[Pd0(N,N-chelate)(oIefin)] complexes containing nitrogen chelates derived from carbohydrates and their use in mild hydrogenation of olefins in water 1

Article abstract of DOI:10.1039/b003167o

New nitrogen chelates of formula 6-R-C₅H₃NCH=NR′-2 and R'N=CHCH=NR' (R = H or Me; R′ = 2,3,4,6-tetra-Oacetyl-β-D-glucopyranose residue) were prepared. The nitrogen donor atom is directly linked to a chiral carbon, i.e. the Cl atom of the sugar ring. The ability of the ligands to induce enantioselective co-ordination of prochiral olefins was assessed by preparing palladium(o) complexes of formula [Pd(N,N-chelate)(olefin)]. Hydrophilic complexes obtained by deprotection of the hydroxy groups of the sugar residue were used for hydrogenating alkenes in water. The course of the reaction is strongly influenced by pH, and homogeneous hydrogenation of the double bond takes place only under basic conditions. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b003167o

View Article Online / Journal Homepage / Table of Contents for this issue
0
[
Pd (N,N-chelate)(olefin)] complexes containing nitrogen chelates
derived from carbohydrates and their use in mild hydrogenation of
olefins in water†
a
a
a
b
a
Carmela Borriello, Maria L. Ferrara, Ida Orabona, Achille Panunzi and Francesco Ruffo*
a
Dipartimento di Chimica, Università di Napoli “Federico II”, via Mezzocannone 4,
I-80134 Napoli, Italy. E-mail: ruffo@unina.it
Facoltà di Agraria and Dipartimento di Chimica, Università di Napoli “Federico II”,
b
via Mezzocannone 4, I-80134 Napoli, Italy
Received 19th April 2000, Accepted 9th June 2000
Published on the Web 6th July 2000
New nitrogen chelates of formula 6-R-C H NCH᎐NRЈ-2 and RЈN᎐CHCH᎐NRЈ (R = H or Me; RЈ = 2,3,4,6-tetra-O-
5
3
᎐
᎐
᎐
acetyl-β--glucopyranose residue) were prepared. The nitrogen donor atom is directly linked to a chiral carbon, i.e.
the C1 atom of the sugar ring. The ability of the ligands to induce enantioselective co-ordination of prochiral olefins
was assessed by preparing palladium(0) complexes of formula [Pd(N,N-chelate)(olefin)]. Hydrophilic complexes
obtained by deprotection of the hydroxy groups of the sugar residue were used for hydrogenating alkenes in water.
The course of the reaction is strongly influenced by pH, and homogeneous hydrogenation of the double bond takes
place only under basic conditions.
The design of new ligands for organic synthesis is an open chal-
1
lenge. The main question connected with this field of research
is the introduction of appropriate functions in a ligand back-
bone, which often involves synthetic difficulties and consump-
tion of resources. These obstacles can effectively be by-passed
if suitable auxiliaries are chosen among the plenty of naturally
occurring compounds. The simplest carbohydrates are good
candidates as they are easily available, chiral, and may afford
2
both lipo- and hydro-soluble ligands depending on whether
or not the hydroxyl groups are protected. Furthermore, the
3
impressive literature on carbohydrates describes several
straightforward procedures for their derivatization.
On these grounds, we recently prepared new nitrogen
chelates (1, 1M* and 2 in Fig. 1) based on α--mannose and
4
α--glucose, where the sugar residue was bound to the
N-donor through the C6 atom. Their co-ordination properties
were investigated by preparing lypo- and hydro-soluble com-
0
0
0
plexes of Pd and Pt with general formula [M (N,N-chelate)-
olefin)]. A significant diastereoselectivity in the co-ordination
(
of both dimethyl fumarate and fumarodinitrile (fdn) was
obtained only with ligands of type 2, while it was found that
5
chelates of type 1 induced poor enantioselectivity. This result
can reasonably be ascribed to the distance of the chiral
auxiliary from the metal center, which may not allow the chiral
information to be effectively transferred to the metal co-
ordination sphere.
Thus, we were stimulated to prepare new N,N-ligands with a
chiral carbon closer to the co-ordinating atoms, i.e. by establish-
ing a linkage between the carbohydrate C1 carbon and the N
donor. In this report we describe the successful approach to the
synthesis of the new ligands 3, 4 and 5 (Fig. 1), and their
palladium(0) complexes [Pd(N,N-chelate)(olefin)] I. The avail-
ability of the hydrophilic form of the ligands prompted us to
contribute to the flourishing chemistry in aqueous media.
Beside safety advantages, water soluble catalysts offer the pos-
sibility of extracting the products with an organic solvent, while
Fig. 1 Ligands of types 1–5; M = mannose.
2a,h,i,6
the active water phase can be recycled for further reaction.
Thus, a new water soluble compound was prepared by
hydrolysis of the acetyl groups in [Pd(3)(fdn)]. The resulting
complex and the previously described [Pd(1M*)(fdn)] were
successfully tested as catalysts in the hydrogenation of olefins in
water. The dependence of the performance of the catalysts on
the pH was also investigated.
4a
†
Dedicated to Professor Paolo Corradini on the occasion of his 70th
birthday.
DOI: 10.1039/b003167o
J. Chem. Soc., Dalton Trans., 2000, 2545–2550
This journal is © The Royal Society of Chemistry 2000
2545

View Article Online
Scheme 1 (i) ϩPMe Ph, ϪN ; (ii) ϩ6-R-C H NCHO-2, ϪO᎐PMe Ph; (iii) ϩ0.5 OHCCHO, ϪO᎐PMe Ph.
