Chiral aminophosphonite-phosphite ligands: Synthesis and use in platinum- and rhodium-based enantioselective hydroformylation of styrene. Crystal structure of [(S)-N-(2,2′biphenoxyphosphino)-2-(2,2′biphenoxyphosphinoxymethyl) pyrrolidine]dichloroplatinum(II)

Article abstract of DOI:10.1016/S0022-328X(01)00768-9

New chiral unsymmetrical ligands containing deficient phosphorous moieties based on (S)-prolinol and (S)-oxoprolinol have been synthesised and used as chiral auxiliaries for the platinum- and rhodium-catalysed enantioselective hydroformylations of styrene. The new ligands are obtained readily by reaction of the amino alcohols with a chlorophosphonite. The beneficial effect induced by the electron-deficient phosphorous ends provided the branched aldehyde in up to 44% ee. A crystal structure of the platinum complex 5 is described.

Full text of DOI:10.1016/S0022-328X(01)00768-9

Journal of Organometallic Chemistry 628 (2001) 114–122
www.elsevier.nl/locate/jorganchem
Chiral aminophosphonite-phosphite ligands: synthesis and use in
platinum- and rhodium-based enantioselective hydroformylation of
styrene. Crystal structure of [(S)-N-(2,2%biphenoxyphosphino)-
2-(2,2%biphenoxyphosphinoxymethyl)pyrrolidine]-
dichloroplatinum(II)
Said Naili a, Isabelle Suisse a, Andre´ Mortreux a, Francine Agbossou-Niedercorn a,*,
Guy Nowogrocki b
a Laboratoire de Catalyse de Lille, Groupe de Chimie Organique Applique´e, ESA 8010, Ecole Nationale Supe´rieure de Chimie de Lille,
B.P. 108-59652 Villeneu6e d’Ascq Cedex, France
b Laboratoire de Cristallochimie et de Physicochimie du Solide, ESA 8012, Ecole Nationale Supe´rieure de Chimie de Lille,
B.P. 108-59652 Villeneu6e d’Ascq Cedex, France
Received 5 January 2001; accepted 5 March 2001
Abstract
New chiral unsymmetrical ligands containing deficient phosphorous moieties based on (S)-prolinol and (S)-oxoprolinol have
been synthesised and used as chiral auxiliaries for the platinum- and rhodium-catalysed enantioselective hydroformylations of
styrene. The new ligands are obtained readily by reaction of the amino alcohols with a chlorophosphonite. The beneficial effect
induced by the electron-deficient phosphorous ends provided the branched aldehyde in up to 44% ee. A crystal structure of the
platinum complex 5 is described. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Platinum; Rhodium; Hydroformylation; Diphosphine; Phosphonite–phosphite
rhodium-based asymmetric hydroformylation involving
chiral bidendate phosphine–phosphites [4] or diphos-
1. Introduction
phites [5] allowed access to highly regio and enantiose-
lecti6e processes. However, until the breakthrough of
Takaya and co-workers with the phosphine–phosphite-
Asymmetric hydroformylation of olefins is a very
attractive process for the synthesis of optically pure
aldehydes. The latter are useful as building blocks for
Rh(I) catalysts [4a], the diphosphane-platinum(II) cata-
the preparation of pharmaceuticals, agrochemicals and
lysts promoted by SnCl2 were the most popular as they
food additives [1]. The 2-arylpropionic acids, readily
obtained by oxidation of the corresponding chiral alde-
led to the highest enantioselectivities [3,6] even if their
utilisation is often accompanied by low regioselectivities
hydes, which constituteaclassofnonsteroidal anti-inflam-
to the branched aldehydes and low reaction rates.
matory drugs with substantial pharmaceutical and com-
We and others [7] have already reported on the
mercial impact [2] are very good examples. Platinum
synthesis, characterisation, and use of aminophos-
and rhodium complexes modified by chiral phosphanes
phine–phosphinites (AMPP) in Rh- and Pt-catalysed
have been the most-employed hydroformylation cata-
styrene hydroformylations. The interest of such ligands
lysts [3]. Importantly, the recent development of
lies in their unsymmetrical nature, which offers distinct
advantages from those of the symmetrical auxiliaries.
Considering the specific ability presented by the bispho-
sphite-based catalytic systems mentioned above, we
sought to synthesise new ligands derived from amino
* Corresponding author. Tel.: +33-3-20434927; fax: +33-3-
20436585.
E-mail address: agbossou@ensc-lille.fr (F. Agbossou-Niedercorn).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)00768-9

S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
115
alcohols, including such electron-deficient — yet achi-
ral — phosphorous moieties. Actually, chiral symmet-
rical diphosphite–platinum catalysts have been
reported recently providing as high as 91% ee in the
hydroformylaton of styrene [6e]. Here, we wish to
present our results on the synthesis of two new ho-
mochiral aminophosphonite–phosphite chelates pos-
sessing bisphenol moieties on the two phosphorous
atoms and on their use in enantioselective hydroformy-
lation of the model substrate styrene. We also describe
an X-ray structure of a platinum complex chelated by
one of these new ligands.
paralleled the route developed for the access to AMPPs.
First, (S)-2-(hydroxymethyl)pyrrolidine (1) was reacted
with 2,2%-bisphenoxyphosphorous chloride (BPPCl) [9]
in the presence of triethylamine in tetrahydrofuran
(THF) at 60°C (Scheme 1).
The ligand formation was monitored by 31P-NMR
spectroscopy. Indeed, the signal of the starting chloride
at 178 ppm decreased in favour of the ligand signals.
Thus, after 18 h stirring at 60°C, two major peaks
remained at 150 and 138.7 ppm attributed to the P(O)
and P(N) phosphorous atoms of (S)-BP-ProNOP (3).
Then, the triethylamine hydrochloride formed was sep-
arated under nitrogen via a simple filtration through
basic alumina providing ligand 3 after solvent removal
as a white powder in 88% yield. Ligand 3 afforded
satisfactory elemental analysis and was characterised by
NMR spectroscopy. Data are summarised in Section 4.
