New chiral diphosphinites: Synthesis of Rh complexes. Heterogenisation on zeolites

Article abstract of DOI:10.1016/S0022-328X(99)00373-3

New diphosphinite ligands derived from (2S,4R), (2S,4S)-1-benzyl-4-hydroxy-4-phenyl-2-(1,1-diphenylmethyl)pyrrolidinylmethanol and (2S,4R), (2S,4S)-1-(3-triethoxysilyl)propylaminocarbonyl-4-hydroxy-4-phenyl-2-(1,1- diphenylmethyl)pyrrolidi nylmethanol have been prepared in high yields (60-80%) and reacted with [RhCl(cod)]₂ to yield cationic [Rh(diphosphinite)(cod)]⁺ complexes. Those metal complexes bearing a triethoxysilyl group were covalently bonded to USY-zeolite by controlled hydrolysis and Rh-heterogenised complexes were obtained. A comparative study (homogeneous versus supported) for the catalytic activity and selectivity in hydrogenation and hydroformylation reactions was made, obtaining an enhanced performance for heterogenised catalysts; moreover, those catalysts can be recycled in successive runs, by a simple filtration, without a significant loss of activity.

Full text of DOI:10.1016/S0022-328X(99)00373-3

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 588 (1999) 186–194
New chiral diphosphinites: synthesis of Rh complexes.
Heterogenisation on zeolites
A. Fuerte a, M. Iglesias a,*, F. Sa´nchez b,1
a Instituto de Ciencia de Materiales de Madrid, CSIC, Cantoblanco, E-28049 Madrid, Spain
b Instituto de Qu´ımica Orga´nica General, CSIC, Juan de la Cier6a 3, E-28006 Madrid, Spain
Received 1 March 1999; received in revised form 20 June 1999
Abstract
New
diphosphinite
ligands
derived
from
(2S,4R),
(2S,4S)-1-benzyl-4-hydroxy-4-phenyl-2-(1,1-diphenylmethyl)-
pyrrolidinylmethanol
and (2S,4R),
(2S,4S)-1-(3-triethoxysilyl)propylaminocarbonyl-4-hydroxy-4-phenyl-2-(1,1-diphenyl-
methyl)pyrrolidinylmethanol have been prepared in high yields (60–80%) and reacted with [RhCl(cod)]2 to yield cationic
[Rh(diphosphinite)(cod)]+ complexes. Those metal complexes bearing a triethoxysilyl group were covalently bonded to USY-zeo-
lite by controlled hydrolysis and Rh-heterogenised complexes were obtained. A comparative study (homogeneous versus
supported) for the catalytic activity and selectivity in hydrogenation and hydroformylation reactions was made, obtaining an
enhanced performance for heterogenised catalysts; moreover, those catalysts can be recycled in successive runs, by a simple
filtration, without a significant loss of activity. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Rhodium; Supported catalysts; Zeolites; Hydrogenation; Hydroformylation
hand, are easier to synthesise and are more stable than
phosphines. Owing to previous results obtained on Rh
complexes supported on inorganic carriers, it can be
worth studying supported complexes with phosphite
ligands to see whether this introduces any catalytic
improvement. Indeed, heterogenised homogeneous cat-
alysts combine the properties of homogeneous catalysts
viz. reactivity, controllability and selectivity, and het-
erogeneous catalysts viz. enhanced stability and
reusability [3]. Moreover, we conducted an extensive
study of the synthesis of new chiral ligands in order to
apply them in asymmetric synthesis. We found that the
catalytic activity could be enhanced as a result of such
immobilisation on a support. For that purpose, special
efforts have been devoted toward the synthesis of easily
accessible ligands based on natural amino acids [4].
These ligands have been applied with success in asym-
metric catalysis, for example, the addition of diethylzinc
to enones [5], hydrogenation [6], cyclopropanation [7],
and oxidation of olefins [8]. Hence, we set out to
explore the synthesis and application of new, closely
related bis(phosphinites) and their corresponding
rhodium complexes exhibiting specific constraints and
electronic properties in order to improve both the activ-
1. Introduction
Asymmetric syntheses with various transition metal
complexes as catalysts have attracted much attention in
the past two decades. Among them, rhodium(I) chiral
diphosphine catalyst complexes have achieved very high
stereoselection (up to 99% ee) in asymmetric hydro-
genation of carbon–carbon double bonds [1], and some
processes based on these catalysts have found industrial
applications [2]. In these systems, the access of reactants
to chiral ligands was essential for the development of
new catalytic systems exhibiting high efficiency and
enantioselectivity. In particular, several classes of chiral
phosphines have been described and when associated
with the rhodium or ruthenium present in the catalytic
precursors, they are able to hydrogenate olefins and
ketones with almost complete enantioselectivity [1,2].
Up to now less work has been done with less electron-
rich ligands such as phosphites, which, on the other
* Corresponding author. Tel.: +34-913-349-032; fax: +34-913-
720-623.
E-mail address: marta.iglesias@icmm.csic.es (M. Iglesias)
1 Also corresponding author.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(99)00373-3

A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
187
2.1. Synthesis of ligands
ity and selectivity of the hydrogenation and hydro-
formylation reactions.
The alcohols were prepared according to the method
previously described [8].
In this paper, we report the synthesis of new bis(pho-
sphinite) ligands, the synthesis and spectroscopic char-
acterisation of their corresponding rhodium complexes
(homogeneous and heterogenised on USY-zeolite) and
their subsequent application in asymmetric hydrogena-
tion and hydroformylation of olefins.
