o-Alkyl-substituted aromatic phosphanes for hydroformylation studies: Synthesis, spectroscopic characterization and ab initio investigations

Article abstract of DOI:10.1016/S1381-1169(03)00022-0

In earlier hydroformylation studies modification of the Rh catalyst with o-methyl-substituted or o-ethyl-substituted phosphane ligands have increased regioselectivity to branched aldehydes. The promising results achieved created a need for further studies. Thus, a wider group of o-substituted arylphosphane ligands were synthesized and tested in Rh-catalyzed hydroformylation to support the previous findings. In the Rh-catalyzed hydroformylation of propylene and 1-hexene the ligands increased the formation of branched aldehydes compared to triphenylphosphane. The increasing size of the o-alkyl-substituent affected favorably iso-selectivity.

Full text of DOI:10.1016/S1381-1169(03)00022-0

Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
o-Alkyl-substituted aromatic phosphanes for hydroformylation
studies: synthesis, spectroscopic characterization and
ab initio investigations
Helena Riihimäkia,∗, Pekka Suomalainenb, Heidi K. Reiniusc, Johanna Suutaric,
Sirpa Jääskeläinenb, A.O.I. Krausec, Tapani A. Pakkanenb, Jouni T. Pursiainena
a
Department of Chemistry, University of Oulu, P.O. Box 3000, FIN-90014 Oulu, Finland
b
Department of Chemistry, University of Joensuu, P.O. Box 111, FIN-80101 Joensuu, Finland
c
Department of Chemical Technology, Helsinki University of Technology, P.O. Box 6100, FIN-02015 Hut, Finland
Received 19 September 2002; received in revised form 21 September 2002; accepted 28 October 2002
Abstract
In earlier hydroformylation studies modification of the rhodium catalyst with o-methyl-substituted or o-ethyl-substituted
phosphane ligands have increased regioselectivity to branched aldehydes. The promising results achieved created a need for
further studies. Hence, a wider group of o-substituted arylphosphane ligands, e.g. (2-cyclohexylphenyl)diphenylphosphane,
(2-isopropylphenyl)diphenylphosphane, (2-methylnaphthyl)diphenylphosphane, (2,5-dimethylphenyl)diphenylphosphaneand
(2-phenylphenyl)diphenylphosphane were synthesized and tested in rhodium-catalyzed hydroformylation to support the pre-
vious findings. Characterization of the ligands was made by NMR spectroscopy (1H, 31P{ H}, 13C{ H}, HSQC, COSY-90
and COLOC). Additional parameters for evaluation of the stereoelectronic properties of the ligands were provided by quantum
mechanical calculations and by synthesizing Rh(acac)(CO)(PR3) complexes. In the rhodium-catalyzed hydroformylation of
propene and 1-hexene the ligands increased the formation of branched aldehydes compared to triphenylphosphane. Addition-
ally the increasing size of the o-alkyl-substituent was found to effect favorably to the iso-selectivity.
© 2003 Elsevier Science B.V. All rights reserved.
1
1
Keywords: Phosphane ligands; o-Alkyl-substituents; Hydroformylation; NMR spectroscopy; Ab initio calculations
1. Introduction
duce regioselectively the linear aldehyde, n-butanal.
Recently, interest has increased in selective formation
of the branched form, isobutanal, which finds use in
the production of polyols, such as neopentyl glycol
[4]. This has extended the importance and also the
synthetic modification scope of phosphane ligands.
The interest in the steric and electronic proper-
ties of phosphanes is based on the assumption that
donor/acceptor properties and the bulkiness of the
ligand strongly influence the course of the reaction
at a transition metal center. Tolman’s electronic and
Hydroformylation of propene yielding n-butanal
and isobutanal is the most important hydroformyla-
tion process on industrial scale, providing about 75%
of all oxo chemicals consumed in the world [1–4].
Traditionally the aim of this process has been to pro-
∗
Corresponding author. Tel.: +358-8-5531647;
fax: +358-8-5531603.
E-mail address: helena.riihimaki@oulu.fi (H. Riihimäki).
1381-1169/03/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S1381-1169(03)00022-0