2
2
5
3
2
2
(
olefin)] complexes depends on the extent of π-back donation to
the metal–alkene bond. Since fdn and ma are both better
9c,10
electron acceptors than dmf,
it is reasonable that their
complexes display higher stability.
Fig. 2 Olefins used in the synthesis of complexes of type I.
NMR spectroscopy
The NMR properties of type I complexes can be summarized
as follows. (i) The most prominent feature of the spectra is the
large high-field shift of the olefin resonances, which is
diagnostic of the above mentioned π-back donation contribu-
Results and discussion
Synthesis of compounds 3, 4 and 5 and of their [Pd(N,N-chelate)-
olefin)] complexes
(
4a,9
tion to the metal-alkene bond.
Since the shift increases
All the ligands display a 1,3-diimine skeleton with 2,3,4,6-tetra-
O-acetyl-β--glucopyranose residues bound to the N atoms
through the C1 carbon atom. More precisely, both nitrogens in
according to the electron-withdrawing ability of the olefin, fdn
11
and ma signals undergo a larger shift.
ii) In the presence of prochiral olefins, such as fdn or dmf,
complexes of type I may exist as diastereomers. Such diastereo-
(
5
(C symmetry) are linked to a sugar moiety, while 3 and 4,
2
4a,9c,d
where one nitrogen belongs to a pyridine unit, display lower
meric pairs are claimed
to interconvert by olefin dissoci-
symmetry (C ). The latter two chelates differ in the presence of
1
ation and subsequent re-co-ordination to the metal center with
the opposite enantioface. When the process is slow on the NMR
H or Me, respectively, in position 6 of the heteroaromatic ring.
The ligands were synthesized by adapting the procedure pre-
12
timescale the signals pertaining to the diastereomers resonate
separately. Alternatively average resonances are observed.
NMR spectra (Table 1) of [Pd(3)(fdn)], [Pd(4)(fdn)], [Pd(3)-
4a
viously used for preparing chelates of type 1 and 2 (Scheme
1
). This involves attainment of an iminophosporane derivative
by reaction of dimethylphenylphosphane with 1-azido-2,3,4,6-
tetra-O-acetyl-β--glucopyranose in dry dichloromethane.
(
dmf)] and [Pd(3)(fdn)] in 0.015 M deuteriochloroform solution
7
disclosed that the olefin exchange is slow at 298 K. Addition of
free olefin (10–30%) caused broadening of the signals due to an
increased exchange rate. Larger amounts of alkene averaged the
signals giving rise to an unique sharp spectral pattern. These
findings are consistent with the generally accepted associative
Subsequent addition of the appropriate carbonyl compound
affords the corresponding ligand, which can be purified by
column chromatography. Complexes of type I (Table 1) with
dimethyl fumarate (dmf), maleic anhydride (ma) or fumaro-
dinitrile (Fig. 2) were prepared in good yield by starting
4a,9c,d
mechanism for olefin exchange in complexes of type I.
from [Pd(dba) ] (dba = dibenzylideneacetone) according to a
2
Accordingly, broadening of the signals is expected to occur also
by simply increasing the complex concentration, and this was
proved in the case of [Pd(3)(fdn)].
4a
well established procedure, eqn. (1). After removal of the
[
Pd(dba) ] ϩ N,N-chelate ϩ olefin →
2
Attempts to obtain the spectra at 328 K resulted in partial
decomposition with formation of palladium, N,N-ligand and
olefin. Only [Pd(4)(fdn)] remained unaltered at this temperature
for several minutes, giving rise to an averaged spectrum.
Comparison of the just mentioned results with those
[Pd(N,N-chelate)(olefin)] I ϩ 2 dba (1)
solvent under vacuum, the complexes were obtained as yellow
to orange microcrystalline solids by column chromatography.
Alternatively, fdn or ma complexes could be prepared through
olefin exchange by starting from the corresponding dmf com-
plexes. Unfortunately, could we not isolate complexes of type
I with ligand 5 (see Experimental). It is worth mentioning
that the corresponding platinum complex [Pt(5)(fdn)] could be
4a
obtained by using ligands of type 1 or 2 discloses that
[Pd(N,N-chelate)(olefin)] complexes with the newly synthesized
nitrogen chelates are less inert and stable. Actually, the spectra
of the complexes containing the former ligands were little
affected by increasing the temperature, and fast olefin exchange
could be observed only in the presence of free olefin.
8
isolated in high yield.
All the isolated complexes could be handled in air at
room temperature and were stored for several weeks at 253 K
without appreciable decomposition. fdn derivatives were found
unaltered after several days in chloroform or dichloromethane
solution, while [Pd(3)(ma)] and dmf complexes of both 3 and 4
slowly decomposed into palladium, olefin and “free” ligand.
(iii) Type I complexes are typically involved in another
dynamic process, i.e. alkene rotation around the metal–olefin
bond. Information on its rate is available only when the olefin
exchange is slow. On the grounds of their rotational symmetry
the two halves of both dmf and fdn are expected to appear
equivalent in the NMR spectra in the presence of fast rotation.
