Synthesis of a tetramethoxy and an amphiphilic tetrahydroxy hemilabile N,P,N-ligand. Coordination behavior towards rhodium(I) and application to hydroformylation of styrene or hydrogenation of trans-cinnamaldehyde

Article abstract of DOI:10.1016/S0022-328X(01)01116-0

A tetramethoxy hemilabile N,P,N-ligand and the corresponding amphiphilic tetrahydroxy ligand have been synthesized via ortho-lithiation of N,N-bis(2-methoxyethyl)-benzenamine and N,N-bis[2-(methoxymethoxy)ethyl]-benzenamine, respectively. The coordination behavior of the ligands towards rhodium(I) in solution was investigated and, according to ¹H-, ¹³C- and ³¹P-NMR data, at room temperature both ligands are coordinated to the metal in a N,P,N-tridentate mode without any Rh-O interaction, whereas at low temperature a P-Rh-N bridged dimeric species was also found together with the most favored tridentate monomer. The rhodium complex of the former ligand was applied to hydroformylation of styrene and the complex of the latter ligand was evaluated in the hydrogenation of trans-cinnamaldehyde.

Full text of DOI:10.1016/S0022-328X(01)01116-0

Journal of Organometallic Chemistry 634 (2001) 90–98
www.elsevier.com/locate/jorganchem
Synthesis of a tetramethoxy and an amphiphilic tetrahydroxy
hemilabile N,P,N-ligand. Coordination behavior towards
rhodium(I) and application to hydroformylation of styrene or
hydrogenation of trans-cinnamaldehyde
Ioannis D. Kostas *
National Hellenic Research Foundation, Institute of Organic and Pharmaceutical Chemistry, Vas. Constantinou 48, 116 35 Athens, Greece
Received 31 May 2001; accepted 9 July 2001
Dedicated to Professor Dr Friedrich Bickelhaupt on the occasion of his 70th birthday and in recognition of his important contributions to
organometallic chemistry and their applications to organic synthesis
Abstract
A tetramethoxy hemilabile N,P,N-ligand and the corresponding amphiphilic tetrahydroxy ligand have been synthesized via
ortho-lithiation of N,N-bis(2-methoxyethyl)-benzenamine and N,N-bis[2-(methoxymethoxy)ethyl]-benzenamine, respectively. The
1
coordination behavior of the ligands towards rhodium(I) in solution was investigated and, according to H-, 13C- and 31P-NMR
data, at room temperature both ligands are coordinated to the metal in a N,P,N-tridentate mode without any Rh–O interaction,
whereas at low temperature a P–Rh–N bridged dimeric species was also found together with the most favored tridentate
monomer. The rhodium complex of the former ligand was applied to hydroformylation of styrene and the complex of the latter
ligand was evaluated in the hydrogenation of trans-cinnamaldehyde. © 2001 Published by Elsevier Science B.V.
Keywords: N,P,N-ligand; Hemilabile ligand; Amphiphilic ligand; Cinnamaldehyde; Hydroformylation; Hydrogenation
1. Introduction
application to homogeneous catalysis have been dis-
cussed elsewhere [10–13].
Phosphorus–nitrogen based hemilabile ligands have
received much attention in recent years due to the
improved catalytic activity of their transition metal
complexes [1]. This class of compounds includes triden-
tate ligands which contain mixed donor sets [1,2].
Amongst these, examples have been reported of N,P,N-
ligands, which have been the subject of some recent
studies in homogeneous catalytic processes, mostly in
hydrogenation [3–5]. In addition, phosphine ligands
possessing methoxy or hydroxy groups have received
particular attention due to their potential properties
[1,6–9]. In recent years, a new concept of functionalised
phosphines has also emerged, namely that of am-
phiphilic phosphines, and the benefits offered by their
Hydroformylation and hydrogenation are well estab-
lished as homogeneous catalytic processes and, al-
though numerous of papers have been reported, they
are still research areas of great significance [14]. The
hydroformylation reaction has been recognized as one
of the most important applications of homogeneous
catalysis and represents the best technology for the
synthesis of aldehydes from olefins [14–16]. The selec-
tive hydrogenation of a,b-unsaturated aldehydes is also
a topic of high interest, and has been studied as a
monophasic or aqueous biphasic homogeneous cata-
lytic process [12,17–19]. In this transformation,
rhodium complexes are well-known to effect easily the
CꢀC hydrogenation and the formation of saturated
aldehydes, in contrast to ruthenium or iridium com-
plexes which effect the selective CꢀO hydrogenation,
yielding the corresponding unsaturated alcohols which
are more valuable products [12,17–19].
* Tel.: +30-1-727-3878; fax: +30-1-727-3877.
E-mail address: ikostas@eie.gr (I.D. Kostas).
0022-328X/01/$ - see front matter © 2001 Published by Elsevier Science B.V.
PII: S0022-328X(01)01116-0

I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
91
We have previously reported the synthesis of
rhodium complexes with P,N-ligands possessing a hy-
droxy or a methoxy group, and their application to
hydroformylation and hydroaminomethylation reac-
tions [20,21]. According to these results, the rhodium
complex possessing the methoxy group gave higher
reaction rates towards hydroformylation of styrene
than that with the hydroxy group [20]. The influence of
the hydroxy groups in phosphorus ligands towards
metal-catalyzed transformations has also been reported
several times by other authors [9]. A very good activity
and selectivity towards hydroformylation of styrene
was also exhibited by rhodium complexes with hemi-
labile nitrogen-containing bis(phosphinite) or bis(phos-
phine) ligands, recently reported by our group [22].
These ligands were synthesized from N-phenyldi-
ethanolamine (1) which used as starting material [22].