2.1.1. Preparation of (2S,4R)-1-benzyl-4-diphenylphos-
phinoxy-4-phenyl-2-(1,1-diphenyl-1-diphenylphosphin-
oxymethyl)pyrrolidine, (2S,4R)-I)
In a typical experiment, (2S,4R)-1-benzyl-4-hydroxy-
4-phenyl-2-(1,1-diphenylmethyl)pyrrolidinylmethanol
(100 mg, 0.23 mmol) was placed in a 100 ml Schlenk
tube (cooled in a ice-water bath) along with diethyl
ether (30 ml) and triethylamine (0.073 ml) and was
reacted with a solution of chlorodiphenylphosphine
(101.4 mg, 0.46 mmol) in diethyl ether (10 ml). The
mixture was stirred at room temperature (r.t.) for 24 h.
Then the reaction mixture was concentrated in vacuo
and ammonium chloride and phosphorus impurities
were removed by filtration through a short column of
basic alumina (eluted with diethyl ether). The solvents
were removed from the filtrate under reduced pressure
to afford (2S,4R)-I as a white solid (170 mg, 90%
yield). No further attempts were made to purify com-
pound (2S,4R-I) as it was found to be amenable to
further reactions. [h]2D5= +22.3° (c=1, CH2Cl2). Anal.
Calc. for C54H47NO2P2: C, 80.7; H, 5.9; N, 1.7; P, 7.7.
Found: C, 80.2; H, 5.7; N, 1.4; P, 7.2%. 1H-NMR
(CDCl3): l 8.0–6.7 (m, 40H, H phenyl); 4.23–4.20 (dd,
1H, NCH); 3.63–3.55 (AB, 2H, CH2Ph); 3.1–2.97 (m,
2H, CH2N); 2.69–2.66 (m, 1H, CHCH2); 2.04–2.00 (m,
1H, CHCH2). 13C-NMR (CDCl3): l 146.8, 146.1, 143.2,
138.7 (CaromꢀR); 128.5, 128.3, 128.2, 127.9, 127.1,
127.0. 126.7, 126.6, 125.8, 125.6, 125.3 (CaromꢀH); 78.2,
78.0 (Ph2COP, PhCOP); 70.6 (CH); 68.6 (CCH2N);
60.4 (PhCH2); 45.1 (CCH2CH). 31P-NMR (CDCl3): l
121.8 ppm.
2. Experimental
All organometallic complexes were prepared under
dinitrogen by standard Schlenk techniques. The starting
complex [Rh(cod)Cl]2 was prepared according to the
method reported in the literature [9]. All solvents were
carefully degassed before use. The silylating agent
OCN(CH2)3Si(OEt)3 obtained from Fluka (96% pure),
was distilled before use. C, H and N analyses were
carried out by the analytical department of the Institute
of Materials Science (CSIC) with a Perkin–Elmer 240C
apparatus. Metal contents were analysed by atomic
absorption using a Unicam Philips SP9 apparatus. IR
spectra were recorded with a Nicolet XR60 spectropho-
tometer (range 4000–200 cm−1) as KBr pellets; H-,
1
13C- and 31P-NMR spectra were taken on Varian
XR300 and Bruker 200 spectrometers. 1H-NMR chemi-
cal shifts are given in ppm using tetramethylsilane as an
internal standard; 31P-NMR chemical shifts are
downfield from 85% H3PO4. Optical rotation values
were measured at the sodium-D line (589 nm) with a
Perkin–Elmer 241 MC polarimeter. Gas chromatogra-
phy analysis was performed using a Hewlett–Packard
5890 II with a flame ionisation detector in a cross-
linked methylsilicone column.
The inorganic support was an ultrastable Y (USY)
zeolite prepared by steam calcination at 1023 K of an
80% ammonium-exchanged NaY (SK40 Union Car-
bide), followed by treatment with a 1 N citric acid
solution at 333 K for 30 min to remove extra-frame-
work species. After this, the zeolite was thoroughly
washed and dried at 403 K for 6 h. The final zeolite had
The phosphinites (2S,4S)-II, (2S,4R)-III and
(2S,4S)-IV were prepared through a procedure similar
to that given for (2S,4R)-I starting from the corre-
sponding dialcohol.
2.1.2. (2S,4S)-1-benzyl-4-diphenylphosphinoxy-4-
phenyl-2-(1,1-diphenyl-1-diphenylphosphinoxymethyl)-
pyrrolidine (2S,4S)-II
a
well-developed supermicropore–mesopore system
,
,
(pore diameter 12–30 A besides the typical ca. 12 A
micropores). The controlled dealumination promotes
destruction of some sodalite units, which allowed direct
communication between a-cages generating cavities
wider than 12 A. The formation of supermicropores
and large mesopores has been detected by N2 adsorp-
tion–desorption. The main characteristics of the resul-
tant zeolite are: unit cell size, 24.40 A; bulk SiO2/Al2O3,
4.2; and crystallinity, greater than 95%. The inorganic
support was dried at 415 K under 0.01 torr before the
anchoring process.
24 h, 82% yield, white powder. [h]2D5= +14.19° (c=
1, CH2Cl2). Anal. Calc. for C54H47NO2P2: C, 80.7; H,
5.9; N, 1.7; P, 7.7. Found: C, 80.4; H, 5.6; N, 1.5; P,
7.5%. 1H-NMR (CDCl3): l 7.9–6.9 (m, 40H, H
phenyl); 5.34–5.24 (dd, 1H, NCH); 4.40–4.36 (m, 1H,
CH2N); 4.00–3.84 (m, 1H, CH2N); 3.64–3.39 (AB, 2H,
CH2Ph); 2.56–2.36 (m, 2H, CHCH2). 13C-NMR
(CDCl3): l 147.7, 145.7, 143.1, 140.0 (CaromꢀR); 128.4,
128.3, 128.0, 127.8, 127.0, 126.6, 126.5, 126.4, 125.7,
125.0 (CaromꢀH); 81.8 (Ph2COP); 72.3 (PhCOP); 66.8
,
,

188
A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
(CH); 61.7 (PhCH2N); 43.8 (CCH2N); 26.9 (CCH2CH).