Since we suspected an inversion of the relative position
of the two 31P-NMR signals [10],1 we carried out a 2D
2. Results and discussion
2.1. Synthesis and characterisation of the ligands and
platinum complexes
1
From our studies on the AMPP ligands, two chiral
precursors have appeared to be particularly suited for
their potential as chiral auxiliaries for enantioselective
processes, i.e. (S)-2-(hydroxymethyl)pyrrolidine (1) and
(S)-2-(hydroxymethyl)pyrrolidinone (2) (Scheme 1) [8].
Consequently, the latter have been selected for investi-
gation in the synthesis of new ligands for asymmetric
hydroformylation. The synthesis of the new ligands
31P, H experiment. Generally, in agreement with the
relative electronegativities of the O and N atoms, in the
31P-NMR spectra, the P(O) signal is downfield from
that of the P(N) one. The 2D 31P-, 1H-NMR cross-
peaks obtained with both phosphorous atoms and the
protons of the chiral skeleton corroborate our assign-
ment of the 31P-NMR signals. Indeed, cross-peaks are
obtained between the P(O) resonance at 150 ppm and
the protons of the NꢀCH1ꢀCH2H3O chain (l 3.85
(H1), 3.85 (H2), and 3.67 (H3)) and the protons of the
bisphenoxy moiety (l 6.95–7.39), whereas the P(N) at
138.7 ppm correlates with the protons on the carbon
atoms adjacent to the nitrogen atom CH2NCH at l
3.85 and with the aryl protons of the bisphenoxy
residue. In addition, a coupling is observed between
both phosphorous atoms and the stereogenic carbon
atom as the signal of the latter appears as a doublet of
doublets (JCP=25 and 3 Hz) in the 13C{1H}-NMR
spectrum of 3.
In a similar manner, the precursor (S)-2-(hydrox-
ymethyl)pyrrolidinone (2) has been reacted with BPPCl
in order to provide (S)-BP-oxoProNOP (4) (Scheme 1).
The workup was carried out as above, furnishing 4 as a
white powder in 86% yield. Derivative 4 presents 31P-
NMR signals as singlets at 145 (P(O)) and 137 ppm
(P(N)). However, ligand 4 is less stable than 3 and only
a small pad of alumina could be used for the filtration
used for the purification. Indeed, the PꢀN bond is more
fragile in 4 than in 3 and the monosubstituted amino-
phosphite derivative HNꢀOP is produced easily. This
substantial fragility did not allow the isolation of an
analytically pure 4. Nevertheless, the purity was high
enough (\95% by 31P-NMR) for an evaluation in
catalysis.
Scheme 1.
1 Such a trend has been observed for ephedrine-based aminophos-
Scheme 2.
phonite–phosphite ligands

116
S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
Table 1
of metal complex, when changing from a phosphine to
a phosphite ligand, a variation in J(M,P) of between 50
and 100% is observed [11,12]. For complex 5 (and 6
Crystal data and structure refinement parameters for [PtCl2{(S)-BP-
ProNOP}] (5·CH2Cl2)
1
below) it is thus not surprising to observe also a
significant increase of J(PtꢀP). In addition, a parallel
Empirical formula
Formula weight
Temperature (K)
Colour and habit
Crystal system
C29H25Cl2NO5P2Pt·CH2Cl2
880.3866
295
Colourless prism
Orthorhombic
P212121 (SG: 19)
1
shortening of the PtꢀP bonds is foreseen and was,
indeed, observed (see the X-ray characterisation below).
Ordinarily, a deshielding is observed for both the P(O)
and P(N) phosphorous atoms upon coordination of
AMPP ligands to rhodium [8b], palladium [13] or
ruthenium [14]. In the case of platinum complexes,
commonly, the P(O) signal is upfield from that of the
free ligands and the P(N) slightly downfield [7c,d]. It is
noteworthy to mention that, here, both phosphorous
resonances of the ligand in complex 5 appear ca. 60
ppm upfield from that of the free ligand. These shield-
ings are in the range usually observed for phosphite-
type complexes when compared to the phosphine-based
platinum species [15].
Space group
Unit cell dimensions
,
a (A)
12.822(4)
13.364(2)
18.382(3)
3150
10
4
1.674
1720
0.30×0.20×0.20
49.9
2–28
ꢀ/2q
,
b (A)
,
c (A)
3
,
V (A )
q range for acc. cell (°)
Z
Dcalc (g cm−3
F(000)
)
Crystal size (mm)
v(Mo–Ka) (cm−1
)
q range for data collection (°)
Scan type
Crystal description
Number of data collected
Number of unique data
hkl Ranges
Irregular, ovoid
8184
7241
0.16, 0.17, −23.23
0.044
(46 10); (16 31); (46 10)
3
5571
407
0.0390
0.045
The reaction of Pt(COD)Cl2 with 4 provided Pt{(S)-
BP-oxoProNOP}Cl2 (6) quantitatively. The latter ex-
hibits 31P-NMR signals as two doublets at 90 and 86
ppm (2J(P,P)=4.2 Hz) accompanied by the 195Pt satel-
lites, the J(Pt,P(O)) and J(Pt,P(N)) being 5658 and 5833
Hz, respectively. Both complexes 5 and 6 are stable in
air in the solid state and in solution.
Rmerge
1
1
Standard reflections
Observability criterion, n [I\n|(I)]
Number of data in refinement
Number of refined parameters
Final R
2.2. Crystal structure of 5
Rw
Largest difference peak and hole
2.1 and −1.5
−3
,
(e A
)
Slow crystallisation from dichloromethane of the
crude complex 5 afforded colourless crystals, which
proved suitable for an X-ray investigation. X-ray data
were collected under the conditions summarised in
Table 1 and atomic coordinates are given in Table 2.
Refinement described in Section 4 gave the structure
shown in Fig. 1 together with the numbering scheme
adopted, and showed the product to be a CH2Cl2
monosolvate monomeric complex. Selected key bond
lengths, bond angles, and deviations from the mean
plane PtꢀClꢀClꢀP(O)ꢀP(N) are listed in Table 3.
