-
stituting the phosphorus and nitrogen atoms with phenyl
and methyl groups, as shown in Figure 6. This corresponds
to real catalysts [Pd/3a] and [Pd/3b], obtained by association
of the palladium center with ligands 3a and 3b.
Ϫ
CH(CO2Me)2
Experimental Section
1
General Remarks: H NMR (200 or 300 MHz), 13C NMR (50 or
75 MHz), and 31P NMR (80 MHz) spectra were obtained with a
Brüker AM 200 or AM 300 spectrometer. Ϫ Optical rotations were
determined with a PerkinϪElmer 241 polarimeter. Ϫ GC analyses
were recorded with a Shimadzu capillary gas chromatograph,
equipped with a ChiraldexTM type BϪPH (30 m ϫ 0.32 mm) capil-
lary column. Ϫ Reactions involving organometallic catalysis were
carried out in Schlenk tubes under an inert gas. Tetrahydrofuran
was distilled from sodium/benzophenone. (2S,3S)-2,3-O-Isopro-
pylidene-1-O-p-toluenesulfonyl--threitol (4) was prepared accord-
ing to a literature procedure.[9]
Figure 6. Modeling of the catalyst Pd(CH3PPh2)(CH3NPhR)(η3-
1,3-diphenylallyl)
The substituents on the phosphorus and nitrogen atoms
are disposed in a geometry that reproduced the X-ray struc-
General Procedure for Preparation of Amino Tosylates 2: A solution
of the lithium derivative of the secondary amine, obtained by mix-
ture of DIOP.[31] We calculated the exo and endo isomers ing the amine (5 mmol) and butyllithium in hexane (5 mmol) in
Eur. J. Inorg. Chem. 2000, 351Ϫ358
355

F. Robert, F. Delbecq, C. Nguefack, D. Sinou
FULL PAPER
THF (5 mL), was slowly added to a solution of ditosylate 1 (2.35 g,
drogenated at room temperature and atmospheric pressure in the
5 mmol) in THF (5 mL) at Ϫ17°C. The solution was stirred for 1 h presence of 20% Pd/C (300 mg). After 48 h, the solution was fil-
at Ϫ17°C, then 12 h at room temperature, and was finally treated tered and the solvent was evaporated to give 2.57 g of amino al-
20
1
with H2O (50 mL). The organic compound was extracted with
CH2Cl2 (3 ϫ 50 mL). Evaporation of the solvent, followed by flash
chromatography gave the amino tosylate 2.
cohol 6 as an oil (100%); [α]D ϭ Ϫ47.3 (c ϭ 1.0, CHCl3). Ϫ H
NMR (200 MHz, CDCl3): δ ϭ 1.32 (s, 3 H, CH3), 1.38 (s, 3 H,
CH3), 2.20Ϫ2.50 (br. s, 3 H, NH2, OH), 2.76 (dd, J ϭ 12.6, 8.0 Hz,
1 H, CH2N), 3.18 (dd, J ϭ 12.6, 4.3 Hz, 1 H, CH2N), 3.70Ϫ3.92
(m, 4 H, CH2O, CHO). C7H15NO3 (161.20): calcd. C 52.16, H 9.38;
found C 51.76, H 9.16.
(2S,3S)-4-Diphenylamino-2,3-O-isopropylidene-1-O-tosylbutane-
1,2,3-triol (2a): 26% yield; Rf ϭ 0.5 (eluent: hexane/ethyl acetate,
20
4:1); [α]D ϭ ϩ8.5 (c ϭ 1.1, CHCl3). Ϫ 1H NMR (200 MHz,
General Procedure for the Preparation of Amino Alcohols 7: The
bromide (12.4 mmol) was slowly added to a mixture of amino al-
cohol 6 (1 g, 6.2 mmol) and K2CO3 (2,57 g, 18.6 mmol) in CH3CN
(6 mL), heated at reflux for 30 min. The reaction mixture was re-
fluxed for 24 h, then cooled at room temperature, and treated with
H2O (10 mL). The solution was extracted with CH2Cl2
(3 ϫ 30 mL). Evaporation of the solvent, followed by flash chroma-
tography gave the corresponding dialkylated amino alcohol 7.
CDCl3): δ ϭ 1.29 (s, 3 H, CH3), 1.38 (s, 3 H, CH3), 2.43 (s, 3 H,
CH3), 3.90Ϫ4.05 (m, 5 H, CH2N, CH2O, CHO), 4.05Ϫ4.20 (m, 1
H, CHO), 6.90Ϫ7.10 (m, 6 H, Harom), 7.15Ϫ7.30 (m, 6 H, Harom),
7.71 (d, J ϭ 8.3 Hz, 2 H, Harom). Ϫ 13C NMR (50 MHz, CDCl3):
δ ϭ 21.6 (CH3), 26.8 (CH3), 27.2 (CH3), 55.0 (d, CH2N), 69.2
(CH2O), 76.0 (CHO), 76.9 (CHO), 110.2 (CMe2), 121.2, 121.9,
128.0, 129.4, 129.8, 132.5, 145.0 and 148.0 (Carom). Ϫ C26H29NO5S
(467.58): calcd. C 66.79, H 6.86; found C 66.65, H 6.57.