The experimental evidence indicates that the alkene rotates
4a,9
As discussed elsewhere,
the stability of [Pd(N,N-chelate)-
2
546 J. Chem. Soc., Dalton Trans., 2000, 2545–2550

View Article Online
1
a
13
b
Table 1 Selected H and C NMR data for complexes of type I
1
13
H
C
c
c
c
CH᎐N
6-R-py
olefin
C1
C2–C5
C6
olefin
[
Pd(3)(fdn)]
8.71
8.89 (1 H, d)
2.9 (2 H, bd)
91.1
74.5 (1 C, b)
72.9 (2 C, b)
61.8
20.0 (2 C, b)
isomer, 80%
6
8.2 (1 C, b)
d
d
d
isomer, 20%
8.68
8.65
8.92 (1 H, d)
8.81 (1 H, d)
3.0 (2 H, b)
3.95 (2 H, ABq)
92.9
93.3
e
[Pd(3)(dmf)]
74.4 (1 C)
73.5 (1 C)
62.2
43.0 (1 C)
42.3 (1 C)
isomer, 70%
7
6
2.1 (1 C)
8.2 (1 C)f
g
isomer, 30%
8.71
8.67
8.78 (1 H, d)
2.88 (3 H)
3.92 (2 H, nm)
2.93 (2 H, ABq)
92.8
91.1
73.1 (1 C)
74.4 (1 C)
72.8 (1 C)
63.0
61.7
[Pd(4)(fdn)]
19.4 (1 C)
18.7 (1 C)
isomer, 80%
7
6
2.7 (1 C)
8.0 (1 C)
g
isomer, 20%
8.63
8.64
2.91 (3 H)
2.78 (3 H)
93.3
91.1
74.6 (1 C)
6
74.3 (1 C)
73.1 (1 C)
62.3
61.8
19.4 (2 C)
f
8.3 (1 C)
[Pd(4)(dmf)]
3.92 (2 H, ABq)
43.2 (1 C)
41.7 (1 C)
isomer, 70%
7
6
2.5 (1 C)
8.2 (1 C)
isomer, 30%
8.68
8.68
2.80 (3 H)
3.88 (2 H, nm)
4.00 (2 H, b)
91.1
74.7 (1 C)
7
6
62.2
62.3
43.0 (1 C)
42.2 (1 C)
3.6 (1 C)f
8.6 (1 C)
h,i
[
[
[
Pd(3)(ma)]
8.83 (1 H, d)
72.5 (1 C)
41.0 (2 C, b)
7
7
6
2.1 (1 C)
1.3 (1 C)
8.0 (1 C)
h, j
Pd(4)(ma)]
8.63
8.67
8.63
2.81 (3 H)
2.88 (3 H)
2.91 (3 H)
4.00 (2 H, b)
90.9
97.2
98.2
74.4 (1 C)
61.9
63.7
63.7
41.9 (1 C)
41.5 (1 C)
7
7
6
2.9 (1 C)
2.8 (1 C)
8.2 (1 C)
k
l
Pd(3*)(fdn)]
isomer, 55%
2.93 (2 H, ABq)
80.4 (1 C)
78.6 (1 C)
7
7
4.3 (1 C)
1.7 (1 C)
g
l
isomer, 45%
80.7 (1 C)
7
7
7
8.9 (1 C)
5.7 (1 C)
2.3 (1 C)
a
Recorded at 298 K and 250 MHz in CDCl (CHCl , δ 7.26, as internal standard). Abbreviations: no attribute, singlet; b, broad signal; bd,
3
3
b
13
broad doublet; d, doublet; nm, narrow multiplet; ABq, AB quartet. Recorded at 298 K and 62.9 MHz in CDCl ( CDCl , δ 77, as internal standard).
3
3
c
d
e 13
13
Of glucoside. Resonances averaged with those of the major isomer.
C NMR spectrum recorded in CD Cl ( CD Cl , δ 53.8, as internal
2 2 2 2
h i 13
f
g
standard). Other resonance/s hidden by those of the major isomer. Hidden by other signals. Mean resonances pertaining to both isomers.
NMR spectrum recorded in DMSO-d (( CD ) SO, δ 39.5, as internal standard). Only the C1 signals are separate. That of the minor isomer
resonates at δ 93.4. In CD OD (CD HOD, δ 3.31, and CD OD, δ 49.3, as internal standards). The olefin carbon atoms resonate at δ 19.6, 18.7,
C
1
3
j
6
3 2
k
13
l
3
2
3
1
7.6, 16.9. Owing to similar diastereomeric abundance the signals cannot unequivocally be attributed.
slowly at 298 K in both diastereomers of [Pd(4)(dmf)]. The
same holds true at least for the most abundant isomer of the
other fdn and dmf complexes. As for the minor ones, no clear
conclusion could be reached since the pertaining resonances
were often hidden by more intense signals.
Table 2 Diastereomeric ratios of type I complexes at 298 K with
ligands 3 and 4
Ratio
Compound
In CDCl3
In CD Cl
2
2
(
iv) According to the orientation of the olefin substituents,
maleic anhydride derivatives of compounds 3 and 4 may exist as
diastereomeric pairs. Beside olefin exchange, also rotation of
the alkene can cause their interconversion. This occurs rapidly
at room temperature as indicated by the presence of broad
average resonances in the NMR spectra.
[
[
[
Pd(3)(fdn)]
Pd(3)(dmf)]
Pd(4)(fdn)]
80:20
70:30
80:20
70:30a
55:45
65:35
80:20
90:10
75:25
[Pd(4)(dmf)]
[Pd(3*)(fdn)]
a
In CD OD.
3
Enantioselectivity of olefin co-ordination
As mentioned above, type I complexes of fdn and dmf exist as
two diastereomers. According to the NMR features the iso-
meric equilibrium is reached rapidly after dissolution of the
the former chelates are within the ranges 90:10–65:35 and
80:20–70:30, respectively in CD Cl and CDCl . The values
2
2
3
13
did not exceed 55:45 in CDCl when type 1 ligands were
3
complexes. Integration of suitable separated CH᎐N peaks
co-ordinated.