Next, the platinum catalytic precursors Pt{(S)-BP-
ProNOP}Cl2 (5) and Pt{(S)-BP-oxoProNOP}Cl2 (6)
were synthesised in dichloromethane at room tempera-
ture by reacting Pt(COD)Cl2 with an equimolar amount
of ligands 3 and 4, respectively (Scheme 2).
The synthesis of 5 proceeded very readily at room
temperature and was almost quantitative, providing,
after workup, an off-white air-stable solid. The charac-
terisation of 5 was carried out through elemental analy-
sis, NMR spectroscopy and X-ray crystallographic
analysis. Data are summarised below and in Section 4.
The 31P-NMR signal trend established a chelation of
ligand 3 to the platinum centre. In fact, the P(O) and
P(N) resonances consist of two sets of doublets of
similar intensity at 85 and 72 ppm. These main signals
are supplemented with the two doublets of doublets
constituting the 195Pt satellites. The large values of the
The cis geometry was established by the X-ray data.
As evidenced from the deviations from the mean plane
PtꢀClꢀClꢀP(O)ꢀP(N), complex 5·CH2Cl2 adopts a con-
formation around Pt close to that of an idealised square
,
planar geometry (average deviation 0.010(2) A). Com-
plex 5·CH2Cl2 exhibits PtꢀP(O) and PtꢀP(N) lengths
,
(2.190(3) and 2.202(2) A) distinctly shorter than those
in other structurally characterised AMPP complexes by
,
an average of 0.027 and 0.029 A, respectively
[7c,d,16,17]. These two characteristics, as mentioned
above, are a consequence of the substitution of the
phosphorous atoms by the electronegative bisphenoxy
moiety. Complex 5·CH2Cl2 exhibits equal PtꢀCl bond
1
1J(PtꢀP), i.e. 5392 Hz for J(Pt,P(O)) and 5792 Hz for
2
1J(Pt,P(N)) associated to the small J(P,P) value (28 Hz) are
consistent with a cis geometry [11,12]. In addition, the
PtꢀP coupling constants are greater than those ob-
served for the AMPP analogues by an average of 1500
Hz of magnitude. It is well known that for a given type
,
lengths (2.331 A) and similar ClꢀPtꢀP angles (89.76 (11)
and 91.51(11)°), indicating that the two BP adducts
induce the same steric hindrance around platinum.

S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
117
There is only a very small widening of the bite angle
P(O)ꢀPtꢀP(N) (91.66(11)°). The latter is overall smaller
than those in other AMPPꢀPt complexes [16]. Other
geometric features were calculated. For example, the
out-of-plane distortion between the aryl rings in both
bisphenol units has been quantified. Thus, the angle
between the planes defined by the adjacent least-square
planes defined by the aryl carbons of the BP units are
38.7 and 38°, respectively, on the P(N) and P(O) sides
of the complex, showing thus a small distortion from
initial bisphenol planarity. The deviations of the aryl
carbon atoms from the corresponding planes were less
359.98°, which shows that the nitrogen atom has an sp2
hybridisation. The sums of the bond angles around
both P(N) and P(O) phosphorous atoms are 307.61 and
313.31°, respectively. These values indicate conforma-
tions intermediate between pyramidal and tetrahedral.
Seven-membered rings can adopt several boat and chair
conformations. Hence, some of them have already been
noticed for AMPP and related complexes, which is
indicative of the overall flexibility of the AMPP ligands
[16,17]. That flexibility potential proves that the AMPP
moieties can accommodate the steric requirement of the
metal and of the other ligands present in the coordina-
tion sphere. Complex 5·CH2Cl2 exhibits an almost pure
twist boat conformation with the oxygen atom of the
P(O) unit very close to the plane defined by
,
than 0.019 A. The sum of the angles around nitrogen is
Table 2
,
Atomic coordinates and equivalent isotropic displacement parameters
for [PtCl2{(S)-BP-ProNOP}] (5·CH2Cl2)
PtꢀClꢀClꢀP(O)ꢀP(N) (deviation 0.234(6) A) (Fig. 2).
The nitrogen atom of the P(N) residue deviates 1.286(6)
,
A from that plane. Such a metallacycle conformation is
generally observed in (S)-Ph,Ph-ProNOP complexes
2
,
Atom
x/a
y/b
z/c
Beq (A )
[7c,14].
Pt(1)
0.03082(2)
0.0044(2)
0.1788(2)
0.00225(3)
0.09258(1)
2.66(1)
4.81(7)
4.25(6)
2.44(4)
2.80(5)
3.3(1)
3.7(2)
3.3(2)
2.8(2)
3.6(2)
4.4(3)
4.6(3)
3.2(1)
3.2(1)
2.9(1)
2.8(1)
3.1(2)
4.2(2)
5.0(3)
5.7(3)
4.4(3)
3.3(2)
3.3(2)
4.7(3)
6.1(4)
5.6(3)
4.2(3)
3.0(2)
3.7(2)
4.6(3)
5.9(4)
5.6(3)
4.1(2)
3.1(2)
3.3(2)
3.8(2)
4.2(3)
4.2(3)
3.7(2)
2.8(2)
10.5(7)
13.7(8)
14.2(8)
Cl(2)
Cl(3)
P(4)
P(5)
O(6)
C(7)
C(8)
N(9)
−0.1349(2)
−0.0802(2)
0.1302(2)
0.0714(2)
0.1841(4)
0.2682(6)
0.2549(6)
0.2366(5)
0.3295(6)
0.3769(7)
0.3547(7)
0.0289(4)
0.0619(4)
0.1292(4)
0.1357(4)
0.1044(6)
0.0324(7)
0.0174(8)
0.0729(9)
0.1419(8)
0.1611(6)
0.2361(6)
0.3245(7)
0.3966(9)
0.3823(8)
0.2928(7)
0.2232(6)
−0.0694(7)
−0.1440(8)
0.1687(1)
0.0499(1)
0.0198(1)
0.1352(1)
0.1156(3)
0.1419(5)
0.1371(4)
0.0605(3)
0.0296(5)
0.0956(5)
0.01574(5)
0.1082(3)
0.2219(3)
−0.0304(3)
−0.0391(3)
−0.1049(4)
−0.1197(5)
−0.1962(6)
0.2483(65)
−0.2291(5)
−0.1568(4)
−0.1369(4)
−0.1772(5)
−0.1611(7)
−0.1052(6)
−0.0653(5)
−0.0821(4)
0.1331(5)
0.0798(5)
0.1008(7)
0.1721(7)
0.2222(6)
0.2025(4)
0.2565(4)
0.3064(4)
0.3578(5)
0.3600(5)
0.3120(4)
0.2623(4)
0.032(1)
2.3. Catalysis
0.0595(2)
−0.1114(2)
−0.1349(5)
−0.0752(8)
0.0423(7)
0.0756(5)
0.1211(8)
The two ligands 3 and 4 have been evaluated in the
platinum- and rhodium-based hydroformylations of
styrene (Table 4). The platinum catalysts were obtained
from the reaction of complexes 5 and 6 with one
equivalent of SnCl2·2H2O as a promoter. Rhodium
catalysts were prepared by mixing Rh4(CO)12 in toluene
with ligands 3 and 4 (Rh–ligand=1.5). The hydro-
formylation reactions were carried out under classical
reaction conditions (Table 4) and provided a mixture of
the branched (b) (2-phenylpropanal) and the linear
(3-phenylpropanal) (n) regioisomers along with the hy-
drogenated side-compound for platinum-based cataly-
ses (Scheme 3).