(2S,3S)-4-Dibenzylamino-2,3-O-isopropylidenebutane-1,2,3-triol
(7a): 85% yield; Rf ϭ 0.3 (eluent: petroleum ether/ethyl acetate,
(2S,3S)-2,3-O-Isopropylidene-4-methyl(phenyl)amino-1-O-tosyl-
butane-1,2,3-triol (2b): 32% yield; Rf ϭ 0.4 (eluent: hexane/ethyl
20
4:1); [α]D ϭ ϩ32.1 (c ϭ 0.8, CHCl3). Ϫ 1H NMR (200 MHz,
20
1
acetate, 4:1); [α]D ϭ ϩ24.6 (c ϭ 1.1, CHCl3). Ϫ H NMR (200
MHz, CDCl3): δ ϭ 1.31 (s, 3 H, CH3), 1.39 (s, 3 H, CH3), 2.45 (s,
3 H, CH3), 2.95 (s, 3 H, CH3), 3.46 (dd, J ϭ 15.0, 5.7 Hz, 1 H,
CH2N), 3.55 (dd, J ϭ 15.0, 4.5 Hz, 1 H, CH2N), 3.90Ϫ4.05 (m, 1
H, CHO), 4.05Ϫ4.20 (m, 3 H, CH2O, CHO), 6.70 (m, 3 H, Harom),
7.25 (m, 2 H, Harom), 7.38 (d, J ϭ 8.5 Hz, 2 H, Harom), 7.76 (d, J ϭ
8.5 Hz, 2 H, Harom). Ϫ C21H27NO5S (405.51): calcd. C 62.20, H
6.71; found C 62.65, H 7.11.
CDCl3): δ ϭ 1.35 (s, 3 H, CH3), 1.40 (s, 3 H, CH3), 2.63 (s, 1 H,
OH), 2.72 (d, J ϭ 5.5 Hz, 2 H, CH2N), 3.60Ϫ3.90 (m, 8 H, CH2O,
CHO, CH2N), 7.30Ϫ7.50 (m, 10 H, Harom). Ϫ C21H27NO3 (341.45):
calcd. C 73.87, H 7.97, N 4.10; found C 73.73, H 8.12, N 4.04.
(2S,3S)-4-Bis(2,4,6-trimethylbenzyl)amino-2,3-O-isopropylidene-
butane-1,2,3-triol (7b): 76% yield; Rf ϭ 0.4 (eluent: petroleum ether/
20
ethyl acetate, 4:1); [α]D ϭ ϩ10.2 (c ϭ 1, CHCl3). Ϫ 1H NMR
(200 MHz, CDCl3): δ 1.35 (s, 3 H, CH3), 1.41 (s, 3 H, CH3), 1.65
(s, 1 H, OH), 2.23 (s, 6 H, CH3), 2.26 (s, 12 H, CH3), 2.58 (m, 2
H, CH2N), 3.40Ϫ3.50 (m, 4 H, CH2O), 3.49 (d, J ϭ 12.5 Hz, 2 H,
CH2N), 3.69 (d, J ϭ 12.5 Hz, 2 H, CH2N), 3.80Ϫ4.00 (m, 2 H,
CHO), 6.75 (s, 4 H, Harom). Ϫ C27H39NO3 (425.62): calcd. C 76.24,
H 9.17; found C 76.21, H 9.34.
(2S,3S)-2,3-O-Isopropylidene-4-p-methoxyphenyl(methyl)amino-1-
O-tosylbutane-1,2,3-triol (2c): 25% yield; Rf ϭ 0.55 (eluent: hexane/
20
ethyl acetate, 3:1); [α]D ϭ ϩ6.3 (c ϭ 0.7, CHCl3). Ϫ 1H NMR
(200 MHz, CDCl3): δ ϭ 1.30 (s, 3 H, CH3), 1.40 (s, 3 H, CH3),
2.88 (s, 3 H, CH3), 2.89 (s, 3 H, CH3), 3.40 (d, J ϭ 5.2 Hz, 2 H,
CH2N), 3.78 (s, 3 H, CH3), 3.90Ϫ4.05 (m, 1 H, CHO), 4.05Ϫ4.20
(m, 3 H, CH2O, CHO), 6.60Ϫ7.00 (m, 4 H, Harom), 7.20Ϫ7.40 (m,
2 H, Harom). Ϫ C22H29NO6S (435.54): calcd. C 60.67, H 6.71; found
C 61.82, H 6.40.
(2S,3S)-4-Bis(2-naphtylmethyl)amino-2,3-O-isopropylidenebutane-
1,2,3-triol (7c): 82% yield; Rf ϭ 0.34 (eluent: petroleum ether/ethyl
20
1
acetate, 4:1); [α]D ϭ ϩ14.0 (c ϭ 0.65, CHCl3). Ϫ H NMR (200
MHz, CDCl3): δ ϭ 1.32 (s, 3 H, CH3), 1.38 (s, 3 H, CH3), 1.61 (s,
1 H, OH), 2.81 (d, J ϭ 5.7 Hz, 2 H, CH2N), 3.73 (d, J ϭ 13.3 Hz,
2 H, CH2N), 3.98 (d, J ϭ 13.3 Hz, 2 H, CH2N), 3.50Ϫ3.70 (m, 4
H, CH2O), 4.00Ϫ4.20 (m, 2 H, CHO), 7.64 (m, 6 H, Harom),
7.80Ϫ7.90 (m, 8 H, Harom). Ϫ C29H31NO3 (441.57): calcd. C 78.91,
H 7.01; found C 80.15, H 7.06.