31P-NMR (CDCl3): l 118.8 ppm.
Anal. Calc. for C62H59F6NO2P3Rh: C, 64.2; H, 5.1; N,
1.2; P, 8.0; Rh, 8.9. Found C, 63.8; H, 4.8; N, 0.9; P,
7.8; Rh, 8.8%. H-NMR (CDCl3): l 8.0–7.2 (m, 40H,
1
H phenyl); 4.3 (br, 4H, CHꢁ); 4.2 (m, 1H, NCH);
3.6–3.5 (AB, 2H, CH2Ph); 3.1 (m, 2H, CH2N); 2.6 (m,
5H, CHCH2, CH2ꢀCHꢁ); 2.0 (m, 1H, CHCH2); 1.8 (m,
4H, CH2ꢀCHꢁ). 13C-NMR (CDCl3): l 146.8–138.7
(CaromꢀR); 129–125 (CaromꢀH); 81.5–81.0 (ꢁCHcod);
78.5–78.0 (Ph2COP, PhCOP); 70.9 (CH); 68.6
(CCH2N); 60.8 (PhCH2); 45.1 (CCH2CH); 30.4
(CHcod). 31P-NMR (CDCl3): l 126.1 (br); −143.8
(1J(PꢀF)=714 Hz).
Complexes 2–4 were prepared following a procedure
similar to that given for 1 using 0.4 equivalents of
[Rh(cod)Cl]2 and two equivalents of the appropriate
ligand (Scheme 1).
2.1.3.
(2S,4R)-1-(3-Triethoxysilyl)propylamino-carbonyl-
4-diphenylphosphinoxy-4-phenyl-2-(1,1-diphenyl-
1-diphenylphosphinoxymethyl)pyrrolidine, (2S,4R)-III
30 h, 67% yield; colourless oil. Anal. Calc. for
C57H62N2O6P2Si: C, 71.2; H, 6.5; N, 2.9; P, 6.4. Found:
1
C, 70.8; H, 6.4; N, 2.5; P, 6.1%. H-NMR (CDCl3): l
7.98–7.13 (m, 35H, H phenyl); 4.12–4.07 (m, 1H,
NCH); 3.85–3.77 (m, 6H, CH2O); 3.76–3.67 (m, 2H,
CH2N(CO)); 3.03–2.99 (m, 2H, CH2N); 2.20–2.10 (m,
2H, CHCH2); 1.61–1.24 (m, NCH2CH2); 1.22 (t, 9H,
CH3); 0.65–0.62 (m, 2H, CH2Si). 13C-NMR (CDCl3): l
158.5 (CO), 145.3, 145.0 (CaromꢀR); 128.4, 128.3, 128.2,
127.4, 127.3, 127.2, 127.0, 124.7 (CaromꢀH); 81.3
(Ph2COP); 79.8 (PhCOP); 67.1 (CH); 63.3 (CH2N);
58.4 (CH3CH2O); 46.1 (CH2CH); 43.5 (CH2NHCO);
23.1 (CH2CH2CH2); 18.3 (CH3CH2O); 7.7 (CH2Si).
31P-NMR (CDCl3): l 122.1 ppm.
2.2.1. [Rh(cod){(2S,4S)-II}]PF6 (2)
Yield 74%. [h]2D5= +6.09° (c=0.55, CH2Cl2).
\(CH3CN) (V−1 cm2 mol−1)=103. Anal. Calc. for
C62H59F6NO2P3Rh: C, 64.2; H, 5.1; N, 1.2; P, 8.0; Rh,
8.9. Found C, 63.9; H, 4.7; N, 1.1; P, 8.2; Rh, 9.3%.
1H-NMR (CDCl3): l 7.7–7.0 (m, 40H, H phenyl); 5.5
(dd, 1H, NCH); 4.3–4.2 (br, 4H, CHꢁ); 4.1, 3.7 (m, 2H,
CH2N); 3.6–3.4 (AB, 2H, CH2Ph); 2.6–2.4 (m, 6H,
CHCH2, CH2ꢀCHꢁ); 1.8–1.6 (m, 4H, CH2ꢀCHꢁ). 13C-
NMR (CDCl3): l 148.0–140.0 (CaromꢀR); 128.4–125.0
(CaromꢀH); 81.7 (Ph2COP); 79.3–79.0 (ꢁCHcod); 72.5
(PhCOP); 66.9 (CH); 61.5 (PhCH2N); 43.8 (CCH2N);
28.2 (CHcod); 26.9 (CCH2CH). 31P-NMR (CDCl3): l
122.0 (br); −143.8 (1J(PꢀF)=714 Hz).
2.1.4. (2S,4S)-1-(3-Triethoxysilyl)propylaminocarbonyl-
4-diphenylphosphinoxy-4-phenyl-2-(1,1-diphenyl-1-
diphenylphosphinoxymethyl)pyrrolidine, (2S,4S)-IV
32 h, colourless oil, 56% yield. Anal. Calc. for
C57H62N2O6P2Si: C, 71.2; H, 6.5; N, 2.9; P, 6.4. Found:
1
C, 70.7; H, 6.4; N, 2.8; P, 6.0%. H-NMR (CDCl3): l
7.89–7.09 (m, 35H, H phenyl); 5.24 (m, 1H, NCH);
4.06–4.05 (m, 2H, CH2N(CO)); 3.69–3.56 (q, 6H,
CH2O); 2.97–2.93 (m, 2H, CH2N); 2.12–2.07 (m, 2H,
CHCH2); 1.60–1.40 (m, NCH2CH2); 1.20 (t, 9H, CH3);
0.82–0.80 (m, 2H, CH2Si). 13C-NMR (CDCl3): l 158.8
(CO); 147.8, 145.5, 143.3, 140.0 (CaromꢀR); 128.4, 128.3,
127.8, 127.0, 126.6, 126.4, 125.7, 125.0 (CaromꢀH); 81.5
(Ph2COP); 72.0 (PhCOP); 66.7 (CH); 63.1 (CH2N);
58.4 (CH3CH2O); 46.4 (CH2CH); 43.3 (CH2NHCO);
23.0 (CH2CH2CH2); 18.5 (CH3CH2O); 7.9 (CH2Si).