For platinum-based reactions, the activity of the
catalytic system obtained in this study was in the
average observed for platinum catalysts [3]. Neverthe-
less, compared to our own previously reported
AMPPꢀPt catalysts [7c], a beneficial effect on both the
activity and the regioselectivity was achieved. The latter
could be ascribed to the bisphenoxy moiety. For in-
stance, the mixture of branched and linear aldehydes
was produced in a b/n ratio ranging from 53/47 to
66/34, whereas earlier ligands provided only ratios in
the 30/70–43/57 range [7c]. In addition, turnover fre-
quencies varying from 15 to 38 h−1 were measured
here. These values have to be compared to those ac-
quired with (S)-Ph,Ph-ProNOP and (S)-Ph,Ph-oxo-
ProNOP ligands (Scheme 4). These two earlier reported
ligands provided lower turnover frequencies, i.e. 2.8
h−1 (b/n=39/59, ee=42%) and 5 h−1 (b/n=29/68,
ee=40%), respectively. These results show clearly that,
for platinum, the reaction rate increases with the lower-
ing of the electron density on the phosphorous moieties
C(10)
C(11)
C(12)
O(13)
O(14)
O(15)
O(16)
C(17)
C(18)
C(19)
C(20)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(27)
C(28)
C(29)
C(30)
C(31)
C(32)
C(33)
C(34)
C(35)
C(36)
C(37)
C(38)
C(39)
C(40)
C(41)
Cl(42)
Cl(43)
0.1744(8)
0.0958(9)
−0.2204(5)
−0.1159(4)
0.1609(4)
−0.0321(4)
0.1678(7)
0.2423(8)
0.2640(9)
0.2130(10)
0.1361(8)
0.1127(7)
0.0337(7)
0.0287(9)
−0.0472(10)
−0.1186(9)
−0.1124(8)
−0.0369(7)
−0.2519(7)
−0.2505(8)
−0.2851(10) −0.2371(8)
−0.3220(9)
−0.3231(8)
−0.2863(7)
−0.2865(6)
−0.3714(6)
−0.3715(8)
−0.2910(9)
−0.2086(8)
−0.2070(6)
0.002(1)
−0.2528(8)
−0.1743(6)
−0.0796(6)
0.0017(8)
0.0093(9)
0.0869(8)
0.1579(8)
0.1510(7)
0.0721(6)
0.686(1)
−0.053(1)
−0.008(2)
0.672(2)
0.582(2)
−0.043(1)
0.077(1)

118
S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
Fig. 1. Solid state structure and atom numbering of 5. Hydrogen atoms are omitted for clarity.
of the ligand. Still, the branched to linear ratios were
generally low with close to equal amounts of both
regioisomers. Complex 6 provides a higher b/n ratio but
with a nevertheless lower enantioselectivity into the
branched aldehyde (15% ee) (run 8). As usually ob-
served, ethylbenzene was formed along with the hydro-
formylation products during the course of the reaction.
Unfortunately, for these two catalytic systems, this
side-reaction is rather important with an average of
35% of ethylbenzene formed. The two solvents used, i.e.
dichloromethane and toluene, did not have much effect
on the properties of the catalyst (runs 1 and 2). Thus,
because of the higher solubility of the starting complex
in dichloromethane, the precatalyst was prepared in
that solvent prior to the catalytic reaction. As pre-
sumed, lowering the partial pressure of hydrogen led to
a smaller amount of hydrogenation (run 4). The highest
asymmetric induction (44% ee) was found when the
temperature was lowered to room temperature (run 7).
This study proves that one chiral source is sufficient in
the ligand to induce enantioselection but the latter is
too low for an application. In addition to that single
chirality, this family of ligands presents an intrinsic
dissymmetry due to its chiral pool origin and the catal-
ysis certainly benefits from that property.