᎐
allowed us to calculate the composition.
This favorable trend could be inferred also in the case of
14
An inspection of Table 2 discloses that our expectation con-
cerning an enhanced stereochemical activity of the new ligands
was correct. In fact, both 3 and 4 discriminate one enantioface
of fdn or dmf more effectively than the closely related ligands
of type 1. Actually, the diastereomeric ratios prompted by
the symmetrical ligand 5. The NMR spectrum of crude
[Pd(5)(fdn)] (see Experimental) disclosed the presence of an
unique diastereomer. We recall that the diastereomeric ratios
prompted by the related type 2 ligands were within 97:3 and
4a
75:25.
J. Chem. Soc., Dalton Trans., 2000, 2545–2550
2547

View Article Online
Synthesis of the water soluble complex [Pd(3*)(fdn)]
olefins displaying different structures or degrees of substitution
is fairly easy.
15
Water soluble palladium(0) compounds are rare. As far as we
In a typical experiment 0.010 mmol of catalyst was dissolved
in 1 mL of water. The flask was evacuated and 0.50 mmol of the
appropriate olefin was added. Hydrogen at atmospheric pres-
sure was then admitted. After stirring at room temperature for
the required time (see below), the organic compounds were
extracted with 1 mL of deuteriochloroform containing 0.10
mmol of hexamethyldisiloxane (HMDS), and the mixture
analysed through NMR spectroscopy. The presence of a known
amount of HMDS us allowed to calculate yields by integration.
By using [Pd(1M*)(fdn)] as catalyst, hydrogenation was com-
plete within two hours and the saturated products were
extracted in high yield (>80–90%). Some decomposition was
however observed, and the residue, tested according to the
know, none contains either a Pd–C bond or a chiral ligand,
4a
while both features are present in the previously described
Pd(1M*)(fdn)]. During this work a new related complex was
[
prepared with the aim of investigating the still unexplored
potential of these species in catalysis (see next paragraph).
Basic hydrolysis of the acetyl groups in [Pd(3)(fdn)] (Scheme
2
) afforded the corresponding hydrophilic [Pd(3*)(fdn)] (3*
16,18
procedure described by Maitlis and co-workers,
was found
to be very active. This result indicates that most of the reaction
occurred under heterogeneous conditions.
As it is known that catalysed hydrogenation of olefins in
6b
water can be influenced by pH, the reactions were run in
aqueous KOH at various concentrations with the same olefin:
catalyst ratio as above. Sugars can epimerize in water in the
presence of base, so the stability of [Pd(1M*)(fdn)] under these
conditions was preliminarily assessed. The complex was found
unaltered after several hours in D O at pH 14.
2
After two hours in 1.0 M KOH, acrylonitrile, methacrylo-
nitrile and crotononitrile were hydrogenated in 70, 50 and 50%
yields, respectively. The metal containing residues recovered by
filtration of the reaction mixture were almost inactive under the
circumstances described by Maitlis and co-workers, which
indicates that hydrogenation is essentially homogeneous under
strongly basic conditions. Reasonably, the suppression of the
heterogeneous contribution determines the observed lower
conversions with respect to those found in water.
When the reaction times were prolonged in order to increase
the yield of products the amount of organic compounds
extracted by chloroform decreased substantially (<30–40%),
while that found in the water phase was negligible. It must be
inferred that undesired concomitant processes occur, which,
however, were not investigated.
For this reason the reactions were run at lower pH values, i.e.
in 0.10 and 0.25 M KOH solutions. In 0.10 M KOH acrylo-
nitrile was hydrogenated within two hours, while methacryl-
onitrile and crotononitrile required three hours to react
completely. In all cases, the organic products were extracted in
good to high yields (70–90%). Analysis of the activity of the
residues disclosed that the homogeneous contribution to the
reactions was substantial. After two hours the yields of satur-
ated products were in the range 60–90%, while the amount of
unchanged olefins was within 0–10%. On the other hand, when
the residues were tested as catalysts under the same conditions
as adopted in the initial runs, the yields were found to be 35–50
and 25–40%, respectively.
Scheme 2
16
indicates the deprotected form of 3). The yellow microcrystal-
line product was isolated in high yield by addition of diethyl
ether to the reaction mixture. The complex, which is fairly
soluble in water and methanol, was characterized through
NMR spectroscopy and elemental analyses. The two expected
diastereomers (in 55:45 ratio) could separately be observed in
CD OD solution at 298 K, while in D O the faster olefin
exchange gave rise to an averaged spectrum. In the former
solvent the presence of four separate signals for the olefin
carbon atoms indicates that also olefin rotation is hindered at
room temperature.
3
2
Hydrogenation reactions
Increasing attention is currently addressed towards homo-
geneous metal catalysed synthesis in aqueous systems.
However, it should be noted that it is difficult in whatsoever
solvent to ascertain whether the reaction proceeds under
2a,h,i,6
16
heterogeneous or homogeneous conditions. It is often found
that both mechanisms operate. As an example, Elsevier and co-
workers showed that [Pd(N,N-chelate)(dmf)] species (N,N
chelate = N,NЈ-diaryldiiminoacenaphthene) promote the hydro-
17a
17b
genation of alkenes
They found
or alkynes
in dry tetrahydrofuran.
17a
that the reaction was homogeneous when
In 0.25 M KOH hydrogenation of acrylonitrile was again
complete within two hours in high yield. The residue was found
almost inactive, i.e. conversion of acrylonitrile into propio-
electron-poor alkenes were employed, while olefins with donor
substituents were mainly hydrogenated by metal particles which
formed in the reacting system.