N-Phenyldiethanolamine (1) was also used for the
synthesis of the tetramethoxy hemilabile N,P,N-ligand
5 and the corresponding amphiphilic tetrahydroxy lig-
and 8, reported in this paper. The solution dynamics of
their rhodium(I) complexes were studied by variable-
temperature NMR spectroscopy, and the different co-
ordination modes of the ligands are discussed. The
complex of the former ligand was evaluated in the
hydroformylation of styrene, whereas that of the latter
ligand was applied to the hydrogenation of trans-
cinnamaldehyde.
tractive reagents for selective carbon–phosphorus bond
formation. It was thus very tempting to try the synthe-
sis of a polyhydroxy phosphorus ligand, containing
additional potential coordination sites, via ortho-lithia-
tion of N-phenyldiethanolamine (1). We have reported
previously the ability of the v-lithioxyalkoxy
(–O(CH2)nOLi) or 2-lithioxyethylamino (–NCH2-
CH2OLi) group to function as an ortho-directing sub-
stituent in the lithiation of aromatic compounds, as a
result of the complex-induced proximity effect process
(CIPE), leading to the ease metallation of v-phenoxy
alcohols and N-(2-hydroxyethyl)-N-methyl-aniline, re-
spectively [20,23,24]. In contrast to these results, the
ortho-lithiation of N-phenyldiethanolamine (1) failed.
The metallation reaction was attempted using a variety
of metallating reagents (n-BuLi, n-BuLi/TMEDA,
s-BuLi, three equivalents) under variable conditions
concerning the solvent (ether, THF, methylcyclohex-
ane) and the temperature (–65 to –40 °C, room tem-
perature or in refluxing solvent), and the reaction
mixture was then quenched with electrophiles, such as
chlorodiphenylphosphine,
dichlorophenylphosphine,
dimethyl disulfide (Scheme 1). Metallation was also
attempted by three equivalents of n-BuLi/TMEDA/
THF under sonication or by formation of the bis-
(sodium or potassium alkoxide) of 1 and subsequent
addition of one equivalent of n-BuLi. Unfortunately,
all attempts led to very discouraging results, as the
work-up procedure gave mixtures containing the start-
ing material 1 and/or a variety of identified by-pro-
ducts. A probable explanation for the unsuccessful
metallation of 1 could be the precipitation of the
bis(alkoxide) which occurred in most cases, thereby
effectively inhibiting the subsequent lithiation, as has
previously been reported for the unsuccessful direct
lithiation of dihydric phenols [25].
2. Results and discussion
2.1. Metallation of N-phenyldiethanolamine. Synthesis
of the ligands
Organolithium compounds are among the most at-
The unsuccessful ortho-lithiation of 1 led us to pro-
tect the hydroxy groups by common protective groups
according to known procedures prior to metallation
[26] (Scheme 1). Unfortunately, the t-butyldiphenylsilyl
ether 2 and the tetrahydropyranyl ether 3 were also
unreactive towards lithiation reagents, probably due to
steric hindrance by the bulky protecting groups. The
most effective substrate towards metallation was found
to be the methyl ether 4. Ligand 5 was prepared in a
69% yield via ortho-lithiation of 4 in ether with one
equivalent of n-BuLi and subsequent reaction of the
resulting aryllithium with 0.5 equivalents of PhPCl2.
The activity of the Me3SiCl/NaI system towards the
cleavage of the methyl groups, however, was too low,
with no synthetic value for the preparation of the
ligand 8.
The most effective protection of the hydroxy groups
in the synthesis of the tetrahydroxy ligand 8, was found
to be the methoxymethyl ether 6 (Scheme 2). ortho-
Lithiation of 6 in ether with one equivalent of n-BuLi
Scheme 1.

92
I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
The solution dynamics of the complexes were studied
1
by variable-temperature H-, 13C- and 31P-NMR spec-
troscopy. In the 31P-NMR spectrum in CD2Cl2 at 300
K, complex 9 shows a doublet at l 40.39 (JRhP=155.7
Hz), which is characteristic for a rhodium complex with
a phosphine which is involved in a N–P five-membered
chelate [20,28]. In the 13C-NMR spectrum of 9 at room
temperature, the four CH2N carbons are represented by
only one peak at l 57.29, indicating their equivalence.
The same occurs for the four OCH3 carbons (l 58.99)
and the four CH2O carbons (l 70.00). The CH2N
resonance at l 57.29 has been slightly shifted (3.02
ppm) to low field compared to that of the free ligand 5,
as a result of Rh–N coordination. In the 1H- and
13C-NMR spectra, the OCH3 resonances, on the other
hand, are almost in the same position as those of the
free ligand, indicating the absence of Rh–O coordina-
tion. According to the above-mentioned data, it can be
concluded that, at ambient temperature, the ligand in
complex 9 is coordinated to rhodium in a N,P,N-tri-
dentate mode without any Rh–O interaction, indicating
that the proposed equilibrium is shifted to the left
(mode 9a, Scheme 4). In the 31P-NMR spectrum of 9,
the decrease of the temperature to 233 K caused a
broadening of the signal (Fig. 1). At 188 K, two
well-separated doublets at l 37.57 (JRhP=156.3 Hz)
and l 48.70 (JRhP=138.1 Hz) with a ratio of about 3:2
were observed, indicating a dynamic equilibrium be-
tween two distinct species in both of which phosphorus
is coordinated to rhodium. The resonance at l 37.57,
with a coupling constant close to that observed at room
temperature, has to be assigned to structure 9a, which
is the most favorable mode even at low temperature. In
the 13C-NMR spectrum at 188 K, the existence of
signals with the above-mentioned chemical shifts corre-
sponding to CH2N, OCH3 and CH2O carbons provide
extra evidence for the presence of mode 9a at low
Scheme 2.
Scheme 3.
Scheme 4.
and subsequent quenching with 0.5 equivalents of
PhPCl2, led to a moderately low yield (30%) of phos-
phine 7, the protecting methoxymethyl groups of which
were removed easily by conc. HCl. Even though ligand
8 possesses four hydrophilic groups (OH), the presence
of the hydrophobic aromatic rings results a very low
solubility of 8 in water, in contrast to the high solubility
in water of an analogous hexahydroxy aminomethyl-
phosphine [27].
temperature.