(2S,3S)-2,3-O-Isopropylidene-4-α-naphthyl(phenyl)amino-1-O-tosyl-
butane-1,2,3-triol (2d): 25% yield; Rf ϭ 0.42 (eluent: hexane/ethyl
20
1
acetate, 5:1); [α]D ϭ ϩ21.6 (c ϭ 0.8, CHCl3). Ϫ H NMR (200
MHz, CDCl3): δ ϭ 1.30 (s, 3 H, CH3), 1.41 (s, 3 H, CH3), 2.38 (s,
3 H, CH3), 3.95Ϫ4.10 (m, 5 H, CH2N, CH2O, CHO), 4.15Ϫ4.25
(m, 1 H, CHO), 7.00Ϫ7.50 (m, 11 H, Harom), 7.60Ϫ7.80 (m, 4 H,
Harom). Ϫ C30H31NO5S (517.64): calcd. C 69.61, H 6.04; found C
69.18, H 6.17.
(2S,3S)-4-(6,7-Dihydro-5H-dibenzo[c,e]azepin-6-yl)-2,3-O-isoprop-
ylidenebutane-1,2,3-triol (7d): 58% yield; Rf ϭ 0.20 (eluent: petro-
leum ether/ethyl acetate, 1:2). Ϫ 1H NMR (200 MHz, CDCl3): δ ϭ
1.27 (s, 3 H, CH3), 1.32 (s, 3 H, CH3), 2.61 (dd, J ϭ 12.7, 5.7 Hz,
1 H, CH2N), 2.87 (s, 1 H, OH), 3.11 (dd, J ϭ 12.7, 3.7 Hz, 1 H,
CH2N), 3.38 (d, J ϭ 12.5 Hz, 2 H, CH2N), 3.47 (d, J ϭ 12.5 Hz,
2 H, CH2N), 3.69 (dd, J ϭ 9.4, 1.9 Hz, 1 H, CH2O), ), 3.87 (dd,
J ϭ 9.4, 3.4 Hz, 1 H, CH2O), 3.90Ϫ4.15 (m, 2 H, CHO), 7.30Ϫ7.50
(m, 8 H, Harom). Ϫ C21H25NO3 (339.44): calcd. C 74.31, H 7.42;
found C 74.24, H 7.27.
(2S,3S)-4-Azido-2,3-O-isopropylidenebutane-1,2,3-triol (5): A mix-
ture of (2S,3S)-2,3-O-isopropylidene-1-O-p-toluenesulfonyl--thre-
itol (4)[9] (11.8 g, 37.5 mmol) and NaN3 (3.5 g, 58 mmol) in DMF
(95 mL) was heated at reflux for 24 h. After being cooled to room
temperature, the solution was treated with H2O (50 mL) and was
then extracted with CH2Cl2 (3 ϫ 50 mL). Evaporation of the sol-
vent, followed by column chromatography (eluent: petroleum ether/
ethyl acetate, 2:1) gave 6.6 g of compound 5 as an oil (95%); Rf ϭ
20
0.3; [α]D ϭ Ϫ65.7 (c ϭ 2.5, CHCl3). Ϫ 1H NMR (200 MHz,
General Procedure for Preparation of Aminotosylates 8: p-Tolu-
enesulfonyl chloride (1.13 g, 5.9 mmol, 1.1 equiv.) was added at
0°C to a solution of amino alcohol 7 (5.3 mmol) in pyridine
(10 mL). After being stirred for 12 h, the solution was treated with
cold H2O (20 mL), and the organic product was extracted with
CH2Cl2 (3 ϫ 10 mL). Evaporation of the solvent gave a residue
which was purified by flash chromatography to give the amino tos-
ylate 8.
CDCl3): δ ϭ 1.36 (s, 3 H, CH3), 1.41 (s, 3 H, CH3), 2.84 (s, 1 H,
OH), 3.34 (dd, J ϭ 13.1, 4.6 Hz, 1 H, CH2N), 3.55 (dd, J ϭ 13.1,
3.7 Hz, 1 H, CH2N), 3.74 (dd, J ϭ 13.8, 3.9 Hz, 1 H, CH2O), 3.86
(dd, J ϭ 13.8, 4.0 Hz, 1 H, CH2O), 4.02 (ddd, J ϭ 11.4, 4.6, 3.7 Hz,
1 H, CHO), 4.07 (ddd, J ϭ 11.4, 4.0, 3.9 Hz, 1 H, CHO). Ϫ
C7H13N3O3 (187.20): calcd. C 44.91, H 6.00; found C 44.96, H
7.28.
(2S,3S)-4-Amino-2,3-O-isopropylidenebutane-1,2,3-triol (6): A solu-
tion of azido alcohol 5 (3 g, 16 mmol) in ethanol (20 mL) was hy-
(2R,3S)-4-Dibenzylamino-2,3-O-isopropylidene-1-O-tosylbutane-
1,2,3-triol (8a): 91% yield; Rf ϭ 0.35 (eluent: petroleum ether/ethyl
356
Eur. J. Inorg. Chem. 2000, 351Ϫ358

Chiral 4-(Diphenylphosphanyl)-1-(dialkylamino)butane Ligands
FULL PAPER
20
acetate, 4:1); [α]D ϭ Ϫ7.0 (c ϭ 1.0, CHCl3). Ϫ 1H NMR (200
CH2N), 4.01 (d, J ϭ 13.6, 4.4 Hz, 1 H, CH2N), 4.00Ϫ4.20 (m, 2
MHz, CDCl3): δ ϭ 1.18 (s, 6 H, CH3), 2.36 (s, 3 H, CH3), 2.56 (d, H, CHO), 7.05Ϫ7.70 (m, 20 H, Harom). Ϫ 31P (80 MHz, CDCl3):
J ϭ 5.5 Hz, 2 H, CH2N), 3.45 (d, J ϭ 13.5, 2 H, CH2N), 3.59 (d, δ ϭ Ϫ22.1. Ϫ C31H32NO2P (481.58): calcd. C 77.32, H 6.70; found
J ϭ 13.5 Hz, 2 H, CH2N), 3.60Ϫ3.80 (m, 2 H, CH2O), 3.85 (m, 1
H, CHO), 4.05 (m, 1 H, CHO), 7.10Ϫ7.30 (m, 12 H, Harom), 7.66
(d, J ϭ 8.1 Hz, 2 H, Harom). Ϫ C28H33NO5S (495.64): calcd. C
67.85, H 6.71; found C 67.80, H 6.70.