31P-NMR (CDCl3): l 119.0 ppm.
2.2.2. [Rh(cod){(2S,4R)-III}]PF6 (3)
Yield 55%. [h]2D5= +13.24° (c=1.2, CH2Cl2).
\(CH3CN) (V−1 cm2 mol−1)=76. Anal. Calc. for
C65H74F6N2O6P3RhSi: C, 59.3; H, 5.6; N, 2.1; P, 7.1;
Rh, 7.8. Found C, 58.8; H, 6.0; N, 1.7; P, 7.5; Rh,
8.4%. 1H-NMR (CDCl3): l 7.8–7.1 (m, 35H, H
phenyl); 4.2 (br, 4H, CHꢁ); 4.1 (br, m, 1H, NCH); 3.8
(m, 6H, CH2O); 3.7 (m, 2H, CH2N(CO)); 3.0 (m, 2H,
CH2N); 2.6 (br, 4 H, CH2ꢀCHꢁ); 2.2 (m, 2H, CHCH2);
1.8 (m, 4H, CH2–CHꢁ); 1.5ꢀ1.3 (m, NCH2CH2); 1.2 (t,
9H, CH3); 0.55 (m, 2H, CH2Si). 13C-NMR (CDCl3): l
158.7 (CO), 145.5–145.0 (CaromꢀR); 128.4–124.7
(CaromꢀH); 81.4 (Ph2COP); 80.0 (PhCOP); 79.8–79.3
(ꢁCHcod); 67.2 (CH); 63.3 (CH2N); 58.5 (CH3CH2O);
46.2 (CH2CH); 43.4 (CH2NHCO); 28.5 (CHcod); 23.0
(CH2CH2CH2); 18.5 (CH3CH2O); 8.0 (CH2Si). 31P-
NMR (CDCl3): l 125.8 (br); −143.6 (1J(PꢀF)=714
Hz).
2.2. Preparation of [Rh(cod){diphosphinite}]PF6
complexes 1–4
Typical procedure for [Rh(cod){(2S,4R)-I}]PF6 (1).
To a yellow solution of [Rh(cod)Cl]2 (0.4 mmol) in
dichloromethane (30 ml) was added a solution of
(2S,4R)-I (0.8 mmol), then ammonium hexafluorophos-
phate (0.8 mmol) was added and the mixture was
stirred for 3 h at 40°C, and filtered. The filtrate was
evaporated under reduced pressure to 2 ml. Careful
addition of diethyl ether caused the precipitation of a
yellow–orange solid which was collected by filtration,
washed with diethyl ether and dried in vacuo to give the
desired cationic complex. Yield 68%. [h[2D5= +8.65°
(c=0.75, CH2Cl2). \(CH3CN) (V−1 cm2 mol−1)=90.
2.2.3. [Rh(cod){(2S,4S)-IV}]PF6 (4)
Yield 60%. [h]2D5= +5.98° (c=1.2, CH2Cl2).
\(CH3CN) (V−1 cm2 mol−1)=68. Anal. Calc. for
C65H74F6N2O6P3RhSi: C, 59.3; H, 5.6; N, 2.1; P, 7.1;
Rh, 7.8. Found C, 58.9; H, 5.2; N, 2.5; P, 6.6; Rh,

A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
189
7.5%. 1H-NMR (CDCl3): l 7.8–7.0 (m, 35H, H
2.3. Preparation of [Rh(CO)2{diphosphinite}]PF6
complexes (5, 6)
phenyl); 5.2 (m, 1H, NCH); 4.2 (br, 4H, CHꢁ); 4.0
(m, 2H, CH2N(CO)); 3.6 (q, 6H, CH2O); 2.9 (m, 2H,
CH2N); 2.5–2.1 (m, 6H, CHCH2, CH2ꢀCHꢁ); 1.8 (m,
4H, CH2ꢀCHꢁ); 1.5 (m, NCH2CH2); 1.2 (t, 9H, CH3);
0.7 (m, 2H, CH2Si). 13C-NMR (CDCl3): l 158.3
(CO); 148.0–140.0 (CaromꢀR); 128.4–125.0 (CaromꢀH);
81.1 (Ph2COP); 80.6–80.1 (ꢁCHcod); 72.2 (PhCOP);
66.5 (CH); 63.0 (CH2N); 58.6 (CH3CH2O); 46.2
(CH2CH); 43.1 (CH2NHCO); 28.5 (CHcod); 23.2
(CH2CH2CH2); 18.8 (CH3CH2O); 8.0 (CH2Si). 31P-
NMR (CDCl3): l 122.0 (br); −143.7 (1J(PꢀF)=714
Hz).
To a solution of [Rh(CO)2Cl]2 (0.4 mmol) in dry
dichloromethane (20 ml) the stoichiometric amount of
the ligand and NH4PF6 was added. The mixture was
heated under reflux for 3 h and the NH4Cl filtered off.
The filtrate was evaporated to dryness, diethyl ether was
added and the yellow–orange powder thus formed
collected, washed with ethyl ether and dried in vacuo.