A dynamic equilibrium between the
N,P,N-tridentate coordination mode 9a and the corre-
sponding monomeric P,N-bidentate mode is unlikely
because the resonance at l 48.70 with a JRhP=138.1
Hz is not characteristic for a N–P five-membered
chelate rhodium complex. The resonance at l 48.70
corresponds to a mode in which a Rh–O interaction
must be excluded for two reasons. The first has to do
with the 13C-NMR spectrum at 188 K, in which two
signals (l 58.18 and 58.43) corresponding to OCH3
carbons, in addition to the signal with chemical shift
corresponding to the OCH3 carbons in mode 9a, are
shifted very close to that of the free ligand 5. The
second reason is that a probable Rh–O coordination
would create an O–P eight-membered chelate complex
with a resonance in the 31P NMR spectrum shifted to
higher field compared to that of the five-membered ring
as a result of the decreased ring strain [29]. The reso-
nance at l 48.70 is in contrast with this consideration,
2.2. Synthesis and characterization of the rhodium
complexes
Treatment of [Rh(COD)2]BF4 with one equivalent of
ligand 5 in dichloromethane or 8 in acetone solution,
yielded the cationic rhodium complexes 9 or 10, respec-
tively, in high yield (Scheme 3). The elemental analyses
are in accordance with the molecular formula of the
complexes, and the observation of the [LRh(COD)]+
ion (L=5 or 8) and the absence of higher aggregates in
the ESI MS spectra suggests their monomeric nature.

I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
93
doublet at l 41.41 (JRhP=150.6 Hz). At room tempera-
ture, in the 1H-NMR spectrum, all the OCH2CH2N
protons are shifted to l 3.59–3.38 and, in the 13C-
NMR spectrum, the four CH2N carbons and the four
CH2O carbons are represented by one peak at l 61.24
and 60.66, respectively. The CH2N resonance has been
slightly shifted (1.45 ppm) downfield compared to that
indicating the absence of Rh–O coordination. In the
13C-NMR spectrum of 9 at 188 K, in addition to the
signals with chemical shifts corresponding to the
OCH2CH2N carbons in mode 9a, we observed several
signals at l 68.42–50.21 for the methylene carbons,
some of which have to be assigned to the NCH2
carbons. These resonances are shifted slightly to lower
or higher field compared with that of the free ligand,
indicating the presence of a mode which contains both
coordinated and non-coordinated nitrogen(s). The
above-mentioned data and analysis indicate that the
mode corresponding to the resonance at l 48.70 in the
31P-NMR spectrum of 9 at 188 K must fill the follow-
ing conditions: (a) N–P five-membered chelation is
ruled out; (b) Rh–O coordination must be excluded; (c)
both coordinated and non-coordinated nitrogen(s) are
present; and (d) phosphorus is fully coordinated to the
metal; in the case of dimeric species, we must postulate
a symmetry in the molecule which makes the two
phosphorus atoms equivalent. These conditions are
fully filled by the proposed mode 9b. In this structure,
it is necessary to consider a P–Rh–N sequence in the
coordination sphere around the metal, as the P–Rh–P
sequence must be excluded due to the lack of P–P
coupling in the 31P-NMR spectrum.
1
of the free ligand 8. In the H-NMR spectrum at room
temperature, the four hydroxy protons have a chemical
shift close to that of the free ligand indicating that there
is no coordination between rhodium and any hydroxy
group [20]. By contrast, the proton resonance of the
coordinated hydroxy group is shifted to a considerably
lower field compared to that of the free ligand [30].
Following the analysis for the analogous complex 9 as
detailed above, these data indicate that the proposed
equilibrium is shifted to the left (mode 10a, Scheme 4).
In the variable-temperature 31P-NMR spectra, the coa-
lescence temperature was observed at 213 K (Fig. 1). At
188 K, two broad signals of different intensity were
observed at approximately 39 and 54 ppm, which have
to be assigned to the modes 10a and 10b, respectively.
In the 1H-NMR spectrum at low temperature, the
resonance corresponding to the four hydroxy protons
has a chemical shift very close to that observed at room
temperature, indicating that association of the hydroxy
group to rhodium does not take place at low
temperature.
Complex 10 displays analogous behavior. In the 31P-
NMR spectrum in acetone-d6 at 300 K, it exhibits a
Fig. 1. Variable-temperature 31P{1H}-NMR spectra for the complexes: (a) 9 and (b) 10. In the spectrum of 10, the singlets at 36.86 and 35.45 ppm,
which, although of very small peak area compared to the adjacent peaks, are prominent in the low temperature spectra, are due to a small amount
of impurities which are barely observable in the room temperature spectrum.

94
I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
Scheme 6.
Scheme 5.
regioselectivities obtained under variable conditions of
pressure and temperature is as expected for hydro-
formylation of styrene using rhodium systems. It is also
of interest to note that, under identical reaction condi-
tions, complex 9 has a lower reactivity than an
analogous five-membered chelate rhodium(I) complex
with a methoxy amino phosphine reported previously
by our group, while the regioselectivity towards the
branched aldehyde was found to be in the same range
[20]. It is possible that the presence of different coordi-
nation modes of 9 in solution could play a role in its
catalytic activity.
The absence of Rh–O coordination in complexes 9
and 10 has also been reported previously by our group
for the analogous rhodium(I) complexes with P,N-lig-
ands possessing a hydroxy or a methoxy group [20]. In
1
the H-NMR spectra of 9 and 10 at room temperature,
of the four COD olefinic protons, only two or three
protons, respectively, are apparent as a broad signal.
This behavior is analogous to that observed for the
above-mentioned P,N-ligands [20]. As explained above,
analytical data of the complexes in the solid state
indicate their monomeric nature. In contrast, there is
evidence for a dynamic equilibrium between two dis-
tinct species for both complexes in solution. The exis-
tence at low temperature of two coordination modes in
the complexes reported in this paper, a monomeric and
another bridged dimeric species, is not usual for di- or
polydentate ligands. However, this behavior is not
unique; it has been reported previously, for example,
that it is possible even at room temperature, for a
2.4. Hydrogenation of trans-cinnamaldehyde
The catalytic activity of complex 10 was tested in the
hydrogenation of cinnamaldehyde (14), chosen as a
model compound of the a,b-unsaturated aldehydes
(Scheme 6; Table 2). The reactions were carried out in
a i-PrOH/H2O (95:5) mixture with a substrate:catalyst
ratio of 497:1. As shown in Table 2, the catalyst
exhibits a good selectivity for C=C hydrogenation,
which yields the hydrocinnamaldehyde (15), over CꢀO
or both CꢀC and CꢀO hydrogenation, yielding the
cinnamyl alcohol (16) or the hydrocinnamyl alcohol
(17), respectively. Decarbonylation products were
present only in traces or non existent. The reaction rate
is temperature and pressure dependent. After 24 h at
P=30 bar, the conversion of cinnamaldehyde was
found to be 42.3, 55.2 and 92.2% at 30, 50 and 80 °C,
repectively (entries 1, 3, 5). Increasing the pressure to
100 bar also increases the reaction rate (entries 6, 7).