The availability of the hydrophilic complexes [Pd(1M*)-
fdn)] and [Pd(3*)(fdn)], both fairly similar to the mentioned
palladium(0) species used by Elsevier, offered the opportunity
of investigating their ability in hydrogenating alkenes in water.
Acrylonitrile, methacrylonitrile and crotononitrile (cis/trans
mixture) were chosen as substrates on the following grounds:
20
nitrile was less than 5% after the same time. Instead, the
clear filtered solution was able to hydrogenate freshly added
acrylonitrile with small loss of activity. These results indicate
that the reaction is essentially homogeneous under these
conditions.
(
Methacrylonitrile and crotonitrile were also hydrogenated
in 0.25 M KOH. After two hours saturated products and
unchanged olefins were recovered in 50–60 and 5–25% yield,
respectively. These ranges became 15–30 and 50–60%, respec-
tively, in the reactions promoted by the corresponding residues.
It should be noted that in none of the experiments hydrolysis
of the nitrile was found to occur. This could easily be checked
(
i) the nitrile group can stabilize the corresponding olefin
palladium(0) complexes by effective electron attraction; (ii)
both the olefins and the corresponding hydrogenated products
are not water soluble and can easily be extracted from
the aqueous phase; (iii) the nitrile substituent is more resistant
than other common organic functions (esters or amide) to
hydrolysis; (iv) a comparison on the behavior of strictly related
13
by recording the C NMR spectra of the products. Actually,
signals at δ 120 were diagnostic of the presence of CN groups,
2
548
J. Chem. Soc., Dalton Trans., 2000, 2545–2550

View Article Online
Table 3 Elemental analyses (%) for the new compounds
Compound
Formula
C found (calc.)
H found (calc.)
N found (calc.)
3
4
5
C H N O
54.83 (55.04)
56.25 (56.00)
50.03 (50.28)
46.59 (46.43)
45.27 (45.46)
47.30 (47.29)
46.07 (46.26)
45.16 (44.98)
46.02 (45.85)
42.27 (42.45)
5.69 (5.54)
5.81 (5.82)
5.52 (5.63)
4.34 (4.22)
4.56 (4.70)
4.54 (4.44)
4.75 (4.89)
4.20 (4.09)
4.42 (4.31)
3.93 (4.01)
6.50 (6.42)
6.36 (6.22)
3.79 (3.91)
8.78 (9.02)
4.05 (4.08)
8.94 (8.82)
4.01 (4.00)
4.41 (4.37)
4.27 (4.28)
12.65 (12.37)
2
0
24
2
9
C H N O
2
1
26
2
9
C H N O
3
0
40
2
18
[Pd(3)(fdn)]
[Pd(3)(dmf)]
[Pd(4)(fdn)]
[Pd(4)(dmf)]
[Pd(3)(ma)]
[Pd(4)(ma)]
[Pd(3*)(fdn)]
C H N O Pd
24 26 4 9
C H N O Pd
2
6
32
2
13
C H N O Pd
2
5
28
4
9
C H N O Pd
2
7
34
2
13
C H N O Pd
2
4
26
2
12
C H N O Pd
2
5
28
2
12
C H N O Pd
1
6
18
4
5
1
3
while no resonances were found in the typical range of carbonyl
frequencies.
Complex [Pd(3*)(fdn)] was also tested in the hydrogenation
of the alkene-1-nitriles. On the grounds of the mentioned
findings the reactions were performed in 0.10 and 0.25 M KOH.
A close analogy with the results obtained by using [Pd(1M*)-
δ 7.26, and CDCl , δ 77, as internal standards). The following
3
abbreviations are used for describing NMR multiplicities: no
attribute, singlet; d, doublet; dd, double doublet; m, multiplet; t,
triplet. Specific optical rotatory powers [α] were measured with
a Perkin-Elmer Polarimeter (Model 141) at 298 K and 589 nm
in dichloromethane (c = 1.0 g per 100 mL). 1-Azido-2,3,4,6-
7
(
fdn)] was found, the only difference consisting in a slightly
tetra-O-acetyl-β--glucopyranose was prepared according to a
improved activity in 0.25 M KOH. Once the experimental
conditions which allowed the catalysis to be mostly homo-
geneous were found an enantioselective reduction of the double
bond was attempted. The substrate α-ethylacrylonitrile was
hydrogenated in 0.25 M KOH by either [Pd(3*)(fdn)] or
literature method. Dichloromethane was distilled from calcium
hydride, methanol from magnesium immediately before use.
Elemental analyses for the new compounds are reported in
Table 3.
21
[
2
Pd(1M*)(fdn)]. In both cases the product of the reaction,
Syntheses
-methylbutyronitrile, did not show significant optical rotatory
22
Compounds 3, 4 and 5. To a stirred solution of PPhMe2
power. The same held true when the reaction was performed
in dry THF by using [Pd(4)(dmf)] as catalyst. The unsuccessful
outcome of these experiments is in keeping with the lack of
literature reports on asymmetric hydrogenations of alkenes
catalysed by palladium complexes. Actually, as far as we know,
this catalysis has been achieved only under heterogeneous con-
(
0.14 g, 1.0 mmol) in 5 mL of dry dichloromethane kept in an
ice-bath a solution of 1-azido-2,3,4,6-tetra-O-acetyl-β--gluco-
pyranose (0.37 g, 1.0 mmol) in 5 mL of dry dichloromethane
was added dropwise. After the addition was complete, the ice-
bath was removed and formation of nitrogen was observed. The
progress of the reaction was monitored by TLC analysis on
silica with ethyl acetate–light petroleum (bp 40–60 ЊC) (1:1) as
eluent. The reaction affording the corresponding iminophos-
phorane was complete within 30 minutes. To the solution was
added the appropriate carbonyl compound in stoichiometric
amount (respectively 1.0 mmol of 6-methyl-2-pyridinecarbalde-
hyde or 2-pyridinecarbaldehyde, and 0.50 mmol of glyoxal as a
23
ditions, while no examples of homogeneous hydrogenation
have so far been reported.