C 76.90, H 6.76.
(2R,3S)-1-Diphenylphosphanyl-2,3-O-isopropylidene-4-methyl(phen-
yl)aminobutane-2,3-diol (3b): 55% yield; Rf ϭ 0.50 (eluent: petro-
20
leum ether/ethyl acetate, 6:1) ): [α]D ϭ Ϫ12.3 (c ϭ 0.9, CHCl3).
1
(2R,3S)-4-Bis(2,4,6-trimethylbenzyl)amino-2,3-O-isopropylidene-1-
O-tosylbutane-1,2,3-triol (8b): 65% yield; Rf ϭ 0.60 (eluent: petro-
Ϫ H NMR (200 MHz, CDCl3): δ ϭ 1.38 (s, 3 H, CH3), 1.39 (s, 3
H, CH3), 2.33 (dd, J ϭ 13.8, 6.0 Hz, 1 H, CH2P), 2.45 (dd, J ϭ
13.8, 6.8 Hz, 1 H, CH2P), 2.88 (s, 3 H, CH3), 3.32 (dd, J ϭ 15.1,
6.8 Hz, 1 H, CH2N), 3.62 (d, J ϭ 15.1, 4.2 Hz, 1 H, CH2N), 3.86
(ddd, J ϭ 11.8, 6.8, 6.0 Hz, 1 H, CHO), 4.11 (ddd, J ϭ 11.8, 6.8,
4.2 Hz, 1 H, CHO), 6.70Ϫ6.85 (m, 3 H, Harom), 7.20Ϫ7.40 (m, 12
H, Harom). Ϫ 31P (80 MHz, CDCl3): δ ϭ Ϫ23.8. Ϫ C26H30NO2P
(419.51): calcd. C 74.44, H 7.21; found C 74.55, H 7.31.
20
leum ether/ethyl acetate, 4:1); [α]D ϭ Ϫ10.2 (c ϭ 1.1, CHCl3). Ϫ
1H NMR (200 MHz, CDCl3): δ ϭ 1.19 (s, 3 H, CH3), 1.22 (s, 3 H,
CH3), 2.23 (s, 6 H, CH3), 2.24 (s, 12 H, CH3), 2.45 (s, 3 H, CH3),
2.56 (d, J ϭ 5.0 Hz, 2 H, CH2N), 3.50 (d, J ϭ 13.1, 2 H, CH2N),
3.58 (d, J ϭ 13.1 Hz, 2 H, CH2N), 3.60Ϫ3.71 (m, 4 H, CH2O,
CHO), 6.78 (s, 4 H, Harom), 7.31 (d, J ϭ 8.1 Hz, 2 H, Harom), 7.70
(d, J ϭ 8.1 Hz, 2 H, Harom). Ϫ C34H45NO5S (579.80): calcd. C
70.43, H 7.82; found C 70.25, H 8.04.
(2R,3S)-1-Diphenylphosphanyl-2,3-O-isopropylidene-4-p-methoxy-
phenyl(methyl)aminobutane-2,3-diol (3c): 27% yield; Rf ϭ 0.34 (elu-
20
(2R,3S)-4-Bis(2-naphtylmethyl)amino-2,3-O-isopropylidene-1-O-
tosylbutane-1,2,3-triol (8c): 70% yield; Rf ϭ 0.40 (eluent: petroleum
ent: petroleum ether/ethyl acetate, 10:1); [α]D ϭ Ϫ6.6 (c ϭ 0.9,
1
CHCl3). Ϫ H NMR (200 MHz, CDCl3): δ ϭ 1.31 (s, 3 H, CH3),
20
ether/ethyl acetate, 4:1); [α]D ϭ Ϫ6.2 (c ϭ 1.2, CHCl3). Ϫ 1H
1.33 (s, 3 H, CH3), 2.27 (dd, J ϭ 13.8, 6.0 Hz, 1 H, CH2P), 2.37
(dd, J ϭ 13.8, 6.5 Hz, 1 H, CH2P), 2.74 (s, 3 H, CH3), 3.19 (dd,
J ϭ 14.8, 6.7 Hz, 1 H, CH2N), 3.40 (d, J ϭ 14.8, 4.2 Hz, 1 H,
CH2N), 3.71 (s, 3 H, CH3), 3.82 (ddd, J ϭ 11.2, 6.5, 6.0 Hz, 1 H,
CHO), 4.01 (ddd, J ϭ 11.2, 6.7, 4.2 Hz, 1 H, CHO), 6.74 (m, 4 H,
Harom), 7.50 (m, 10 H, Harom). Ϫ 31P (80 MHz, CDCl3): δ ϭ Ϫ23.7.
ϪC27H32NO3P (449.53): calcd. C 72.14, H 7.18; found: C 72.16,
H 6.23.