After 3 h at P=30 bar and T=50 °C, there was only
a low conversion of cinnamaldehyde (6.1%) with a
quantitative selectivity to hydrocinnamaldehyde 15, as
rhodium
phino)pentane [31].
complex
with
a,v-bis(diphenylphos-
2.3. Hydroformylation of styrene
The rhodium complex 9 was applied to the hydro-
formylation of styrene (11) with a high styrene:9 ratio
(1455:1), under variable conditions of pressure and
temperature (Scheme 5; Table 1). A study of its cata-
lytic activity towards hydroformylation is of interest,
since the species present under these conditions are
unknown. Complex 9 displays good turnover numbers
(up to 1407), a high chemoselectivity for aldehydes
(over 97%) with a very good branched:linear ratio (up
to 95.2%). The variation observed in this work in the
Table 1
Hydroformylation of styrene catalyzed by rhodium complex 9
a
b
c
Entry
T (°C)
P
(bar)
Time (h)
Conversion (%)
Rc (%)
Rbr (%)
TON d
1
2
3
4
5
30
40
40
60
40
100
100
100
100
30
22
22
4
4
22
64.0
97.7
34.4
98.2
72.5
98.3
99.0
99.5
97.0
97.7
95.2
94.2
94.7
86.9
87.6
915
1407
498
1386
1031
A 4 mM solution in CH2Cl2. Styrene:catalyst=1455:1.
a P, initial total pressure of CO/H2 (1/1).
b Rc, chemoselectivity towards aldehydes 12 and 13.
c Rbr, regioselectivity towards branched aldehyde 12.
d Turnover no. (TON)=aldehydes fraction×substrate:catalyst ratio.

I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
95
4. Experimental
a result of the favored CꢀC over CꢀO hydrogenation
(entry 4). When the substrate:catalyst ratio was de-
creased to 99:1, cinnamaldehyde was converted in high
yield (97.8%) after 24 h, under mild reaction conditions
(P=30 bar; T=30 °C) (entry 2). Taking into consider-
ation the selectivity of the reaction for conversions over
92%, we can conclude that the increased temperature
yields a lower selectivity towards 15, which at P=30
bar was found to be 87.7 and 71.0% at 30 and 80 °C,
respectively (entries 2, 5). The hydrogen pressure on the
other hand, does not significantly affect the selectivity
of the reaction. At 30 °C, the selectivity for 15 was
about 88% for a pressure of 30 or even 100 bar (entries
2, 6).
4.1. General
n-BuLi was prepared from lithium metal and n-BuCl
in methylcyclohexane. [Rh(COD)2]BF4 was prepared
according to literature procedure [32–34]. All the other
chemicals were commercially available. All prepara-
tions and catalysis were carried out under argon by
using dry and degassed solvents. THF and diethyl ether
were distilled from 9-fluorenylpotassium and Na/ben-
zophenone, respectively. Hydroformylation and hydro-
genation studies were performed in a stainless steel
autoclave (300 ml) with magnetic stirring. Syngas: CO/
H2 (1/1) (CO, 1.8; H2, 3.0). NMR: Bruker AC 300
(300.13, 75.47 and 121.50 MHz for H, 13C and 31P,
1
1
respectively); H- and 13C-NMR shifts were referenced
to the solvents and the 31P-NMR shifts were referenced
to external 85% H3PO4 in D2O. The distinction of the
CH, CH2 and CH3 carbons in the 13C-NMR spectra
was performed by DEPT-NMR experiments. ESI MS:
Finnigan MAT TSQ 7000. GC–MS (EI): Varian Sat-
urn 2000 with a 30 m×0.25 mm DB5-MS column. GC:
Varian Star 3400 CX with a 30 m×0.53 mm DB5
column. Elemental analyses for C, H, N: Perkin–Elmer
PE 2400 II.
3. Conclusions
Two cationic rhodium(I) complexes with hemilabile
N,P,N-ligands possessing four methoxy or four hy-
droxy goups, respectively, have been synthesized. The
tetramethoxy ligand was prepared via ortho-lithiation
of N,N-bis(2-methoxyethyl)-benzenamine. Metallation
of N-phenyldiethanolamine (1) was found to be unsuc-
cessful and, for that reason, the analogous tetrahydroxy
ligand was prepared via ortho-lithiation of the
methoxymethyl ether of 1. At room temperature both
ligands are bound to rhodium in a N,P,N-tridentate
coordination mode without any Rh–O interaction,
whereas at low temperature a P–Rh–N bridged
dimeric species was also found together with the most
favored tridentate monomer. The complexes were ex-
amined as catalysts for the hydroformylation of styrene
or hydrogenation of cinnamaldehyde, providing a good
activity and selectivity towards the formation of 2-
phenyl-propanal or hydrocinnamaldehyde, respectively.
4.2. N,N-bis[2-[(t-Butyldiphenylsilyl)oxy]ethyl]-
benzenamine (2)
To a mixture of 1 (8 g, 44.14 mmol) and imidazole
(15.2 g, 223.27 mmol), N,N-dimethylformamide (160
ml) was added with an ice-water bath cooling. The
reaction mixture was stirred at room temperature (r.t.)
for 1 h and subsequently t-butyldiphenylchlorosilane
(27 ml, 103.83 mmol) was added while cooling at 0 °C.