Conclusion
This paper describes new chiral nitrogen chelates based on
carbohydrates and the corresponding palladium(0) complexes
of formula [Pd(N,N-chelate)(olefin)]. The nitrogen donor is
directly linked to the chiral C1 atom of a glucose moiety. This
feature enhances the ability of the ligands to induce enantio-
selective co-ordination of prochiral olefins with respect to that
displayed by related chelates where the N-donor is bonded to
the sugar via the methylene C6 atom.
A new water soluble palladium(0) complex was also pre-
pared. This compound and a previously reported related species
were examined as catalysts in the hydrogenation of unsaturated
nitriles in water. It was found that the pH of the solvent plays
a crucial role in determining both yield and reaction course.
More precisely, in distilled water hydrogenation is fast, but
essentially heterogeneous. In 1.0 M KOH hydrogenation is
homogeneous, but parallel undesired reactions lower the yield
of the products. Best conditions for the homogeneous hydro-
genation of acrylonitrile were found in 0.25 M KOH, where the
yield of propionitrile was high. Under these conditions the
reaction involving methacrylonitrile and crotononitrile also
proceeds readily, but a heterogeneous contribution is present.
This latter finding can be rationalized by considering that the
stability of [Pd(N,N-chelate)(olefin)] complexes grows when the
steric hindrance and/or the donor properties of the alkene
decrease. Thus, with substituted alkenenitriles formation of metal
particles may occur faster than in the presence of acrylonitrile.
4
0% w/w water solution). When glyoxal was employed sodium
sulfate was also added. After 15 minutes of stirring the volume
was reduced under vacuum and the mixture filtered on Florisil
(
15 × 1.5 cm) with ethyl acetate–light petroleum (1:1) as eluent.
Removal of the solvents under vacuum afforded the micro-
crystalline products in 70–80% yield. H NMR: 3, δ 8.61
1
(
5
(
2
(
4
1 H, d), 8.51 (1 H), 7.98 (1 H, d), 7.33 (1 H, dd), 7.32 (1 H, t),
.37 (1 H, t), 5.14 (1 H, t), 4.99 (1 H, t), 4.93 (1 H, dd), 4.28
1 H, dd), 4.18 (1 H, dd), 3.89 (1 H, m), 2.07 (3 H), 2.02 (3 H),
.01 (3 H) and 1.99 (3 H); 4, 8.50 (1 H), 7.80 (1 H, d), 7.61
1 H, t), 7.18 (1 H, d), 5.36 (1 H, t), 5.14 (1 H, t), 4.99 (1 H, t),
.88 (1 H, dd), 4.28 (1 H, dd), 4.18 (1 H, dd), 3.89 (1 H, m), 2.57
(
(
3 H), 2.07 (3 H), 2.02 (6 H) and 2.01 (3 H); 5, 8.00 (2 H, d), 5.31
2 H, t), 5.11 (2 H, t), 4.90 (2 H, t), 4.84 (2 H, dd), 4.25 (2 H, dd),
4
.16 (2 H, dd), 3.83 (2 H, m), 2.06 (6 H), 2.04 (6 H), 2.01 (6 H)
13
and 2.00 (6 H). Selected C resonances (at 62.9 MHz): 3,
δ 162.5 (CH᎐N), 149.4 (C6 of pyridine), 136.6, 125.4, 121.5
᎐
(
(
C3–C5 of pyridine), 92.3 (C1 of glucose), 73.7, 73.3, 71.8, 68.3
C2–C5 of glucose) and 62.1 (C6 of glucose); 4, 162.7 (CH᎐N),
᎐
1
7
36.9, 125.1, 118.6 (C3–C5 of pyridine), 92.3 (C1 of glucose),
3.7, 73.4, 71.9, 68.4 (C2–C5 of glucose), 62.1 (C6 of glucose)
and 24.2 (6-Me of pyridine); 5, 161.2 (CH᎐N), 91.7 (C1 of
glucoses), 73.7, 73.3, 71.3, 68.1 (C2–C5 of glucoses) and 61.9
᎐
3
Ϫ1
Ϫ1
(
C6 of glucoses); [α]: 3, Ϫ16; 4, Ϫ45; 5, Ϫ19 deg cm g dm .
[
Pd(N,N-chelate)(olefin)] complexes I. To a suspension of
Experimental
[
Pd(dba) ] (0.58 g, 1.0 mmol) in 5 mL of dry toluene were added
2
NMR spectra were recorded with a 250 MHz spectrometer
the N,N-chelate (1.5 mmol) and the olefin (1.5 mmol). After 1 h
(
Bruker Model AC-250). The solvent was CDCl3 (CHCl3,
of stirring the solvent was removed under vacuum. The residue
J. Chem. Soc., Dalton Trans., 2000, 2545–2550
2549

View Article Online
was separated by chromatography on silica gel with dichloro-
methane (to remove dba) and then with dichloromethane–
methanol (40:1). The solvents were removed under vacuum
from the collected yellow-orange fractions affording the prod-
uct as a yellow-orange microcrystalline solid (yield: 75–85%).