Table 3
Selected bond lengths (A), bond angles (°), and deviations from the
,
,
Pt(1)ꢀCl(2)ꢀCl(3)ꢀP(4)ꢀP(5) mean plane (A)
Bond length
Pt(1)ꢀCl(3)
Pt(1)ꢀP(4)
P(4)ꢀN(9)
2.331(3)
2.202(2)
1.620(7)
1.595(6)
1.600(6)
1.492(9)
1.485(11)
1.412(9)
Pt(1)ꢀCl(2)
Pt(1)ꢀP(5)
P(5)ꢀO(6)
P(5)ꢀO(13)
P(5)ꢀO(14)
O(6)ꢀC(7)
O(13)ꢀC(29)
O(14)ꢀC(40)
2.331(2)
2.190(3)
1.578(6)
2.190(3)
1.589(7)
1.444(11)
1.449(11)
1.392(9)
1.472(12)
1.472(12)
1.521(14)
P(4)ꢀO(15)
P(4)ꢀO(16)
N(9)ꢀC(8)
N(9)ꢀC(10)
O(15)ꢀC(17)
O(16)ꢀC(28)
C(10)ꢀC(11)
C(11)ꢀC(12)
C(12)ꢀC(8)
1.413(C10) C(22)ꢀC(23)
1.530(13)
1.548(14)
1.546(13)
C(34)ꢀC(35)
C(7)ꢀC(8)
Bond angles
Cl(2)ꢀPtꢀCl(3)
Cl(2)ꢀPtꢀP(5)
87.07(10)
89.76(11)
118.71(27)
118.09(26)
95.83(39)
114.98(26)
108.91(21)
111.47(38)
122.91(45)
109.67(65)
P(4)ꢀPtꢀP(5)
91.66(11)
91.51(11)
110.66(21)
107.24(37)
Cl(3)ꢀPtꢀP(4)
Pt(1)ꢀP(5)ꢀO(14)
O(6)ꢀP(5)ꢀO(14)
O(14)ꢀP(5)ꢀO(13) 104.57(37)
Pt(1)ꢀP(4)ꢀO(15)
N(9)ꢀP(4)ꢀO(15)
O(15)ꢀP(4)ꢀO(16) 102.02(35)
P(4)ꢀN(9)ꢀC(10) 127.40(52)
P(4)ꢀO(15)ꢀC(17) 127.9(52)
P(5)ꢀO(6)ꢀC(7) 124.43(50)
P(5)ꢀO(13)ꢀC(29) 118.10(49)
Pt(1)ꢀP(5)ꢀO(6)
Pt(1)ꢀP(5)ꢀO(13)
O(6)ꢀP(5)ꢀO(13)
Pt(1)ꢀP(4)ꢀN(9)
Pt(1)ꢀP(4)ꢀO(16)
N(9)ꢀP(4)ꢀO(16)
P(4)ꢀN(9)ꢀC(8)
C(10)ꢀN(9)ꢀC(8)
118.77(23)
99.82(36)
On the other hand, for rhodium-based hydroformyla-
tions, compared to the previously reported ephedrine-
based aminophosphonite–phosphite ligands [7e,10],
modest regioselectivities in favour of the branched alde-
hyde and low enantioselectivities were obtained (runs
9–11). In addition, the reaction rates are significantly
lower, too. Thus, not surprisingly, bisphenoxy moieties
are not good substituents for phosphorous atoms in
ligands designed for rhodium-catalysed enantioselective
hydroformylations.
P(4)ꢀO(16)ꢀC(28) 116.66(46)
P(5)ꢀO(14)ꢀC(40) 123.65(45)
De6iations from the mean plane
Pt(1)
Cl(2)
Cl(3)
P(4)
P(5)
O(6)
0.007(0)
0.010(2)
0.013(2)
0.009(2)
−0.013(2)
0.234(6)
N(9)
1.286(6)
−1.377(6)
1.037(5)
0.050(5)
−1.368(5)
O(13)
O(14)
O(15)
O(16)

S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
119
Scheme 3.
Fig. 2. View of the structure of 5 presenting the seven-membered
metallacycle in a Newman-type projection.
Scheme 4.
4. Experimental
4.1. General procedures
3. Conclusions
All reactions and manipulations were carried out
under an atmosphere of dry N2 using standard Schlenk-
type glassware. Toluene and THF were distilled from
benzophenone ketyl. Dichloromethane was distilled
from calcium hydride. All solvents were freeze–thaw-
degassed (three times) prior to use. The starting com-
pounds Pt(COD)Cl2 [18] and BPPCl [9] were
synthesised according to the literature procedures. Sty-
rene was filtered through alumina and freshly distilled
just before use.
The syntheses of aminophosphonite–phosphite lig-
ands and of the corresponding platinum complexes are
readily carried out. The platinum complexes are stable
under normal conditions and catalyse the asymmetric
hydroformylation of styrene. These new catalysts are
significantly more active than the previously reported
PtꢀAMPP. Nevertheless, only modest enantioselectivi-
ties are obtained and the hydrogenation side-reaction is
observed in quite large amounts. The corresponding
rhodium complexes present significantly poorer cata-
lytic properties in styrene hydroformylation.
Elemental analyses were carried out by Laboratoires
Wolff, Clichy, France. NMR spectra were obtained at
room temperature in CD2Cl2 on a Bruker AC300 spec-
Table 4
Enantioselective hydroformylation of styrene involving Pt (5 and 6) a and rhodium (Rh-3 and Rh-4) b precatalysts
e
Run
Complex
Solvent
Temperature
(°C) (PCO/PH
(atm))
Conversion
(%) c
TOF
Aldehydes
(mol %) e
Select. b/n
PhEt
ee g
d
(h−1
)
(mol. %) f
2
1
2
3
4
5
6
7
8
5
5
5
5
5
5
5
6
PhCH3
CH2Cl2
50 (45/90)
50 (45/90)
50 (45/90)
50 (60/60)
50 (45/90)
30 (45/90)
20 (45/90)
20 (45/90)
80 (6/6)
97
98
97
98
98
54
41
30
100
15
20
38
35
36
38
55
20
15
15
15
1.3
1.7
70
65
66
80
65
66
66
65
53/47
55/45
55/45
53/47
53/47
53/47
53/47
66/34
63/37
88/12
72/28
30
35
34
20
35
34
34
35
0
26
30
30
30
30
38
44
15
5
h
h
h
h
h
h
CH2Cl2/PhCH3
CH2Cl2/PhCH3
CH2Cl2/PhCH3
CH2Cl2/PhCH3
CH2Cl2/PhCH3
CH2Cl2/PhCH3
toluene
9
10
11
Rh-3
Rh-3
Rh-4
100
100
100
toluene
toluene
40 (6/6)
40 (6/6)
0
0
19
15
a Conditions : platinum complex/SnCl2 1/1, substrate/catalyst=1000, 18 h, 30 ml of solvent.
b Rhodium based hydroformylations were carried out in a stainless steel autoclave in toluene (30 ml) [Rh]=5.41×10−3 mol L−1; precatalyst
Rh-3: in situ mixing of Rh4(CO)12 and 3, Rh-4: in situ mixing of Rh4(CO)12 and 4, ligand/Rh=1.5; S/Rh=400; t=48 h.
c Styrene conversion.
d Turn over frequency: moles of aldehydes (b+n) formed per mole of platinum per hour.
e Selectivity into 2-(b) and 3-phenylpropanol (n) determined by GLC analysis.
f Selectivity into ethylbenzene determined by GLC analysis.
g See Section 4, all configurations are (S).
h Mixture of CH2Cl2/PhCH3 2/28 ml.