After an overnight stirring, DMF was removed by
Table 2
Hydrogenation of cinnamaldehyde catalyzed by rhodium complex 10
a
Entry
T (°C)
P
(bar)
Time (h)
Conversion (%)
Selectivity (%) b
15 16
96.4
TON c
17
3.6
8.5
5.3
–
9.0
9.4
11.5
1
30
30
50
50
80
30
50
30
30
30
30
30
24
24
24
3
24
24
24
42.3
97.8
55.2
6.1
92.2
94.6
95.0
–
210
97
274
30
458
470
472
2d
3
87.7
88.9
100.0
71.0
87.9
86.3
3.8
5.8
–
20.0
2.7
2.2
4
5
6
7
100
100
A 2 mM solution in i-PrOH/H2O (95:5). Cinnamaldehyde/catalyst=497:1.
a P, initial pressure of H2 at r.t.
b (Product/converted 14)×100.
c Turnover no. (TON)=fraction of products (15+16+17)×substrate/catalyst ratio.
d Cinnamaldehyde/catalyst=99:1.

96
I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
evaporation, water was added and the product was
extracted with dichloromethane (3×150 ml). The com-
bined organic layers were dried over Na2SO4, filtered
and evaporated to dryness, yielding 37.2 g of a yellow
viscous oil, which solidified upon standing. It was re-
crystallized from ethanol, yielding 2 (26.3 g, 91%) as a
by evaporation, water was added, the product was
extracted with dichloromethane (3×100 ml), dried over
Na2SO4, filtered and evaporated to dryness. Purification
was carried out by distillation, yielding 4 (28.2 g, 84%),
1
b.p. 105 °C (0.1 mmHg). H-NMR (CHCl3-d, l ppm):
7.23–7.20 (m, 2H, Ar); 6.74–6.67 (m, 3H, Ar); 3.59–
3.54 (m, 8H, OCH2CH2N); 3.37 (s, 6H, OCH3).
13C{1H}-NMR (CHCl3-d, l ppm): 147.75, 129.26,
116.05 and 111.73 (Ar); 70.04 (CH2O); 58.94 (OCH3);
50.88 (CH2N). GC–MS (EI): m/z (relative intensity)
209 ([M+], 23), 164 (100), 132 (16), 106 (12), 91 (6), 77
(15), 59 (25). Anal. Found: C, 68.81; H, 9.24; N, 6.61.
Calc. for C12H19NO2 (209.29): C, 68.87; H, 9.13; N,
6.69%.
1
white solid, m.p. 103–105 °C. H-NMR (CHCl3-d, l
ppm): 7.67 (d, 3J=6.7 Hz, 8H, Ar); 7.47–7.35 (m, 12H,
3
3
Ar); 7.02 (t, J=7.4 Hz, 2H, Ar); 6.57 (t, J=7.1 Hz,
3
3
1H, Ar); 6.31 (d, J=7.9 Hz, 2H, Ar); 3.77 (t, J=6.4
Hz, 4H, CH2O); 3.48 (t, J=6.4 Hz, 4H, CH2N); 1.08
3
(s, 18H, C(CH3)3). 13C{1H}-NMR (CHCl3-d, l ppm):
147.59, 135.59, 133.49, 129.66, 129.10, 127.68, 115.45
and 111.22 (Ar); 60.80 and 52.93 (OCH2CH2N); 26.83
(C(CH3)3); 19.09 (C(CH3)3). ESI MS: m/z 658.5 ([M+
H]+). Anal. Found: C, 76.36; H, 7.85; N, 2.04. Calc. for
C42H51NO2Si2 (658.04): C, 76.66; H, 7.81; N, 2.13%.
4.5. 2,2%-(Phenylphosphinidene)-
bis[N,N-bis(2-methoxyethyl)-benzenamine] (5)
4.3. N,N-bis[2-[(Tetrahydro-2H-pyran-2-yl)oxy]ethyl]-
benzenamine (3) [35]
n-BuLi (1.80 M in methylcyclohexane, 26 ml, 46.80
mmol) was added to a solution of 4 (9 g, 43.06 mmol)
in ether (35 ml), under argon. The reaction mixture was
subsequently stirred at r.t. for 29 h, and then, a solution
of dichlorophenylphosphine (3 ml, 22.12 mmol) in ether
(25 ml) was added dropwise with ice-water bath cooling
and the mixture stirred at r.t. overnight. The volatile
materials were removed by evaporation, water (100 ml)
was added, the product was extracted with toluene (3
× 100 ml), the organic phase washed with water, dried
over Na2SO4, filtered and evaporated to dryness, yield-
ing 13.65 g of an orange oil. Purification was carried
out by column chromatography over silica gel. During
the elution, the eluent was gradually changed in 10%
increments from pure hexane to pure ethyl acetate,
yielding 5 (7.8 g, 69%) as a viscous oil. 1H-NMR
(CHCl3-d, l ppm): 7.29–7.25 (m, 7H, Ar); 7.21–7.17
(m, 2H, Ar); 7.00–6.96 (m, 2H, Ar); 6.67–6.64 (m, 2H,
Ar); 3.27–3.22 (m, 8H, CH2O); 3.19–3.09 (m, 8H,
CH2N); 3.14 (s, 12H, OCH3). 13C{1H}-NMR (CHCl3-d,
l ppm): 154.93–123.87 (Ar); 70.96 (CH2O); 58.46
(OCH3); 54.27 (CH2N). 31P{1H}-NMR (CHCl3-d, l
ppm): –22.63 (s). ESI MS: m/z 525 ([M+H]+). Anal.
Found: C, 68.55; H, 7.92; N, 5.24. Calc. for
C30H41N2O4P (524.64): C, 68.68; H, 7.88; N, 5.34%.