NMR (200 MHz, CDCl3): δ ϭ 1.25 (s, 3 H, CH3), 1.27 (s, 3 H,
CH3), 2.37 (s, 3 H, CH3), 2.75 (d, J ϭ 5.9 Hz, 2 H, CH2N), 3.73
(d, J ϭ 13.1, 2 H, CH2N), 3.88 (d, J ϭ 13.1 Hz, 2 H, CH2N),
3.70Ϫ3.90 (m, 2 H, CHO), 4.00Ϫ4.20 (m, 2 H, CH2O), 7.20 (d,
J ϭ 8.1 Hz, 2 H, Harom), 7.50Ϫ7.70 (m, 6 H, Harom), 7.62 (d, J ϭ
8.1 Hz, 2 H, Harom), 7.70Ϫ7.90 (m, 8 H, Harom). Ϫ C36H37NO5S
(595.76): calcd. C 72.58, H 6.26; found C 72.60, H 6.21.
(2R,3S)-4-(6,7-Dihydro-5H-dibenzo[c,e]azepin-6-yl)-2,3-O-iso-
propylidene-1-O-tosylbutane-1,2,3-triol (8d): Yield: 63%; Rf ϭ 0.65
(eluent: petroleum ether/ethyl acetate, 1:1). Ϫ 1H NMR (200 MHz,
CDCl3): δ ϭ 1.38 (s, 3 H, CH3), 1.44 (s, 3 H, CH3), 2.42 (s, 3 H,
CH3), 2.66 (dd, J ϭ 13.1, 4.8 Hz, 1 H, CH2N), 2.85 (dd, J ϭ 13.1,
6.5 Hz, 1 H, CH2N), 3.34 (d, J ϭ 12.4 Hz, 2 H, CH2N), 3.46 (d,
J ϭ 12.4 Hz, 2 H, CH2N), 3.96 (ddd, J ϭ 8.3, 4.7, 3.9 Hz, 1 H,
CHO), 4.10Ϫ4.30 (m, 12 H, CHO), 4.19 (dd, J ϭ 10.6, 4.7 Hz, 1
H, CH2O), 4.30 (dd, J ϭ 10.6, 3.9 Hz, 1 H, CH2O), 7.30Ϫ7.60 (m,
10 H, Harom), 7.82 (d, J ϭ 8.2 Hz, 2 H, Harom). Ϫ C28H31NO5S
(493.62): calcd. C 68.13, H 6.33; found C 68.47, H 6.90.
(2R,3S)-1-Diphenylphosphanyl-2,3-O-isopropylidene-4-α-naphthyl-
(phenyl)aminobutane-2,3-diol (3d): 82% yield; Rf ϭ 0.5 (eluent:
20
petroleum ether/ethyl acetate, 6:1); [α]D ϭ Ϫ35.2 (c ϭ 0.9,
1
CHCl3). Ϫ H NMR (200 MHz, CDCl3): δ ϭ 1.39 (s, 3 H, CH3),
1.41 (s, 3 H, CH3), 2.28 (dd, J ϭ 13.8, 6.0 Hz, 1 H, CH2P), 2.38
(dd, J ϭ 13.8, 7.0 Hz, 1 H, CH2P), 3.80Ϫ4.00 (m, 2 H, CH2N),
4.00Ϫ4.20 (m, 2 H, CHO), 6.90Ϫ7.60 (m, 22 H, Harom). Ϫ 31P (80
MHz, CDCl3): δ ϭ Ϫ23.9. Ϫ C35H34NO2P (531.64): calcd. C 78.51,
H 6.38; found C 78.54, H 6.50.
(2R,3S)-4-Dibenzylamino-1-diphenylphosphanyl-2,3-O-isopropyl-
idenebutane-2,3-triol (9a): 40% yield; Rf ϭ 0.22 (eluent: petroleum
(2S,3S)-2,3-O-Isopropylidene-1-O-tosyl-4-tosylaminobutane-1,2,3-
triol (10): 65% yield; Rf ϭ 0.7 (eluent: petroleum ether/ethyl acetate,
20
1
ether/ethyl acetate, 19:1); [α]D ϭ Ϫ21.8 (c ϭ 0.9, CHCl3). Ϫ H
NMR (200 MHz, CDCl3): δ ϭ 1.19 (s, 3 H, CH3), 1.29 (s, 3 H,
CH3), 2.16 (ddd, J ϭ 13.9, 7.7, 1.4 Hz, 1 H, CH2P), 2.22 (ddd, J ϭ
13.9, 4.5, 1.9 Hz, 1 H, CH2P), 2.47 (dd, J ϭ 13.7, 5.5 Hz, 1 H,
CH2N), 2.60 (dd, J ϭ 13.7, 4.3 Hz, 1 H, CH2N), 3.46 (d, J ϭ
13.6 Hz, 2 H, CH2N), 3.61 (d, J ϭ 13.6 Hz, 2 H, CH2N), 3.60Ϫ3.75
(m, 1 H, CHO), 3.90Ϫ4.00 (m, 1 H, CHO), 7.10Ϫ7.30 (m, 20 H,
20
1:1); [α]D ϭ ϩ5.6 (c ϭ 1.1, CHCl3). Ϫ 1H NMR (200 MHz,
CDCl3): δ ϭ 1.22 (s, 3 H, CH3), 1.26 (s, 3 H, CH3), 2.37 (s, 3 H,
CH3), 2.39 (s, 3 H, CH3), 2.90Ϫ3.20 (m, 2 H, CH2N), 3.80Ϫ4.10
(m, 6 H, CH2O, CHO), 4.64 (t, J ϭ 6.5 Hz, 1 H, NH), 710Ϫ7.30
(m, 5 H, Harom), 7.60Ϫ7.80 (m, 5 H, Harom), 7.71 (d, J ϭ 8.3 Hz,
2 H, Harom). Ϫ C21H27NO7S2 (469.57): calcd; C 53.72, H 5.80;
found C 53.20, H 5.84.