2.3.1. [Rh(CO)2{(2S,4R)-I}]PF6 (5)
Yield 72%. [h]2D5= +18.65° (c=1.3, CH2Cl2).
Scheme 1.

190
A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
\(CH3CN) (V−1 cm2 mol−1)=88. Anal. Calc. for
C56H46F6NO4P3Rh: C, 60.7; H, 4.3; N, 1.3; P, 8.4; Rh,
9.3. Found C, 60.9; H, 4.4; N, 0.9; P, 8.0; Rh, 8.8%. 1H-
NMR (CDCl3): l 8.0–7.2 (m, 40H, H phenyl); 4.2 (m,
1H, NCH); 3.6–3.4 (AB, 2H, CH2Ph); 3.0 (m, 2H,
CH2N); 2.5 (m, 3H, CHCH2); 1.9 (m, 1H, CHCH2). 13C-
Table 1
Asymmetric hydrogenation of Z-(a)-ethyl acetamidocinnamate (cat./
subs.=1/100, T=333 K; PH =5 atm)
2
Catalyst
Time (h) (% conv.) a
TOF b
ee (%) c
1
2
7(90)
3(85)
5(80)
3(90)
1654
1667
3871
3325
7
12
8
NMR (CDCl3):
l 190.5 (br, CO); 147.0–138.5
Zeol-3
Zeol-4
(CaromꢀR); 129.0–125.0 (CaromꢀH); 78.4–78.0 (Ph2COP,
PhCOP); 71.3 (CH); 68.8 (CCH2N); 60.6 (PhCH2); 45.3
(CCH2CH). 31P-NMR (CDCl3): l 127.1 (br); −143.8
(1J(PꢀF)=714 Hz).
14
a Total conversion was achieved in 9 h.
b mmol substrate/mmol [Rh] h.
c S configuration, ee measured by optical rotation.
2.3.2. [Rh(CO)2{(2S,4R)-III}]PF6 (6)
Yield 65%. [h]2D5= +8.3° (c=1.2, CH2Cl2).
\(CH3CN)(V−1 cm2 mol−1)=66. Anal. Calc. for
C59H62F6N2O8P3RhSi: C, 56.0; H, 4.9; N, 2.1; P, 7.4; Rh,
8.1. Found C, 56.8; H, 5.0; N, 1.7; P, 7.5; Rh, 8.4%. 1H-
NMR (CDCl3): l 7.8–7.1 (m, 35H, H phenyl); 4.2 (br,
m, 1H, NCH); 3.9 (m, 6H, CH2O); 3.7 (m, 2H,
CH2N(CO)); 3.0 (m, 2H, CH2N); 2.1 (m, 2H, CHCH2);
1.5–1.3 (m, NCH2CH2); 1.2 (t, 9H, CH3); 0.57 (m, 2H,
CH2Si). 13C-NMR (CDCl3): l 190.6 (br, CO), 146.0–
145.0 (CaromꢀR); 128.8–124.5 (CaromꢀH); 81.6
(Ph2COP); 80.2 (PhCOP); 67.4 (CH); 63.3 (CH2N); 58.6
(CH3CH2O); 46.4 (CH2CH); 43.5 (CH2NHCO); 23.2
(CH2CH2CH2); 18.7 (CH3CH2O); 8.0 (CH2Si). 31P-
NMR (CDCl3): l 126.5 (br); −143.8 (1J(PꢀF)=714
Hz).
at 338 K and 5 atm of dihydrogen pressure. The molar
ratio catalyst/substrate was 1/100. The olefin (2 mmol)
was added to a solution (homogeneous catalysts) or sus-
pension (heterogenised catalysts) of the catalyst in
ethanol (45 ml) and the reactor was purged with nitro-
gen. The reaction mixture was heated to 338 K, pres-
surised with dihydrogen and stirred. The results were
monitored by HPLC on a C18 reversed phase using 1:1.1
methanol–water as eluent with UV detection at 228 nm.
After hydrogenation, the homogeneous catalyst was re-
moved by filtration through a short column of Celite,
whilst the heterogenised supported was easily separated
by filtration, and used in a new run. Hydrogenation re-
sults are shown in Table 1.
2.5.2. Hydroformylation reaction, general procedure
Hydroformylation experiments were performed in a
100 ml stainless steel autoclave; toluene or
dichloroethane was used as the solvent. Typical reaction
conditions were catalyst/substrate molar ratio of 1/700,
50 ml toluene as solvent, temperature of 343 K, 20 bar
1:2 H2–CO.
The autoclave was charged with the catalyst, solvent
and substrate. When thermal equilibrium was reached,
the gas mixture was introduced until the desired pressure
was reached, and then samples were taken at regular
time periods and analysed by GLC. Optical yield was de-
termined by GLC using a cyclodextrine capillary column
[10]. The hydroformylation activities are represented by
the turnover frequencies (Table 2).
2.4. Heterogenisation of Rh(I) complexes on
USY-zeolite
The supported Rh(I) complexes (Zeol-3, Zeol-4, Zeol-
6) were prepared as we have previously described [6a,7a].
Thus, a solution of 3 (0.2 mmol) in dry dichloromethane
(2 ml) was added to a well-stirred toluene suspension (40
ml) of the inorganic support (modified USY-zeolite dried
at 140°C/0.1 mm/Hg for 3–4 h, 1 g) and the mixture was
stirred at r.t. for 24 h. The solid was then filtered and
Soxhlet-extracted with 1:2 dichloromethane–ethyl ether
for 7–24 h to remove the remaining non-supported com-
plex from heterogenised catalyst. The remaining pale-
yellow solid was dried in vacuo and analysed.