To a solution of 1 (6.5 g, 35.86 mmol) and freshly
distilled dihydropyran (30 ml, 331.67 mmol) in
dichloromethane (150 ml), piridinium-toluene-4-sulfo-
nate (2.0 g, 7.96 mmol) was added and stirred at r.t. for
6 days. The solution was then diluted by the addition of
dichloromethane (100 ml), washed with half-saturated
brine and then with water. The organic phase was dried
over Na2SO4, filtered and evaporated to dryness, yield-
ing 12.65 g of a purple oil. Isolation of pure product
was carried out by column chromatography over silica
gel using a mixture of hexane/THF (5:1) as eluent,
1
yielding 3 (11.5 g, 92%) as a colorless oil. H-NMR
3
(CHCl3-d, l ppm): 7.20 (t, J=7.8 Hz, 2H, Ar); 6.74
3
3
(d, J=8.3 Hz, 2H, Ar); 6.66 (t, J=7.2 Hz, 1H, Ar);
4.59 (s, 2H, CH–THP); 3.91–3.80 (m, 4H, CH2); 3.61–
3.56 (m, 6H, CH2); 3.50–3.46 (m, 2H, CH2); 1.87–1.43
(m, 12H, CH2–THP). 13C{1H}-NMR (CHCl3-d, l
ppm): 147.58, 129.08, 115.70 and 111.48 (Ar); 98.95
(CH); 64.67, 62.05, 50.89, 30.48, 25.29 and 19.34 (CH2).
GC–MS (EI): m/z (relative intensity) 349 ([M+], 7),
221 (36), 150 (100), 85 (51), 77 (12), 55 (54). Anal.
Found: C, 68.70; H, 9.16; N, 3.74. Calc. for C20H31NO4
(349.47): C, 68.74; H, 8.94; N, 4.01%.
4.4. N,N-bis(2-Methoxyethyl)-benzenamine (4) [36]
4.6. N,N-bis[2-(Methoxymethoxy)ethyl]-benzenamine
(6)
A solution of 1 (29 g, 160 mmol) in THF (100 ml)
was added dropwise to a suspension of NaH (previ-
ously washed twice with THF to remove the mineral
oil, 13 g, 80% in NaH, 433 mmol) in THF (100 ml)
cooled with an ice-water bath, and then stirred at r.t.
for 1 h. A solution of dimethyl sulfate (35 ml, 368
mmol) in THF (100 ml) was then added dropwise at
0 °C. After stirring at r.t. overnight, THF was removed
A solution of 1 (12.7 g, 70.07 mmol) in THF (60 ml)
was added dropwise to a suspension of NaH (previ-
ously washed twice with THF to remove the mineral
oil, 5.7 g, 80% in NaH, 190 mmol) in THF (100 ml) and
refluxed for 1 h. After cooling at r.t., a solution of
chloromethyl methyl ether (13 ml, 174.39 mmol) in

I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
97
THF (50 ml) was added dropwise and the resulting
mixture refluxed for 2 h. THF was removed by evapo-
ration, water was added, the product was extracted
with dichloromethane (3×100 ml), dried over Na2SO4,
filtered and evaporated to dryness, yielding 17.1 g of a
brown oil. Purification was carried out by distillation,
yielding 6 (14.4 g, 76%), b.p. 135–140 °C (0.01
mmHg). H-NMR (CHCl3-d, l ppm): 7.22 (t, J=8.0
Hz, 2H, Ar); 6.75–6.67 (m, 3H, Ar); 4.63 (s, 4H,
OCH2O); 3.73–3.69 (m, 4H, CH2O); 3.64–3.60 (m, 4H,
NCH2); 3.36 (s, 6H, OCH3). 13C{1H}-NMR (CHCl3-d,
l ppm): 147.50, 129.25, 116.12 and 111.69 (Ar); 96.60
(OCH2O); 65.04 (CH2O); 55.16 (OCH3); 50.97 (NCH2).
GC–MS (EI): m/z (relative intensity) 269 ([M+], 24),
238 (7), 207 (17), 194 (100), 134 (72), 91 (11), 77 (26), 45
(52). Anal. Found: C, 62.61; H, 8.73; N, 5.43. Calc. for
C14H23NO4 (269.34): C, 62.43; H, 8.61; N, 5.20%.
50 ml), the organic phase washed with water, dried over
Na2SO4, filtered and evaporated to dryness, yielding 8
(0.89 g, 80%, essential pure according to 1H- and
31P-NMR spectroscopy) as a yellow jelly, which soli-
dified
upon
standing.
Recrystallization
from
dichloromethane at –20 °C yielded a white solid, m.p.
1
161–165 °C. H-NMR (acetone-d6, l ppm): 7.41–7.36
1
3
(m, 7H, Ar); 7.26–7.21 (m, 2H, Ar); 7.11–7.06 (m, 2H,
Ar); 6.76–6.72 (m, 2H, Ar); 3.77 (s, 2H, OH); 3.70 (sl
3
br, 2H, OH); 3.35–3.30 (m, 8H, CH2O); 3.10 (t, J=
6.1 Hz, 8H, CH2N). 13C{1H}-NMR (acetone-d6, l
ppm): 157.74–126.50 (Ar); 61.35 and 59.79
(OCH2CH2N). 31P{1H}-NMR (acetone-d6, l ppm): –
27.50 (s). ESI MS: m/z 469.2 ([M+H]+, 7). Anal.
Found: C, 67.12; H, 7.25; N, 6.02. Calc. for
C26H33N2O4P (468.53): C, 66.65; H, 7.10; N, 5.98%.