H
arom). Ϫ 13C NMR (50 MHz, CDCl3): δ ϭ 27.2 (CH3), 27.3
(CH3), 32.4 (d, 15.0 Hz, CH2P), 54.8 (CH2N), 59.0
J
ϭ
(2 ϫ CH2N), 77.3 (d, J ϭ 13.1 Hz, CHO), 80.7 (d, J ϭ 9.0 Hz,
CHO), 108.8 (CMe2), 127.0Ϫ139.3 (Carom). Ϫ 31P (80 MHz,
CDCl3): δ ϭ Ϫ23.1. Ϫ C33H36NO2P (509.63): calcd. C 77.78, H
7.12, N 2.75, P 6.08; found C 77.34, H 7.12, N 2.90, P 6.84.
General Procedure for Preparation of Amino Phosphanes 3 and 9: A
solution of PPh2Li, prepared under argon in THF (80 mL) at 0°C
from ClPPh2 (8.5 g, 38.5 mmol) and Li (10 equiv.), was slowly ad-
ded to amino tosylate 2 or 8 (12.6 mmol), dissolved in THF
(20 mL), until the red color of the solution remained. The solution
was stirred for 2 h and was then hydrolyzed with H2O (50 mL).
The solvent was evaporated and the residue was treated with H2O
(50 mL) and CH2Cl2 (3 ϫ 50 mL). Evaporation of the solvent, fol-
lowed by flash chromatography gave the amino phosphane 3 or 9.
(2R,3S)-4-Bis(2,4,6-trimethylbenzyl)amino-1-diphenylphosphanyl-
2,3-O-isopropylidenebutane-2,3-triol (9b): 45% yield; Rf ϭ 0.6 (elu-
20
ent: petroleum ether/ethyl acetate, 9:1); [α]D ϭ Ϫ1.6 (c ϭ 1.3,
1
CHCl3). Ϫ H NMR (200 MHz, CDCl3): δ ϭ 1.19 (s, 3 H, CH3),
1.31 (s, 3 H, CH3), 2.00Ϫ2.30 (m, 2 H, CH2P), 2.23 (s, 6 H, CH3),
2.25 (s, 12 H, CH3), 2.45Ϫ2.60 (m, 2 H, CH2N), 3.42 (d, J ϭ
12.4 Hz, 2 H, CH2N), 3.63 (d, J ϭ 12.4 Hz, 2 H, CH2N), 3.72Ϫ3.90
(m, 2 H, CHO), 7.30Ϫ7.70 (m, 10 H, Harom). Ϫ 31P (80 MHz,
CDCl3): δ ϭ Ϫ23.6. Ϫ C39H48NO2P (593.79): calcd. C 78.89, H
(2R,3S)-1-Diphenylamino-4-diphenylphosphanyl-2,3-O-isoprop-
ylidenebutane-2,3-diol (3a): 40% yield; Rf ϭ 0.70 (eluent: petroleum
20
ether/ethyl acetate, 6:1); [α]D ϭ Ϫ11.3 (c ϭ 1.0, CHCl3). Ϫ 1H
NMR (200 MHz, CDCl3): δ ϭ 1.33 (s, 3 H, CH3), 1.39 (s, 3 H,
CH3), 2.35Ϫ2.45 (m, 2 H, CH2P), 3.88 (dd, J ϭ 13.6, 6.7 Hz, 1 H, 8.15; found C 78.25, H 8.03.
Eur. J. Inorg. Chem. 2000, 351Ϫ358
357

F. Robert, F. Delbecq, C. Nguefack, D. Sinou
FULL PAPER
Asymmetric Catalysis in Organic Synthesis, Wiley, New York,
1994.
(2R,3S)-4-Bis(2-naphtylmethyl)amino-1-diphenylphosphanyl-2,3-O-
isopropylidenebutane-2,3-triol (9c): 20% yield; Rf ϭ 0.22 (eluent:
[2] [2a]
[2b]
A. Pfaltz, Acc. Chem. Res. 1993, 26, 339Ϫ345. Ϫ
O.
20
petroleum ether/ethyl acetate, 7:1); [α]D ϭ Ϫ28.6 (c ϭ 0.6,
Reiser, Angew. Chem. 1993, 105, 576Ϫ578; Angew. Chem. Int.
1
[2c]
CHCl3). Ϫ H NMR (200 MHz, CDCl3): δ ϭ 1.29 (s, 3 H, CH3),
Ed. Engl. 1993, 32, 547Ϫ549. Ϫ
B. M. Trost, D. L. Van
Vranken, Chem. Rev. 1996, 96, 395Ϫ422. Ϫ [2d] J. M. J. Williams,
Synlett 1996, 705Ϫ710.
1.36 (s, 3 H, CH3), 2.10Ϫ2.30 (m, 2 H, CH2P), 2.71Ϫ2.80 (m, 2 H,
CH2N), 3.62 (d, J ϭ 12.0 Hz, 2 H, CH2N), 3.73 (d, J ϭ 12.0 Hz,
2 H, CH2N), 4.05Ϫ4.20 (m, 2 H, CHO), 7.20Ϫ7.80 (m, 24 H,
Harom). Ϫ 31P (80 MHz, CDCl3): δ ϭ Ϫ23.5. Ϫ C41H40NO2P
(609.75): calcd. C 80.76, H 6.61; found C 80.25, H 6.03.