2.5. Catalytic experiments
The catalytic properties of the above homogeneous
and heterogenised Rh complexes were studied under
conventional conditions for batch reactions in a reactor
(Autoclave Engineers) of 100 ml capacity at tempera-
tures, pressures and Rh/substrate molar ratios indicated
below in each particular case. The results were moni-
tored by GLC or HPLC.
3. Results and discussion
3.1. Synthesis and characterisation of diphosphinite
ligands
Phosphinite ligands are generally synthesised by reac-
tion of an amino alcohol with the appropriate
chlorophosphine [11]. Accordingly, the diphosphinite
ligands were synthesised by following an analogous pro-
cedure. The starting alcohols (2S,4R), (2S,4S)-1-
2.5.1. Hydrogenation of prochiral olefins
Ethyl (Z)-a-acetylaminocinnamate, selected as a
model compound, was hydrogenated in a batch reactor

A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
191
and (2S,4R)-VI, could not be fully characterised. IR
spectra of freshly prepared samples were obtained. The
most characteristic IR bands are those of the 1200–950
cm−1 region, and the characteristic band of free
diphenylphosphinic acid at 955 cm−1 is of medium
intensity. The 31P-NMR shows the signals mentioned
before (a broad signal at 26.2 (V), 22.5 (VI) ppm) and
a sharp singlet at 28.64 ppm corresponding to DPA).
benzyl - 4 - hydroxy - 4 - phenyl - 2 - (1,1 - diphenylmethyl)-
pyrrolidinylmethanol and (2S,4R), (2S,4S)-1-(3-tri-
ethoxysilyl)propylaminocarbonyl-4-hydroxy-4-phenyl-
2-(1,1-diphenylmethyl)pyrrolidinylmethanol were ob-
tained from the corresponding optically active amino
acid (L-hydroxyproline) following procedures previ-
ously described in the literature [8]. The diols were
reacted with two equivalents of chlorodiphenylphos-
phine in dry diethyl ether, under nitrogen, in the pres-
ence of triethylamine (r.t., 24 h) to give the
corresponding ligands in 60–80% yields (Scheme 2).
The benzyl pyrrolidine derivatives are white powders,
and the ligands with the triethoxysilyl group are colour-
less oils; they were characterised by microanalysis, and
NMR (1H-, 13C-, 31P-) spectroscopy (Section 2). These
products are air-sensitive at r.t. and slowly oxidise to
give OP(O)Ph2 derivatives. The RꢀOP(O)Ph2 bonds are
prone to hydrolysis and mixtures of composition
RꢀOPPh2·DPA (DPA=diphenyl phosphinic acid,
HOP(O)Ph2) were formed by hydrolysis of the partially
oxidised products, from which colourless crystals of
DPA could be isolated. Totally oxidised and only par-
tially hydrolysed products were also isolated.
3.2. Synthesis of rhodium complexes
The dimeric [Rh(cod)Cl]2, two equivalents of
NH4PF6 and two equivalents of freshly prepared lig-
ands (I–IV) were reacted (CH2Cl2, 313 K, 3 h) leading
to the corresponding complexes 1–4 (Scheme 2). The
complexes were isolated as yellow and relatively air-sta-
1
ble solids. Microanalytical, conductivity, IR, H- and
31P-NMR spectroscopic data given in Section 2 confirm
the proposed structures. Some precautions, i.e. low
temperatures and the exclusion of water and hydroxylic
solvents, have to be taken, especially in the case of
triethoxysilane derivatives, in order to avoid hydrolysis
and subsequent polymerisation. Since no suitable crys-
tals could be obtained for carrying out the X-ray work,
the structures of the complexes were determined by IR
and NMR spectroscopy and elemental analyses.
1H resonance patterns of the ligands were complex
because they are strongly coupled systems. Neverthe-
less, complete assignment of almost all resonances was
2
1
possible on the basis of D-NMR (1H-, H-COSY). The
31P-{1H}-NMR spectra of freshly prepared samples
were obtained. The spectrum of (2S,4R)-I exhibits a
broad signal at l=121.8 ppm while l=118.8 ppm
for (2S,4S)-II. In all samples, some signals for the
oxidised and partially hydrolysed compounds (RꢀOP-
(O)PPh2·DPA) appear at 26.2 ppm for the (2S,4R)-I
derivative and 22.5 ppm for (2S,4S)-II, whereas the
corresponding signal of DPA is present as a sharp
singlet at 28.6 ppm.
The elemental analyses are in agreement with the
stoichiometry proposed for bidentate structures with
1:1 Rh–P-donor ratio. The IR spectra show bands for
a coordinated diene (cod) and for the P-donors that
appear at their expected frequencies. A band at ꢀ845
cm−1 that corresponds to w(PꢀF) was also present.
The 1H- and 13C-NMR spectra show the signals
corresponding to pyrrolidine protons and carbons
slightly downfield shifted with respect to the free lig-
ands. The signals of the atoms close to the metal are
remarkably broadened due to metal interactions and to
Because of their sensitivity to hydrolysis (2S,4R)-V
Table 2
Styrene hydroformylation using diphosphinite–rhodium catalysts a
Catalyst
t (h)
Conv. (%) b
Aldehydes (%)
Branched c
TOF d
Normal
1
94
72 e
96
80
36
96
86
83
10
78
75
87
78
80
36
77
43
45
51
53
49
63
32
57
55
40
42
51
838
24
933
2104
7336
2188
2620
2
5
Zeol-3
Zeol-4
Zeol-6
a Conditions: cat./substrate=1/700, P, 20 bar CO/H2, T=343 K.
b Conversion into aldehydes.
c eeB10% in all cases.
d mmol substrate/mmol [Rh] h.
e T=313 K.