4.9. [5Rh(COD)]BF4 (9)
4.7. 2,2%-(Phenylphosphinidene)-
bis[N,N-bis[2-(methoxymethoxy)ethyl]-benzenamine] (7)
A solution of the ligand 5 (0.1727 g, 0.33 mmol) in
dichloromethane (15 ml) was added dropwise to the
dark red solution of [Rh(COD)2]BF4 (0.1338 g, 0.33
mmol) in dichloromethane (5 ml) with a dry ice/acetone
cooling bath. The reaction mixture was warmed slowly
to r.t. within 1 h, and stirred at this temperature for an
additional 1 h. The resulting light orange solution was
evaporated to dryness, the residue washed with ether
and dried, yielding 9 (0.2471 g, 91%) as a yellow solid,
n-BuLi (1.80 M in methylcyclohexane, 24 ml, 43.20
mmol) was added dropwise to a solution of 6 (10.6 g,
39.41 mmol) in ether (35 ml), cooled with ice-water
bath, and then stirred at r.t. for 25 h and refluxed for 5
h. After cooling with ice-water bath, a solution of
dichlorophenylphosphine (2.8 ml, 20.65 mmol) in ether
(20 ml) was added dropwise and the mixture stirred at
r.t. overnight. The volatile materials were removed by
evaporation, water (100 ml) was added, the product
was extracted with dichloromethane (3×100 ml), the
organic phase washed with water, dried over Na2SO4,
filtered and evaporated to dryness, yielding 12.05 g of
an orange viscous oil. Purification was carried out by
column chromatography over silica gel. During the
elution, the eluent was gradually changed from hexane/
ethyl acetate (3:1) to (1:1), yielding 7 (3.8 g, 30%) as a
1
m.p. (dec.) 171–174 °C. H-NMR (CH2Cl2-d2, l ppm):
7.94–7.88, 7.61–7.38 and 7.27–7.22 (3×m, 13H, Ar);
5.43 (br s, 2H, COD–CH); 3.55 and 3.31 (2×br m,
16H, NCH2CH2O); 3.18 (s, 12H, OCH3); 2.54 (br s,
4H, COD–CH2); 2.22–2.11 (br m, 4H, COD–CH2).
13C{1H}-NMR (CH2Cl2-d2, l ppm): 154.28–124.87
(Ar); 111.76–111.58 (m, COD–CH); 71.67 (d, JRhC
=
13.4 Hz, COD–CH); 70.00 (s, CH2O); 58.99 (s, OCH3);
57.29 (sl br s, CH2N); 32.03 and 28.33 (2×s, COD–
CH2). 31P{1H}-NMR (CH2Cl2-d2, l ppm): 40.39 (d,
1
viscous oil. H-NMR (CHCl3-d, l ppm): 7.29–7.25 (m,
7H, Ar); 7.20–7.16 (m, 2H, Ar); 7.00–6.96 (m, 2H, Ar);
6.67–6.64 (m, 2H, Ar); 4.42 (s, 8H, OCH2O); 3.33–3.26
(m, 16H, OCH2CH2N); 3.26 (s, 12H, OCH3). 13C{1H}-
NMR (CHCl3-d, l ppm): 154.72–123.83 (Ar); 96.39
(OCH2O); 66.04 (CH2O); 55.01 (OCH3); 54.33 (CH2N).
31P{1H}-NMR (CHCl3-d, l ppm): –22.53 (s). ESI MS:
m/z 645 ([M+H]+); 613 ([M−OCH3]+). Anal.
Found: C, 63.24; H, 7.93; N, 4.04. Calc. for
C34H49N2O8P (644.74): C, 63.34; H, 7.66; N, 4.34%.
J
RhP=155.7 Hz). ESI MS: m/z 735 ([M−BF4]+).
Anal. Found: C, 55.48; H, 6.49; N, 3.04. Calc. for
C38H53BF4N2O4PRh (822.53): C, 55.49; H, 6.49; N,
3.41%.
4.10. [8Rh(COD)]BF4 (10)
A solution of the ligand 8 (0.1107 g, 0.24 mmol) in
acetone (15 ml) was added dropwise to the dark red
solution of [Rh(COD)2]BF4 (0.0960 g, 0.24 mmol) in
acetone (5 ml) with a dry ice/acetone cooling bath. The
reaction mixture was warmed slowly to r.t. within 1 h,
and stirred at this temperature for an additional 1 h.
The resulting light orange solution was evaporated
under reduced pressure to a minimal volume, and addi-
tion of dichloromethane caused the precipitation of a
yellow solid. The mixture was evaporated to dryness,
yielding 10 (0.1650 g, 90%), m.p. (dec.) 198–203 °C.
4.8. 2,2%-(Phenylphosphinidene)-
bis[N,N-bis(2-hydroxyethyl)-benzenamine] (8)
Concentrated HCl (2 ml) was added to the solution
of 7 (1.54 g, 2.39 mmol) in methanol (30 ml) and
refluxed for 1 h. Methanol was evaporated, the residue
was alkalinized by a 10% aqueous solution of NaOH,
the product was extracted with dichloromethane (3×

98
I.D. Kostas / Journal of Organometallic Chemistry 634 (2001) 90–98
1H-NMR (acetone-d6, l ppm): 7.86–7.76, 7.66–7.61,
7.50–7.45 and 7.32–7.28 (4×m, 13H, Ar); 4.12 (br s,
3H, COD–CH); 3.73 (sl br s, 4H, OH, exchangeable
with D2O); 3.59–3.38 (m, 16H, NCH2CH2O); 2.59 (m,
4H, COD–CH2). Four of the COD–CH2 protons are
not observed and are assumed to be obscured by the
acetone-d5 signal. 13C{1H}-NMR (acetone-d6, l ppm):
156.59–126.71 (Ar); 69.97 (COD–CH); 61.24 and 60.66
(NCH2CH2OH). The COD–CH2 carbons are not ob-
served and are assumed to be obscured by the acetone-
d6 signal. 31P{1H}-NMR (acetone-d6, l ppm): 41.41 (d,
References
[1] C.S. Slone, D.A. Weinberger, C.A. Mirkin, Prog. Inorg. Chem.
48 (1999) 233.
[2] J.W. Faller, G. Mason, J. Parr, J. Organomet. Chem. 626 (2001)
181 and references cited therein.
[3] Y. Jiang, Q. Jiang, G. Zhu, X. Zhang, Tetrahedron Lett. 38 (1997)
215.
[4] P. Braunstein, M.D. Fryzuk, F. Naud, S.J. Rettig, J. Chem. Soc.
Dalton Trans. (1999) 589.
[5] P. Braunstein, F. Naud, A. Pfaltz, S.J. Rettig, Organometallics 19
(2000) 2676.
[6] A. Bader, E. Lindner, Coord. Chem. Rev. 108 (1991) 27.
[7] J. Holz, M. Quirmbach, A. Bo¨rner, Synthesis (1997) 983.
[8] K.V. Katti, H. Gali, C.J. Smith, D.E. Berning, Acc. Chem. Res.
32 (1999) 9.
J
RhP=150.6 Hz). ESI MS: m/z 679 ([M−BF4]+).