120
S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
1
trometer operating at 300.1, 121.5 and 75 MHz for H,
31P{1H}, and 13C{1H}, respectively. Proton and carbon
chemical shifts were referenced to the deuterated sol-
vent employed relative to tetramethylsilane (l 0 ppm).
Phosphorous chemical shifts are reported with positive
values downfield from external 85% H3PO4 in D2O (l 0
ppm). The compositions of the catalytic reaction mix-
tures were determined by GLC analysis with a Delsi 30
gas chromatograph equipped with a flame ionisation
detector using a 25 m×0.25 mm CPSil 5-CB fused
silica capillary column. The determination of the enan-
tiomeric excesses was performed on a 25 m×0.32 mm
Chirasil-Dex column on the alcohols obtained after
reduction of the aldehydes.
Hz), 130.3 (d, JCP=14.7 Hz), 130.9 (s), 148.0 (d, JCP=
4.3 Hz), 149.7 (d, JCP=5.2 Hz), 149.8 (d, JCP=4.7
Hz), 150.4 (d, JCP=3.2 Hz), 182.6 (d, JCP=15.9 Hz).
31P{1H}-NMR (l ppm): 137 (s), 145 (s).
4.1.3. [(S)-N-(2,2%biphenoxyphosphino)-
2,2%biphenoxyphosphinoxymethyl)pyrrolidine]
dichloroplatinum(II) (5)
A Schlenk tube was charged with Pt(COD)Cl2 (0.1 g,
0.27 mmol), CH2Cl2 (10 ml) and a stir bar. A solution
of 3 (0.155 g, 0.29 mmol, 1.1 equivalents) in CH2Cl2 (5
ml) was added via cannula with stirring. After 30 min
vigorous stirring at room temperature (r.t.), the solvent
was evaporated from the colourless solution under vac-
uum. Then, the white residue was washed with diethyl
ether (2×10 ml) and finally dried under vacuum
providing compound 5 in 97% yield (0.202 g, 0.25
mmol). Anal. Found: C, 42.31; H, 3.38; N, 1.60; O,
9.60. Calc. for C29H25Cl2NO5P2Pt·0.5CH2Cl2: C, 42.29;
4.1.1. Synthesis of [(S)-N-(2,2%biphenoxyphosphino)-
2-(2,2%biphenoxyphosphinoxymethyl) pyrrolidine (3)
Under nitrogen, a 100-ml Schlenk flask was charged
with a magnetic stir bar, (S)-2-hydroxymethyl proline
(1) (0.52 g, 5.16 mmol), THF (25 ml), and NEt3 (5 ml).
Then BPPCl (3.1 g, 12.38 mmol, 2.4 equivalents) in
solution in THF (15 ml) was slowly added via cannula
with stirring. A white precipitate formed instanta-
neously. The mixture was stirred vigorously and heated
at 60°C overnight. The end of the reaction was checked
by 31P-NMR on an aliquot taken from the reaction
medium. After cooling, the resulting crude reaction
mixture was filtered under nitrogen through a pad of
basic alumina (1×3 cm) and washed with Et2O (2×20
ml). The solvents were evaporated from the combined
filtrates under reduced pressure. The residue was fur-
ther dried under oil pump vacuum, affording 3 as a
white solid in 88% yield (2.4 g, 4.5 mmol). Anal.
Found: C, 65.52; H, 4.90; N, 2.72. Calc. for
C29H25NO5P2: C, 65.79; H, 4.75; N, 2.65%. 1H-NMR (l
ppm): 1.60 (m, 2H), 1.71 (m, 1H), 1.81 (m, 1H), 2.69
(m, 1H), 3.11 (m, 1H), 3.67 (m, 1H), 3.85 (m, 2H),
6.95–7.39 (m, 16H). 13C{1H}-NMR (l ppm): 24.9 (s),
28.4 (d, JCP=3.7 Hz), 45.1 (s), 58.4 (dd, JCP=25 and
3 Hz), 67.3 (s), 122.0 (s), 122.2 (s), 124.8 (d, JCP=8.2
Hz), 125.4 (s), 129.4 (s), 129.6 (d, JCP=4.3 Hz), 129.9
(d, JCP=14.6 Hz), 130.2 (s), 150.0 (d, JCP=5.2 Hz),
150.2 (d, JCP=5.8 Hz), 151.8 (d, JCP=5.3 Hz), 152.1
(d, JCP=5.2 Hz). 31P{1H}-NMR (l ppm): 138.7 (s),
150.0 (s).
1
H, 3.12; N, 1.67; O, 9.55%. H-NMR (l ppm): 1.73 (m,
1H), 1.80 (m, 1H), 2.12 (m, 2H), 2.69 (m, 2H), 3.3 (m,
1H), 3.9 (m, 1H), 4.6 (m, 1H). 31P{1H}-NMR (l ppm):
72 (d, JPP=28, JPtꢀP=5792 Hz, satellites), 85 (d, JPP
28, JPtꢀP=5392 Hz, satellites).
=
4.1.4. [(S)-N-(2,2%biphenoxyphosphino)-
2,2%biphenoxyphosphinoxymethyl)pyrrolidinone]
dichloroplatinum(II) (6)
This compound was synthesised in the same way as
5. Yield 98%. 31P{1H}-NMR (l ppm): 86 (d, JPP=4.2,
JPtꢀP=5833 Hz, satellites), 90 (d, JPP=4.2, JPtꢀP=5658
Hz, satellites).