192
A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
Scheme 2.
the conformational non-rigidity on the NMR time
scale. The cyclooctadiene and aromatic rings give rise
to resonances in the expected positions. 31P-NMR spec-
tra showed a broad signal centred at 126 ppm for 1 and
122 ppm for 2. The signal for the PF6 anion appears at
diphosphinite and anion that appear in the expected
positions. The 1H- and 13C-NMR spectra show the
signals corresponding to pyrrolidine protons and car-
bons slightly downfield shifted as well as broadened
signals for the atoms spacially close to the metal. The
two carbonyl ligands appear as a sharp single resonance
at l=190.6 ppm.
1
−143.8 ppm with J(PꢀF)=714 Hz.
Partially oxidised compounds react with [RhCl(cod)]2
to yield complexes with phosphine oxide ligands (7, 8).
IR for these compounds show w(PꢁO) at 1200 cm−1. In
the case of RꢀOP(O)Ph2 the peak at 26.2 ppm in the
31P-NMR of the free ligand is shifted to 32.1 ppm when
the Rh complex was formed. The shift indicates a
coordination of the ligand with the rhodium through
the phosphine oxide. These compounds could not be
fully characterised due to the presence of diphenylphos-
phinic acid as contaminant.
3.3. Heterogenisation of Rh(I) complexes on
USY-zeolite
Preparations of zeolite heterogenised complexes for
complexes 3, 4, and 6 bearing a triethoxysilylpropyl
group were carried out by controlled hydrolysis of
SiꢀOEt bonds and reaction with the free silanol
(SiꢀOH) on the surface of a USY-zeolite. The resulting
catalytic material is very stable and the species are
covalently bonded to the surface. IR and UV–vis spec-
troscopy confirmed the fact that structures of the start-
ing complexes are maintained when attached to the
surface. The elemental analysis of C, H, N and Rh also
confirm the 1:1 Rh–ligand stoichiometry. It is unlikely
that the nature of the complex is substantially altered
under the relatively mild conditions of the anchoring
reaction [12]. The loading of metal is always ca. ꢀ1–
2% (90.1%) measured by atomic absorption of metal
of the digested samples. These values have been used
for calculating the ratio catalyst/substrate in the reac-
tion tests.
The same form, starting from binuclear rhodium
complex [{Rh(CO)2Cl}2] and ligands (2S,4R)-1-benzyl-
4-diphenylphosphinoxy-4-phenyl-2-(1,1-diphenyl-1-
diphenylphosphinoxymethyl)pyrrolidine (2S,4R)-I, or
(2S,4R) - 1 - (3 - triethoxysilyl)propylaminocarbonyl - 4-
diphenylphosphinoxy - 4 - phenyl - 2 - (1,1 - diphenyl - 1-
diphenylphosphinoxymethyl)pyrrolidine (2S,4R)-III, in
CH2Cl2 and in the presence of NH4PF6, mononuclear
complexes [Rh(CO)2(ligand)]PF6 (5, 6) were obtained in
good yield. The compounds synthesised were character-
1
ised by elemental analysis, IR and H- and 13C-NMR
spectroscopies. IR spectra show two intense w(CO)
bands at 2072 and 1998 cm−1, assigned to two cis-di-
carbonyl vibrations and bands corresponding to

A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
193
vents (e.g. methanol) (Z=83.6 kcal mol−1) led to
medium conversions of olefin to aldehydes. Only non-
coordinating solvents of low polarity such as CH2Cl2
and toluene (Z=54.0 kcal mol−1) gave high conver-
sions to aldehydes with medium branched/linear ratio
(c.f. the same behaviour observed in some other hy-
droformylation systems [14]).
The temperature has a remarkable effect on reac-
tion rate and selectivity. The rate of hydroformylation
increases when increasing the reaction temperature,
but the selectivity strongly decreases [15]. Optimal re-
sults were obtained at 343 K and therefore this was
chosen as the standard temperature in most of our
experiments. The reaction can be performed at r.t. in
excellent selectivity but the reaction rate is too low
(Table 2). Thus, the selected conditions for testing the
catalysts for hydroformylation are 70°C and 20 bar,
and a ratio of 14:6 CO–H2. In these conditions high
conversion (\75%) and a branched/linear ratio near
unity was obtained (Table 2).
An examination of Table 2 shows that in all cases
high conversions and yields of oxo products, low re-
gioselectivities and poor enantioselectivities (B10%),
were obtained. The racemization phenomena probably
take place during the long time required to obtain
high substrate conversion.
When transition metal complexes have been sup-
ported on USY-zeolite, no induction period was
observed and the turnover numbers for hydro-
formylation increases. With these rhodium systems,
ethylbenzene, the hydrogenation product of styrene
was not detected. These catalysts can be recover-
ed and reused retaining most of their catalytic
activity.
4. Reactivity
4.1. Hydrogenation of dehydroaminoacid deri6ati6es
The homogeneous and heterogenised Rh catalysts
(1, 2 and Zeol-3, Zeol-4, respectively) were tested for
enantioselective hydrogenation of ethyl (Z)-a-ac-
etamido- and a-benzamidocinnamates (1/100 catalyst/
substrate ratio) in mild conditions yielding the
corresponding phenylalanine derivatives with quantita-
tive conversion and up to 14% ee (Table 1). For the
two substrates tested the homogeneous complexes
present an induction period due to slow formation of
catalytic active species. On the contrary, when com-
plexes were supported on a modified USY-zeolite, us-
ing the same amount of rhodium as homogeneous
catalysts, no induction period can be detected, proba-
bly as a consequence of the known strong capability
of zeolites to absorb H2 on their surfaces, which in-
creases the local concentration of hydrogen. This be-
haviour was found in other zeolite supported Rh
complexes [6b,13]. When transition metal complexes
have been supported on carriers such as polymers or
silica, etc. it is generally accepted that a moderate to
strong reduction in the reactivity has been observed;
however, in our case the turnover numbers for hydro-
genation increase slightly indicating the cooperative
effect of the support. The enantioselectivities obtained
were low and close to the results for homogeneous
catalysts. The low enantioselectivity is presumably due
to the loss of the bidentate coordination of the ligand
with the metal because of the opening of the seven–
membered ring.