Anal. Found: C, 52.63; H, 5.83; N, 3.06. Calc. for
C34H45BF4N2O4PRh (766.42): C, 53.28; H, 5.92; N,
3.66%.
[9] A. Bo¨rner, Eur. J. Inorg. Chem. (2001) 327.
[10] G. Papadogianakis, R.A. Sheldon, in: B. Cornils, W.A. Herrmann
(Eds.), Tenside Ligands in Aqueous-Phase Organometallic Catal-
ysis, Ch. 3.2.4, Wiley–VCH, Weinheim, 1998.
[11] A. Buhling, P.C.J. Kamer, P.W.N.M. van Leeuwen, J.W.
Elgersma, K. Goubitz, J. Fraanje, Organometallics 16 (1997) 3027.
[12] E. Valls, J. Suades, R. Mathieu, Organometallics 18 (1999) 5475.
[13] E. Valls, J. Suades, R. Mathieu, J.F. Piniella, A. Alvarez-Larena,
J. Organomet. Chem. 626 (2001) 139.
[14] C.D. Frohning, C.W. Kohlpaintner (Hydroformylation, Ch.
2.1.1), H. Brunner (Hydrogenation, Ch. 2.2), in: B. Cornils, W.A.
Herrmann (Eds.), Applied Homogeneous Catalysis with
Organometallic Compounds, vol. 1, VCH, Weinheim, 1996.
[15] P.W.N.M. van Leeuwen, C. Claver (Eds.), Rhodium Catalyzed
Hydroformylation, Kluwer Academic Publishers, Dordrecht,
2000.
[16] Survey of hydroformylation for 1999: F. Ungva´ry, Coord. Chem.
Rev. 213 (2001) 1.
[17] J.M. Grosselin, C. Mercier, G. Allmang, F. Grass, Organometal-
lics 10 (1991) 2126.
[18] D.J. Darensbourg, N.W. Stafford, F. Joo´, J.H. Reibenspies, J.
Organomet. Chem. 488 (1995) 99.
4.11. Hydroformylation of styrene catalyzed by
rhodium complex 9
In a typical experiment, styrene (2 ml, 17.456 mmol)
and a 4 mM solution of rhodium complex 9 in
dichloromethane (3 ml, 0.012 mmol) were placed under
argon in an oven-dried autoclave, which was then
closed, pressurized with syngas (CO/H2=1:1) to 100
bar and brought to the corresponding temperature.
After the required reaction time, the autoclave was
cooled to room temperature, the pressure was carefully
released and the solution was passed through celite and
1
analyzed by GC, GCMS and H NMR spectra. Con-
versions were determined by GC.
[19] A. Fukuoka, W. Kosugi, F. Morishita, M. Hirano, L. McCaffrey,
W. Henderson, S. Komiya, Chem. Commun. (1999) 489.
[20] I.D. Kostas, C.G. Screttas, J. Organomet. Chem. 585 (1999) 1.
[21] I.D. Kostas, J. Chem. Res. (S) (1999) 630.
[22] I.D. Kostas, J. Organomet. Chem. 626 (2001) 221.
[23] I.D. Kostas, C.G. Screttas, C.P. Raptopoulou, A. Terzis, Tetrahe-
dron Lett. 38 (1997) 8761.
[24] C.S. Salteris, I.D. Kostas, M. Micha-Screttas, G.A. Heropoulos,
C.G. Screttas, A. Terzis, J. Org. Chem. 64 (1999) 5589.
[25] J.M. Saa´, J. Morey, G. Sun˜er, A. Frontera, A. Costa, Tetrahedron
Lett. 32 (1991) 7313.
[26] T.W. Greene, P.G.M. Wuts (Eds.), Protective Groups in Organic
Synthesis, Ch. 2, 2nd ed., Wiley, New York, 1991.
[27] J.G.E. Krauter, M. Beller, Tetrahedron 56 (2000) 771.
[28] I. Le Gall, P. Laurent, E. Soulier, J.-Y. Salau¨n, H. des Abbayes,
J. Organomet. Chem. 567 (1998) 13.
4.12. Hydrogenation of trans-cinnamaldehyde catalyzed
by rhodium complex 10
A typical experiment was performed in an autoclave,
in which air was evacuated and replaced with argon. A
2 mM solution of rhodium complex 10 in i-PrOH/H2O
(95:5) (4 ml, 0.008 mmol) and cinnamaldehyde (0.5 ml,
3.972 mmol) were placed under argon in the autoclave,
which was then closed, pressurized with hydrogen (30
or 100 bar) and brought to the corresponding tempera-
ture. After the required reaction time, the work-up and
the analysis of the products was performed as described
above for hydroformylation.
[29] S. Borns, R. Kadyrov, D. Heller, W. Baumann, A. Spannenberg,
R. Kempe, J. Holz, A. Bo¨rner, Eur. J. Inorg. Chem. (1998) 1291.
[30] A. Bo¨rner, A. Kless, J. Holz, W. Baumann, A. Tillack, R.
Kadyrov, J. Organomet. Chem. 490 (1995) 213.
[31] D.H.M.W. Thewissen, K. Timmer, J.G. Noltes, J.W. Marsman,
R.M. Laine, Inorg. Chim. Acta 97 (1985) 143.
Acknowledgements
[32] G. Giordano, R.H. Crabtree, Inorg. Synth. 28 (1990) 88.
[33] M.A. Bennett, J.D. Saxby, Inorg. Chem. 7 (1968) 321.
[34] M. Green, T.A. Kuc, S.H. Taylor, J. Chem. Soc. (A) (1971) 2334.
[35] It is a known compound; however no experimental details and
analytical data were given: S. Gervat, E. Le´onel, J-Y. Barraud, V.
Ratovelomanana, Tetrahedron Lett. 34 (1993) 2115.
[36] Known compound: T. Ibaya, T. Inagi, T. Mizutani, JP 03181447
A2 (1991); Chem Abstr. 116 (1992) 41064.
The author thanks Marianna Kyttari and Maria
Vojatzi for technical assistance and Eleni Siapi for
elemental analyses and ESI MS measurements. The
investigation was supported by the National Hellenic
Research Foundation.