4.1.5. Crystal structure of 5·CH2Cl2
Suitable crystals were obtained as colourless prisms
by slow crystallisation from dichloromethane. The crys-
tal selected was mounted for data collection on an
Enraf–Nonius CAD-4 diffractometer, as summarised in
Table 1. The intensities of three representative reflec-
tions, which were measured after every 2 h, remained
constant throughout data collection, indicating crystal
and electronic stability. Data were corrected for
Lorentz and polarisation effects but not (since the faces
were not identified easily) for absorption. The atomic
coordinates of the platinum atom were deduced from
the Patterson function, and the positions of chlorine,
phosphorous, oxygen, nitrogen, and carbon atoms were
identified from successive difference-Fourier syntheses.
All positional and thermal parameters were refined with
SHELX-76 [19]. The atomic scattering factors for neutral
atoms, as well as the anomalous dispersion correction
coefficients, were taken from Ref. [20]. The presence of
a solvent molecule highly disordered in the structure
did not permit us to localise the hydrogen atoms.
Refinement of all atoms yielded a final R=0.039 and
Rw=0.045, with w=1/[|2(F)+0.00369F2] in the final
stages.
4.1.2. Synthesis of [(S)-N-(2,2%biphenoxyphosphino)-
2-(2,2%biphenoxyphosphinoxymethyl) pyrrolidinone (4)
This compound was synthesised in the same way as
3. The final filtration was carried out through a small
1
pad of alumina (1×0.5 cm). Yield 86%. H-NMR (l
ppm): 1.77 (m, 2H), 2.17 (m, 2H), 3.77 (m, 2H), 3.8 (m,
2H), 6.9–7.5 (m, 16H). 13C{1H}-NMR (l ppm): 24.3
(s), 31.2 (s), 46.3 (s), 57.4 (dd, JCP=4.3 and 7.9 Hz),
67.8 (d, JCP=3.7 Hz), 122.1 (s), 122.4 (s), 125.5 (d,
J
CP=6.6 Hz), 125.9 (s), 129.1 (s), 129.7 (d, JCP=5.2

S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
121
4.1.6. Catalytic asymmetric platinum-based
hydroformylations
free of charge from The Director, CCDC, 12 Union
Road, Cambridge CB12 1EZ, UK (Fax: +44-1223-
336033; e-mail: deposit@ccdc.cam.ac.uk or www: http:/
/www.ccdc.cam.ac.uk).
All catalytic reactions were conducted in a magneti-
cally stirred and double-walled 60-ml stainless steel
reactor. In a typical experiment, a solution of the
cocatalyst SnCl2·2H2O (0.1 mmol) and of the platinum
complex (0.1 mmol) in CH2Cl2 (2 ml) was introduced in
the reactor under dinitrogen. Then, styrene (100 mmol)
and the internal standard (n-decane) in toluene (28 ml)
were introduced. The reactor was sealed, pressurised to
130 atm with CO–H2 (1/2) and heated at 50°C. The
reaction was monitored by GLC analysis of aliquots of
the reaction mixture. When the reaction was stopped,
the reactor was cooled to r.t. and depressurised. The
pale-yellow solution was analysed by GLC to determine
the selectivities into hydroformylated (branched and
normal) products and ethylbenzene. Pentane was then
added to the crude reaction mixture to precipitate the
catalyst. The resulting solid was filtered off and the
remaining solution was concentrated by rotatory evap-
oration. The crude mixture was then dissolved in di-
ethylether (20 ml) and an excess of LiAlH4 was added.
After 1 h stirring at r.t., the medium was cautiously
quenched with aqueous HCl (1 N) (10 ml). The
aqueous layer was extracted two times with diethylether
(10 ml). The combined organic phases were washed
with water (2×20 ml), dried over MgSO4, concentrated
by rotatory evaporation, and analysed by chiral GC.
Acknowledgements
This work has been supported financially by the
Centre National de la Recherche Scientifique and the
Ministe`re de l’Enseignement Supe´rieur. We thank
Olivier Lot and Dr Herve´ Bricout for initial experi-
ments, and Catherine Me´liet for skilful NMR
assistance.
References
[1] C. Botteghi, S. Paganelli, A. Schionato, M. Marchetti, Chirality
3 (1991) 355.
[2] (a) J.P. Rieu, G. Boucherle, H. Cousse, G. Mouzin, Tetrahedron
42 (1986) 4095. (b) H.R. Sonawane, N.S. Bellur, J.R. Ahuja,
D.G. Kulkarni, Tetrahedron: Asymmetry 3 (1992) 163.
[3] Review articles on asymmetric hydroformylation: (a) S. Gladiali,
J. Bayon, C. Claver, Tetrahedron: Asymmetry 7 (1995) 1453. (b)
F. Agbossou, J.-F. Carpentier, A. Mortreux, Chem. Rev. 95
(1995) 2485. (c) G. Consiglio, in: I. Ojima (Ed.), Catalytic
Asymmetric Syntheses, VCH, Weinheim, 1993, p. 273. (d) M.
Beller, B. Cornils, C.D. Frohning, V.W. Kohlpainter, J. Mol.
Catal. 104 (1995) 17. (e) K. Nozaki, H. Takaya, T. Hiyama,
Top. Catal. 4 (1997) 175.
[4] (a) N. Sakai, S. Mano, K. Nozaki, H. Takaya, J. Am. Chem.
Soc. 115 (1993) 7033. (b) N. Sakai, K. Nozaki, H. Takaya, J.
Chem. Soc. Chem. Commun. (1994) 395. (c) T. Horiuchi, T.
Ohta, K. Nozaki, H.Takaya, J. Chem. Soc. Chem. Commun.
(1996) 155.
[5] (a) N. Sakai, K. Nozaki, K. Mashima, I. Takaya, Tetrahedron:
Asymmetry 3 (1992) 583. (b) J.E. Babbin, G.T. Whiteker,
WOUS patent 911 518, 1992 (to Union Carbide Corp.), Chem.
Abstr. 119 (1993) 159872h. (c) G.J.H. Buisman, P.C.J. Kramer,
P.W.N.M. van Leeuwen, Tetrahedron: Asymmetry 4 (1993)
1625. (d) G.J.H. Buisman, M.E. Martin, E.J. Vos, A. Klootwijk,
P.C.J. Kramer, P.W.N.M. van Leeuwen, Tetrahedron: Asymme-
try 7 (1995) 719.