4.2. Hydroformylation of styrene
Hydroformylation of styrene with a 2:1 mixture of
carbon monoxide and hydrogen was carried out mak-
ing use of the zeolite–metal complexes (Zeol-3, Zeol-
4, Zeol-6). The results were compared with those
obtained with homogeneous metal complexes (1, 2, 5).
We have studied the influence of the different parame-
ters on the reaction, especially the role of the support
on the intrinsic activity of the heterogenised catalysts,
as well as the surface concentration effect and the
geometrical constraints.
In the following, chemical yield, selectivity and op-
tical yield obtained with varying parameters such as
nature of the solvent, reaction temperature and cata-
lyst type, are given.
The nature of the solvent is critical to the rate and
selectivity of the hydroformylation results. Solvents
such as acetone with medium rated polarity (Z=65.7
kcal mol−1) and coordinating properties gave low
conversions to aldehydes, while very polar protic sol-
5. Conclusions
A facile procedure for the synthesis of bis-phos-
phinites has been found. The new rhodium–phos-
phinite complexes are active for the hydrogenation
and hydroformylation of olefins. Heterogenisation on
USY-zeolite (containing supermicropores and a large
quantity of silanol groups) increases the activity of
the homogeneous catalysts for different substrates.
The heterogenised complexes are significantly more
stable than their corresponding homogeneous com-
plexes over prolonged reaction times.
To summarise, zeolite-supported complexes show
interesting catalytic properties in hydrogenation and
hydroformylation reactions and these properties are
related to the changes in the microenvironment of the
ligand metal-complex, caused by the support. These
catalysts can be recovered and reused, retaining most
of their catalytic activity.

194
A. Fuerte et al. / Journal of Organometallic Chemistry 588 (1999) 186–194
(Eds.), Comprehensive Coordination Chemistry, vol. 2, Perga-
Acknowledgements
mon, Oxford, 1987, p. 739.
[5] A. Corma, M. Iglesias, M.V. Mart´ın, J. Rubio, F. Sa´nchez,
Tetrahedron: Asymmetry 3 (1992) 845.
The authors are grateful for the financial support
from Direccio´n General de Investigacio´n Cient´ıfica y
Te´cnica of Spain (Project MAT97-1016-c02-02).
[6] (a) A. Corma, A. Carmona, M. Iglesias, A. San Jose´, F.Sa´nchez,
J. Organomet. Chem. 492 (1995) 11. (b) A. Corma, M. Iglesias,
C. Del Pino, F. Sa´nchez, J. Organomet. Chem. 431 (1992) 233.
[7] (a) A. Corma, A. Carmona, M. Iglesias, F. Sa´nchez, Inorg.
Chim. Acta 244 (1996) 239. (b) A. Corma, A. Carmona, M.
Iglesias, F.Sa´nchez, Inorg. Chim. Acta 244 (1996) 79.
[8] A. Corma, A. Fuerte, M. Iglesias, F. Sa´nchez, J. Mol. Catal. A:
Chem. 107 (1996) 225.
[9] (a) J. Chatt, L.M. Venanzi, J. Chem. Soc. (1957) 4715. (b) D.
Drew, J.R. Doyle, Inorg. Synth 13 (1972) 48. (c) G. Giordano,
R.H. Crabtree, Inorg. Synth 28 (1990) 88.
[10] E. Miranda, F. Sa´nchez, J. Sanz, M.I. Jimenez, I. Martinez-
Castro, J. High Resol. Chromatogr. 21 (1998) 225.
[11] A. Mortreux, F. Petit, G. Buono, G. Peiffer, Bull. Soc. Chim. Fr.
4 (1987) 631.
References
[1] Reviews: (a) H. Takaya; T. Ohta; R. Noyori, in: I. Ojima (Ed.),
Catalytic Asymmetric Synthesis, VCH, New York, 1993, Ch. 1.
(b) R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wi-
ley, New York, 1994, p. 16. (c) J.M. Brown, Chem. Soc. Rev.
(1993) 25. (d) H. Brunner, in: E.L. Eliel, S.H. Wilen (Eds.),
Topics in Stereochemistry, vol. 18, Wiley, New York, 1988, p.
129. (e) I. Ojima, N. Clos, C. Bastos, Tetrahedron 45 (1989)
6901.
[2] S. Akutagawa in: A.N. Collins, G.N. Sheldrake, J. Crosby
(Eds.), Chirality in Industry, Wiley, Chichester, UK, 1992, p.
325. (b) S. Kohta, Tetrahedron 50 (1994) 3639. (C) S. Akuta-
gawa, Appl. Catal. A: Gen. 128 (1995) 171.
[12] L.L. Murrell, in: J.J. Burton, R.L. Garten (Eds.), Advanced
Materials in Catalysis, Academic Press, New York, 1977, Ch. 8.
[13] A. Corma, M. Iglesias, F. Sa´nchez, Catal. Lett. 32 (1995) 313.
[14] I. Amer, H. Alper, J. Am. Chem. Soc. 112 (1990) 3674.
[15] R. Lazzaroni, G. Ucello-Barretta, M. Benetti, Organometallics 8
(1989) 2323.
[3] U. Nagel, E. Kinzel, J. Chem. Soc. Chem. Commun. (1986)
1089.
[4] S.M. Laurie, in: G. Wilkinson, R.R. Gillard, J.A. McCleverty
.