[3]
T. Hayashi, M. Konishi, M. Fukushima, K. Kanehira, T. Hioki,
M. Kumada, J. Org. Chem. 1983, 48, 2195Ϫ2202.
[4] [4a]
T. Hayashi, M. Tajika, K. Tamao, M. Kumada, J. Am.
Chem. Soc. 1976, 98, 3718Ϫ3719. Ϫ [4b] T. Hayashi, M. Konishi,
M. Fukushima, T. Mise, M. Kagotani, M. Tajika, M. Kumada,
J. Am. Chem. Soc. 1982, 104, 180Ϫ186.
(2R,3S)-4-(6,7-Dihydro-5H-dibenzo[c,e]azepin-6-yl)-1-diphenyl-
phosphanyl-2,3-O-isopropylidenebutane-2,3-triol (9d): 80% yield,
[5] [5a]
H. Kubota, K. Koga, Tetrahedron Lett. 1994, 35,
[5b]
6689Ϫ6694. Ϫ
42, 543Ϫ547.
H. Kubota, K. Koga, Heterocycles 1996,
1
Rf ϭ 0.32 (eluent: petroleum ether/ethyl acetate, 5:1). Ϫ H NMR
[6] [6a]
(200 MHz, CDCl3): δ ϭ 1.30 (s, 3 H, CH3), 1.38 (s, 3 H, CH3),
2.31 (m, 2 H, CH2P), 2.72 (m, 2 H, CH2N), 3.32 (d, J ϭ 12.3 Hz,
2 H, CH2N), 3.38 (d, J ϭ 12.3 Hz, 2 H, CH2N), 4.00Ϫ4.20 (m, 2
H, CHO), 7.20Ϫ7.60 (m, 18 H, Harom). Ϫ 13C NMR (50 MHz,
CDCl3): δ ϭ 27.4 (CH3), 27.5 (CH3), 32.8 (d, J ϭ 16.3 Hz, CH2P),
55.9 (2 ϫ CH2N), 57.4 (CH2N), 77.6 (d, J ϭ 14.8 Hz, CHO), 80.7
(d, J ϭ 7.8 Hz, CHO), 109.3 (CMe2), 127.6Ϫ141.3 (Carom). Ϫ 31P
(80 MHz, CDCl3): δ ϭ Ϫ23.8. Ϫ C33H34NO2P (507.62): calcd. C
78.08, H 6.75; found C 79.10, H 6.59.
J. P. Cahill, F. M. Bohnen, R. Goddard, C. Krüger, P. J.
Guiry, Tetrahedron: Asymmetry 1998, 9, 3831Ϫ3839. Ϫ [6b] J. P.
Cahill, P. J. Guiry, Tetrahedron: Asymmetry 1998, 9,
[6c]
4301Ϫ4306. Ϫ
J. P. Cahill, A. P. Lighfoot, R. Goddard, J.
Rust, P. J. Guiry, Tetrahedron: Asymmetry 1998, 9, 4307Ϫ4312.
[7] [7a]
P. Wimmer, M. Widhalm, Tetrahedron: Asymmetry 1995, 6,
[7b]
657Ϫ660. Ϫ
P. Wimmer, M. Widhalm, Monatsh. Chem.
[7c]
1996, 127, 669Ϫ681. Ϫ
M. Boughida, M. Widhalm, Tetra-
hedron: Asymmetry 1998, 9, 1073Ϫ1083. Ϫ [7d] M. Widhalm, K.
Mereiter, M. Bourghida, Tetrahedron: Asymmetry 1998, 9,
2983Ϫ2986.
[8]
[9]
F. Robert, N. Gaillard, D. Sinou, J. Mol. Catal., in press.
W. R. Jackson, C. G. Lovel, Aust. J. Chem. 1982, 35,
2069Ϫ2075.
(2R,3S)-1-Diphenylphosphanyl-2,3-O-isopropylidene-4-tosylamino-
butane-2,3-diol (11): 73% yield; Rf ϭ 0.25 (eluent: petroleum ether/
[10]
20
ethyl acetate, 4:1); [α]D ϭ Ϫ4.2 (c ϭ 0.3, CHCl3). Ϫ 1H NMR
U. Leutenegger, G. Umbricht, C. Fahrni, P. von Matt, A. Pfaltz,
Tetrahedron 1992, 48, 2143Ϫ2156.
(200 MHz, CDCl3): δ ϭ 1.28 (s, 3 H, CH3), 1.34 (s, 3 H, CH3),
2.42 (s, 3 H, CH3), 2.30Ϫ2.50 (m, 2 H, CH2P), 3.00Ϫ4.20 (m, 2 H,
CH2N), 4.79 (t, J ϭ 6.1 Hz, 1 H, NH), 3.80Ϫ4.20 (m, 2 H, CHO),
7.20Ϫ7.40 (m, 10 H, Harom), 7.42 (d, J ϭ 8.2 Hz, 2 H, Harom), 7.83
(d, J ϭ 8.2 Hz, 2 H, Harom). Ϫ 31P (80 MHz, CDCl3): δ ϭ Ϫ24.6.
Ϫ C26H30NSO4P (483.57): calcd. C 64.58, H 6.25, N 2.89; found
C 63.96, H 6.09, N 2.87.
[11] [11a] B. E. R. Schilling, R. Hoffmann, J. W. Faller, J. Am. Chem.
[11b]
Soc. 1979, 101, 592Ϫ598. Ϫ
1996, 15, 2836Ϫ2838.