[6] (a) G. Consiglio, S.C.A. Nefkens, A. Borner, Organometallics 10
(1991) 2046. (b) L. Kolar, P. Sandor, G. Szalontai, J. Mol. Catal.
67 (1991) 191. (c) A. Scrivanti, S. Zeggio, V. Boeghetto, U.
Matteoli, J. Mol. Catal. 101 (1995) 217. (d) J.K. Stille, H. Su, P.
Brechor, G. Parrinello, L.S. Hegedus, Organometallics 10 (1991)
1183. (e) S. Csere´pi-Szu¨cs, J. Bakos, Chem. Commun. (1997)
635.
[7] (a) S. Mutez, A. Mortreux, F. Petit, Tetrahedron Lett. 29 (1988)
1911. (b) Y. Pottier, A. Mortreux, F. Petit, J. Organomet. Chem.
370 (1989) 333. (c) S. Naili, J.-F. Carpentier, F. Agbossou, A.
Mortreux, G. Nowogrocki, J.-P. Wignacourt, Organometallics
14 (1995) 401. (d) A.M. Bandini, G. Banditelli, E. Cesarotti, G.
Minghetti, B. Bovio, Inorg. Chem. 31 (1992) 391. (e) R. Ewalds,
E.B. Eggeling, A.C. Hewat, P.C.J. Kramer, P.W.N.M. van
Leeuwen, D. Vogt, Chem. Eur. J. 6 (2000) 1496.
4.1.7. Catalytic asymmetric rhodium-based
hydroformylations
In a typical experiment, a Schlenk tube was charged
with a stir bar, the ligand (0.216 mmol), Rh4(CO)12
(26.9 mg, 0.036 mmol), and toluene (10 ml). After 15
min stirring at r.t., the solution was transferred via
cannula to a 60-ml stainless steel double-walled auto-
clave equipped with a stir bar. After a N2 purge, styrene
(6 g, 57.6 mmol), diluted in toluene (10 ml) (sub-
strate=2.16 mol l−1), was also added via cannula. The
autoclave was pressurised to 12 bar of syn-gas (H2–CO
1/1) and then heated to the desired temperature. The
pressure was kept constant throughout the whole reac-
tion by using a gas reservoir along with a pressure
regulator. The workup and determination of conver-
sion, chemo-, regio-, and enantioselectivities were car-
ried out as reported above for platinum-based catalytic
runs.
5. Supplementary material
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC No. 154572 for PtCl2{(S)-BP-
ProNOP}. Copies of this information may be obtained
[8] (a) A. Roucoux, L. Thieffry, J.-F. Carpentier, M. Devocelle, C.
Me´liet, F. Agbossou, A. Mortreux, Organometallics 15 (1996)
2440. (b) F. Agbossou, J.-F. Carpentier, C. Hatat, N. Kokel, A.
Mortreux, Organometallics 14 (1995) 2480.

122
S. Naili et al. / Journal of Organometallic Chemistry 628 (2001) 114–122
[9] G.J.H. Buisman, P.C.J. Kramer, P.W.N.M. van Leeuwen, Tetra-
hedron: Asymmetry 4 (1993) 1625.
[15] S. Csere´pi-Szu¨cs, G. Huttner, L. Zsolnai, J. Bakos, J.
Organomet. Chem. 586 (1999) 70.
[10] O. Lot, I. Suisse, A. Mortreux, F. Agbossou, J. Mol. Catal. 164
(2000) 125.
[11] P.S. Pregosin, R.W. Kunz, in: P. Diehl, E. Fluck, R. Kosfeld
(Eds.), NMR 16 Basic Principles and Progress: 31P and 13C
NMR of Transition Metal Phosphine Complexes, Springer,
Berlin, 1979.
[12] (a) T.T. Derencse´nyi, Inorg. Chem. 20 (1981) 665. (b) G.G.
Mather, A. Pidcock, G.J.N. Rapsey, J. Chem. Soc. Dalton
Trans. (1973) 2095. (c) L.M. Green, Y. Park, D.W. Meek, Inorg.
Chem. 27 (1988) 1658. (d) J.A. Rahn, L. Baltusis, J.H. Nelson,
Inorg. Chem. 29 (1990) 750. (e) A.W. Verstuyft, D.A. Redfield,
L.W. Cary, J.H. Nelson, Inorg. Chem. 15 (1976) 1128.
[13] E. Cesarotti, M. Grassi, L. Prati, J. Chem. Soc. Dalton Trans. 1
(1989) 161.
[16] Square planar complexes bearing AMPP ligands — Platinum:
(a) S. Naili, J.-F. Carpentier, F. Agbossou, A. Mortreux, New J.
Chem. 21 (1997) 919. (b) Ref. [7c]. Rhodium: (d) V.A. Pavlov,
E.I. Klabunovsky, Y.T. Struchkov, A.A. Voloboev, A.I.
Yanovsky, J. Mol. Catal. 14 (1988) 119. (e) E. Cesarotti, A.
Chiesa, G. Ciani, A. Sironi, J. Organomet. Chem. 251 (1983) 79.
Palladium: (f) E. Cesarotti, M. Grassi, L. Prati, F. Demartin, J.
Organomet. Chem. 370 (1989) 407.
[17] Ruthenium complexes bearing AMPP ligands: (a) E. Cesarotti,
L. Prati, A. Sironi, G. Ciani, C. White, J. Chem. Soc. Dalton
Trans. 5 (1987) 1149. (b) Ref. [14].
[18] D. Drew, J.R. Doyle, A.G. Shaver, Inorg. Synth. 13 (1972) 47.
[19] G. Sheldrick, SHELX-76, System of Crystallographic Computer
Programs, University of Cambridge, Cambridge, UK, 1976.
[20] D.T. Cromer, International Tables for X-ray Crystallography,
vol. IV, Kynoch Press, Birmingham, UK, 1974.
[14] F. Hapiot, F. Agbossou, C. Me´liet, A. Mortreux, G.M. Rosair,
A.J. Welch, New J. Chem. 21 (1997) 1161.
.