Alternatively, fdn or ma complexes could be prepared by start-
ing from the corresponding dmf complexes, according to the
following example: to a solution of [Pd(3)(dmf)] (0.69 g, 0.10
mmol) in 1 mL of dichloromethane was added fdn (0.012 g,
1999, 64, 5593; (i) R. Selke, J. Holz, A. Riepe and A. Börner, Chem.
Eur. J., 1998, 4, 769; (j) R. Kadyrov, D. Heller and R. Selke,
Tetrahedron: Asymmetry, 1998, 9, 329; (k) J. Costamagna, N. P.
Barroso, B. Matsuhiro and M. Villagràn, Inorg. Chim. Acta, 1998,
2
73, 191; (l) M. Stolmàr, C. Floriani, G. Gervasio and D. Viterbo,
J. Chem. Soc., Dalton Trans., 1997, 1119; (m) M. Tschoerner,
G. Trabesinger, A. Albinati and P. S. Pregosin, Organometallics,
1997, 16, 3447; (n) T. V. RajanBabu, A. L. Casalnuovo and
T. A. Ayers, in Advances in Catalytic Processes, ed. M. P. Doyle,
Jai Press Ltd, London, 1997, vol. 2, ch. 1, pp. 1–41.
3
4
Carbohydrate Analysis, A Practical Approach, eds. M. F. Chaplin
and J. F. Kennedy, IRL Press, Oxford, 1994.
(a) M. L. Ferrara, F. Giordano, I. Orabona, A. Panunzi and
F. Ruffo, Eur. J. Inorg. Chem., 1999, 1939; (b) M. L. Ferrara,
M. Funicello, I. Orabona, A. Panunzi and F. Ruffo, Organo-
metallics, 1998, 17, 3832.
0
.15 mmol). The resulting solution was chromatographed as
described above affording pure [Pd(3)(fdn)] (yield 80–85%).
Attempts to prepare [Pd(5)(fdn)] according to eqn. (1) led to
copious formation of a black precipitate. The deuteriochloro-
form extract was found to contain mainly 5, dba and fdn, while
only 10–20% of the material consisted of the desired complex.
The mother-liquor was found to contain dba in substantial
amounts. Attempts to purify the crude product were unsuccess-
ful. Better results were not obtained either by changing the
solvent of the reaction, or by using other alkenes, i.e. acrylo-
nitrile and dimethyl fumarate.
5
6
This is in keeping with the acknowledged higher discriminating
ability of C symmetrical ligands with respect to those displaying C
2
1
symmetry. See, J. K. Whitesell, Chem. Rev., 1989, 89, 1581.
For examples, see: (a) A. J. Sandee, V. F. Slagt, J. N. H. Reek, P. C. J.
Kamer and P. W. N. M. van Leeuwen, Chem. Commun., 1999, 1633;
(b) K. C. Tin, N. B. Wong, R. X. Li, Y. Z. Li, J. Y. Hu and X. J. Li,
J. Mol. Catal. A, 1999, 137, 121; (c) H. Jiang, Y. Xu, S. Liao, D. Yu,
H. Chen and X. Li, J. Mol. Catal. A, 1999, 142, 147; (d) B. E.
Hanson, Coord. Chem. Rev., 1999, 185–186, 795; (e) D. J.
Darensbourg, J. B. Robertson, D. L. Larkins and J. H. Reibenspies,
Inorg. Chem., 1999, 38, 2473; ( f ) A. Fukuoka, W. Kosugi,
F. Morishita, M. Hirano, L. McCaffrey, W. Henderson and
S. Komiya, Chem. Commun., 1999, 489; (g) L. Johansson, O. B.
Ryan and M. Tilset, J. Am. Chem. Soc., 1999, 121, 1974; (h) M. S.
Goedheijt, J. N. H. Reek, P. C. J. Kamer and P. W. N. M. van
Leeuwen, Chem. Commun., 1999, 2431; (i) J. C. Shi, C. H. Yueng,
D. X. Wu, Q. T. Liu and B. S. Kang, Organometallics, 1999, 18,
[
Pd(3*)(fdn)]. To a stirred suspension of [Pd(3)(fdn)] (0.012
g, 0.20 mmol) in 2.5 mL of dry methanol was added a catalytic
amount of sodium methoxide in the same solvent. After 30
minutes formation of a yellow solution ensued, and the product
was crystallized by addition of diethyl ether. The complex was
separated, washed with diethyl ether (3 × 3 mL) and dried
under vacuum (yield: 90%).
3
796; (j) G. Verspui, G. Papadogianakis and R. A. Sheldon, Chem.
Hydrogenation reactions
Commun., 1998, 401; (k) A. N. Ajjou and H. Alper, J. Am. Chem.
Soc., 1998, 120, 1466; (l) S. Otto, G. Boccaletti and J. B. F. N.
Engberts, J. Am. Chem. Soc., 1998, 120, 4238; (m) F. Joó and
Á. Kathó, J. Mol. Catal. A, 1997, 116, 3; (n) W. A. Herrmann and
C. W. Kohlpaintner, Angew. Chem., Int. Ed. Engl., 1993, 32, 1524.
W. A. Szarek, O. Achmatowicz, Jr., J. Plenkiewicz and B. K.
Radatus, Tetrahedron, 1978, 34, 1427.