T. R. Ward, Organometallics
[12]
[13]
[14]
M. D. Curtis, O. Eisenstein, Organometallics 1984, 3, 887Ϫ895.
B. Bigot, F. Delbecq, New J. Chem. 1990, 14, 659Ϫ669.
S. Sakaki, M. Nishikawa, A. Ohyoshi, J. Am. Chem. Soc. 1980,
102, 4062Ϫ4069.
[15]
[16]
P. E. Blöchl, A. Togni, Organometallics 1996, 15, 4125Ϫ4132.
`
K. J. Szabo, E. Hupe, A. L. E. Larsson, Organometallics 1997,
(2S,3S)-1,4-Bis(diphenylamino)-2,3-O-isopropylidenebutane-2,3-diol
16, 3779Ϫ3785.
[17] [17a]
[17b]
20
(12): [α]D ϭ Ϫ5.2 (c ϭ 0.8, CHCl3). Ϫ 1H NMR (300 MHz,
`
K. J. Szabo, J. Am. Chem. Soc. 1996, 118, 7818Ϫ7826. Ϫ
[17c]
K. J. Szabo, Chem. Eur. J. 1997, 3, 592Ϫ600. Ϫ
K. J.
CDCl3): δ ϭ 1.40 (s, 6 H, CH3), 3.85 (d, J ϭ 4.8 Hz, 4 H, CH2),
4.20 (m, 2 H, CH), 6.90Ϫ7.02 (m, 12 H, Harom), 7.18Ϫ7.32 (m, 8
H, Harom). Ϫ 13C NMR (75 MHz, CDCl3): δ ϭ 27.5 (CH3), 55.7
(CH2), 77.8 (s, CH), 109.7 (CMe2), 121.5, 121.8, 129.4 and 148.2
(C6H5). Ϫ C31H32N2O2 (464.61): calcd. C 80.14, H 6.94, N 6.03;
found C 80.08, H 6.85, N 6.15.
Szabo, Organometallics 1996, 15, 1128Ϫ1133.
[18]
[19]
[20]
S. Sakaki, H. Satoh, H. Shono, Y. Ujino, Organometallics 1996,
15, 1713Ϫ1720.
S. Sakaki, K. Takeuchi, M. Sugimoto, Organometallics 1997,
16, 2995Ϫ3003.
A. Aranyos, K. J. Szabo, A. M. Castano, J. E. Bäckvall, Or-
ganometallics 1997, 16, 1058Ϫ1064.
[21]
[22]
[23]
A. E. Smith, Acta Crystallogr., Sect. A 1969, 25, S161.
R. Mason, D. R. Russel, Chem. Commun. 1966, 26.
S. Liu, J. F. K. Müller, M. Neuburger, S. Shaffner, M. Zehnder,
J. Organomet. Chem. 1997, 549, 283Ϫ293.
R. Hoffmann, J. Chem. Phys. 1963, 39, 1397Ϫ1412.
O. Eisenstein and R. Hoffmann, J. Am. Chem. Soc. 1981, 101,
4308Ϫ4320.
J. H. Ammeter, H. B. Burgi, J. C. Thibeault, R. Hoffmann, J.
Am. Chem. Soc. 1978, 100, 3686.
Asymmetric Alkylation. ؊ Typical Procedure: A solution of the so-
dium salt of dimethyl malonate, prepared from dimethyl malonate
(118.9 mg, 0.900 mmol) and sodium hydride in THF (2 mL) at
room temperature, was added to a mixture of Pd2(dba)3 (55 mg,
6.10Ϫ3 mmol), ligand (12.10Ϫ3 mmol) and rac-(E)-1,3-di-
phenylprop-2-enyl acetate (75.7 mg, 0.300 mmol) in THF (2 mL).
The solution was stirred for 24 h and was then quenched with a
small amount of water. Evaporation of the solvent, followed by
column chromatography gave the alkylated product. The conver-
sion and enantiomeric purity was determined by HPLC with a chi-
ral stationary phase column Chiraldex ODϪH (hexane/2-propa-
nol ϭ 98:2). The absolute configuration was determined by polari-
metry.[10]
[24]
[25]
[26]
[27]
M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.
Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Peter-
son, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V.
G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B.
Stefanov, A. Nanayakkara, M. Challlacombe, C. Y. Peng, P. Y.
Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S. Replogle, R.
Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees,
J. Baker, J. P. Stewart, M. Head-Gordon, C. Gonzalez, J. A.
Pople, Gaussian 94, Revision B.1, Gaussian, Inc., Pittsburgh
PA, 1995.
[28]
T. G. Appleton, H. C. Clark, L. E. Manzer, Coord. Chem. Rev.
1973, 10, 335Ϫ422.
Acknowledgments
[29]
[30]
A. E. Smith, Acta Crystallogr. 1965, 18, 331Ϫ340.
P. Sautet, O. Eisenstein, K. M. Nicholas, Organometallics 1987,
6, 1845Ϫ1849.
´
ˆ
R. F. thanks the MESRT and Region Rhone-Alpes for a fellowship.
[31]
S. Brunie, J. Mazan, N. Langlois, H. B. Kagan, J. Organomet.
Chem. 1976, 114, 225Ϫ232.
[1]
[1a] H. Brunner, W. Zettlmeier, Handbook of Enantioselective Ca-
[1b]
talysis, VCH, Weinheim, 1993. Ϫ
I. Ojima, Catalytic Asym-
Received July 26, 1999
[1c]
metric Synthesis, VCH, Weinheim, 1993. Ϫ
R. Noyori,
[I99281]
358
Eur. J. Inorg. Chem. 2000, 351Ϫ358