In a typical experiment 0.010 mmol of either [Pd(1M*)(fdn)]
or [Pd(3*)(fdn)] was dissolved in 1 mL of water (or appropriate
KOH solution) in a 50 mL flask with a rubber stopper and an
outlet connected to a vacuum line. The flask was evacuated and
7
0
.50 mmol of the appropriate olefin was added with a syringe
8
9
F. Ruffo, unpublished work.
through the rubber stopper. The vacuum line was replaced by
an hydrogen reservoir at atmospheric pressure and the gas
introduced into the flask. After stirring at room temperature the
organic compounds were extracted with 1 mL of deuterio-
chloroform containing 0.10 mmol of hexamethyldisiloxane,
and the mixture analysed through NMR spectroscopy.
(a) A. Selvakumar, M. Valentini, M. Wörle and P. S. Pregosin,
Organometallics, 1999, 18, 1207; (b) R. A. Klein, P. Witte, R. van
Belzen, J. Fraanje, K. Goubitz, M. Numan, H. Schenk, J. M.
Ernsting and C. J. Elsevier, Eur. J. Inorg. Chem., 1998, 319;
(
c) L. Canovese, F. Visentin, P. Uguagliati and B. Crociani, J. Chem.
Soc., Dalton Trans., 1996, 1921; (d) R. van Asselt, W. J. J. Smeets
and A. L. Spek, Inorg. Chem., 1994, 33, 1521.
0 Ts. Ito, S. Hasegawa, Y. Takahashi and Y. Ishii, J. Organomet.
1
1
1
1
Chem., 1974, 73, 401.
Acknowledgements
1
13
1 Average H ( C) NMR ∆δ values for the olefin nuclei are: fdn, 3.3
100); ma, 3.1 (95); dmf, 3.0 (91).
(
The authors thank the Consiglio Nazionale delle Ricerche,
Ministero dell’ Università e della Ricerca Scientifica e Tecno-
logica (Programmi di Ricerca Scientifica di Rilevante Interesse
Nazionale, Cofinanziamento 1998–1999, PRIN 9803243241)
and the CIRAM, Università di Napoli “Federico II” for finan-
cial support, and the Centro Interdipartimentale di Metod-
ologie Chimico-Fisiche, Università di Napoli “Federico II” for
NMR facilities.
2 Henceforth a process which is slow (or fast) on the NMR time scale
will simply be indicated as slow (or fast).
3 The only exception was provided by complex [Pd(1G-Me)(fdn)]
which gave rise to a diastereomeric ratio of 85:15.
1
14 Selected H NMR resonances (δ, in CDCl ): 8.48 (2 H, CH᎐N) 2.98
᎐
3
(2 H, HC᎐CH).
1
5 D. J. Darensbourg, T. J. Decuir, N. W. Stafford, J. R. Robertson,
J. D. Draper, J. H. Reibenspies, A. Kathó and F. Joó, Inorg. Chem.,
1
997, 36, 4218; J. W. Ellis, K. N. Harrison, P. A. T. Hoye, A. G.
Orpen, P. G. Pringle and M. P. Smith, Inorg. Chem., 1992, 31,
026.
3
References
1
6 J. E. Hamlin, K. Hirai, A. Millan and P. M. Maitlis, J. Mol. Catal.,
1980, 7, 543.
1
Proceedings of the 10th IUPAC Symposium on Organo-Metallic
Chemistry Directed Towards Organic Synthesis (OMCOS 10),
Versailles, 1999.
17 (a) R. van Asselt and C. J. Elsevier, J. Mol. Catal., 1991, 65, L13;
(b) M. W. van Laren and C. J. Elsevier, Angew. Chem., Int. Ed., 1999,
38, 3715.
2
For recent examples of ligands based on carbohydrates, see:
1
7a,19
(
a) S. Shin and T. V. RajanBabu, Org. Lett., 1999, 1, 1229; (b) J. C.
18 We found that use of thiophene as suppressor
heterogeneous contribution was unreliable in our case.
19 E. W. Stern and P. K. Maples, J. Catal., 1972, 27, 120.
of the
Shi, Y. Zheng, D. X. Wu, X. Y. Huang, Q. T. Liu and B. S. Kang,
Inorg. Chim. Acta, 1999, 290, 121; (c) S. Yano, S. Inoue, Y. Yasuda,
T. Tanase, Y. Mikata, T. Kakuchi, T. Tsubomura, M. Yamasaki,
I. Kinoshita and M. Doe, J. Chem. Soc., Dalton Trans., 1999, 1851;
20 Appreciable amounts (15–20%) of HOCH CH CN also formed.
2
2
21 J. Kovács and C. S. Marvel, J. Polym. Sci. A, 1967, 5, 563. The olefin
was accompanied by an equimolar amount of acetonitrile.
22 α (589 nm, 293 K, (S)-(ϩ)-2-methylbutyronitrile, neat) = ϩ31Њ.
23 J. R. G. Perez, J. Malthete and J. Jacques, C. R. Acad. Sci., Ser. 2,
1985, 500, 169; K. Harada and T. Yoshida, Naturwissenshaften,
1970, 57, 131; Y. I. Petrov, E. I. Klabunovskii and A. A. Balandin,
Kinet. Katal., 1967, 8, 1384.
(
d) K. Yonehara, T. Hashizume, K. More, K. Ohe and S. Uemura,
Chem. Commun., 1999, 415; (e) H. Junicke, C. Bruhn, R. Kluge,
A. S. Serianni and D. Steinborn, J. Am. Chem. Soc., 1999, 121, 6232;
(
f ) O. Pamies, G. Net, A. Ruiz, S. Castillón and C. Claver, ref. 1,
VSL-15; (g) L. Jessen and J. Heck, ref. 1, P-212; (h) K. Yonehara,
T. Hashizume, K. Mori, K. Ohe and S. Uemura, J. Org. Chem.,
2
550
J. Chem. Soc., Dalton Trans., 2000, 2545–2550