70
H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
steric parameters (ν and θ, respectively) have formed
the basis for the description of steric and electronic
properties of phosphanes during the last 30 years
[5–7]. High hydroformylation rates of olefins have
been found with Rh-catalysts containing bulky o-tert-
butyl-substituted phenylphosphites with large cone
angles (175–180◦) [8–11]. As a result from the large
cone angles, in HRh(CO)n(PR3)m-type catalyst, only
one phosphite coordinates to the Rh-center. The ster-
ically hindered phosphite ligands with strongly ele-
ctron-withdrawing nature (χ-value ∼28) have strong
aptitude for CO dissociation compared with the triph-
enylphosphane (145◦, χ-value ∼13), and the formed
16e− complex easily binds to the olefin, thus initiat-
ing a very fast reaction cycle. Due to the large space
available in comparison with the HRh(CO)2(PPh3)2
system, the reaction favor branched aldehyde prod-
ucts, resulting in a moderate overall linearity [8–10].
The same trend to form 16e− complexes has been no-
ticed with bulky nonlinear alkylphosphane-modified
catalysts [12]. The high rates of isomerization caused
by bulky o-tert-butyl-substituted phenylphosphites
and thereby the relatively high rates of hydroformy-
lation of internal olefins explain also the high i/n
ratios [10,11]. The improvements in NMR techniques
and molecular modeling have given new possibili-
ties to study the properties of phosphane and phos-
phite ligands [9,13–18]. Today, molecular modeling
techniques are more and more used to rationalize
structure/reactivity relationships, while ab initio cal-
culations give an insight into the electronic structure
of organometallic complexes [18].
Although regioselectivity in the hydroformylation
of alkenes depends on a number of factors, one of the
most important in the corresponding reaction system
are the effects induced by the catalyst. The prop-
erty of aromatic phosphane ligands in combination
with o-substituents has been studied mainly with het-
eroatom containing substituent groups [19–26]. These
potential bidentate ligands have proved to be useful in
modifying the electronic and steric properties of reac-
tive metal centers. However, the heteroatom modifi-
cation (–SCH3, –N(CH3)2, –OCH3 and –CF3) turned
out to be unsuccessful in propene hydroformylation
decreasing the conversion markedly, and in some cases
the hydroformylation reaction was totally hindered.
The preparation and utilization of non-polar o-alkyl-
substituted phosphanes has been studied only in a few
reports [4,21,27–29]. o-Methylphenyldiphenylphos-
phane ligand as a part of PtCl2(cod)/SnCl2 catalyst
in the hydroformylation of 2-butene has enhanced
selectivity to the branched aldehyde [28]. Moreover,
in the hydroformylation of propene, we achieved enh-
anced selectivity to isobutanal relative to triphenyl-
phosphane through modification of rhodium catalyst
with o-alkyl-substituted triphenyphosphane [4]. High
i/n ratios were obtained with (o-methylphenyl)diphen-
ylphosphane (o-MeP), (o-ethylphenyl)diphenylphos-
phane (o-EtP) and (2,4,5-trimethylphenyl)diphenylp-
hosphane (2,4,5-MeP), but unfortunately the activities
remained low [4].
The promising results containing o-methyl-substitu-
ted and o-ethyl-substituted phenylphosphanes for con-
trolling the product formation upon hydroformylation
of propene and 1-hexene achieved created a need for
further systematic investigations. Ligands with still
bulkier o-substituents have now been examined as a
modifiers for rhodium catalysts designed for hydro-
formylation. The present paper describes the prepa-
ration and characterization by NMR spectroscopy of
(2-cyclohexylphenyl)diphenylphosphane (1), (2-isop-
ropylphenyl)diphenylphosphane (2), (2-methylnapht-
hyl)diphenylphosphane (3), (2,5-dimethylphenyl)dip-
henylphosphane (4) and (2-phenylphenyl)diphenylph-
osphane (5) ligands (see Scheme 1). The ligands
were tested in the hydroformylation of propene and
1-hexene. In spite of the different experimental se-
tups for the two alkenes comparison was considered
reasonable because the effects of the different ligands
were examined relative to the reaction with PPh3
modified catalyst. Furthermore, the use of different
Rh-species were sought to be justified, as the forma-
tion of actual catalyst species, trigonal bipyramidal
rhodium carbonyl hydride can be formed from various
rhodium precursors.
The crystal structures of trans-Rh(CO)Cl(o-MeP)2
complex (o-MeP = o-methylphenyldiphenylphosp-
hane) showed earlier that the o-alkyl-substituents were
very close to the metal center [29]. The interesting
finding concerning the location of o-alkyl-substituents
inspired to additional studies and so the stereoelec-
tronic effect of the ligands on the hydroformylation
reaction was probed by experimental and theoretical
methods. Rh(acac)(CO)(PR3) complexes were pre-
pared and characterized in order to investigate the
coordination properties of the ligands. Geometrical

H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
71
Scheme 1. Schematic structures of the reported ligands (numbering corresponds to the NMR data). Cone angles (◦) are given in parentheses.
arrangement and steric size (cone angle) of the free
and coordinated ligands were studied theoretically
by ab initio Hartree–Fock and DFT methods. Data
obtained are discussed in relation to the results of the
hydroformylation of propene and 1-hexene.
(99%) was from Fluka. Diethyl ether and tetrahydro-
furan were distilled over sodium-benzophenone ketyl
under nitrogen before use.
Spectroscopy: The characterization of the ligands
1
was based on H-, 13C{1H}-, 31P{1H}-, COSY-90-,
HSQC- and COLOC-NMR spectroscopy. NMR spec-
tra were recorded on a Bruker DPX400 spectrometer
at room temperature in CDCl3 (99.8% D, 0.03%
TMS). NMR spectra of the metal complexes were
recorded on a Bruker AM250 spectrometer. 1H-NMR:
reference SiMe4. 13C{1H}-NMR: CDCl3 set to
77.0 ppm. 31P{1H}-NMR: external standard 85%
H3PO4. 13C{1H}- and 31P{1H}-NMR shifts of the
free ligands are presented in Tables 1 and 2, respec-
tively. Exact mass peaks for ligands were determined
on a Micromass LCT, using ESI+ method.
2. Experimental
General remarks: The used phosphanes are air- and
moisture-sensitive, both as pure compounds and in sol-
ution, and unprotected show observable oxidation
within 1 day. Thus, all reactions were carried out with
standard Schlenk techniques under nitrogen or argon
atmosphere. The ligands were recrystallized from
ethanol or a mixture of ethanol and hexane. 2-Bromo-
toluene (99%), 1-bromo-2-methylnaphthalene (90%),
bromobenzene (99%), 2-bromobiphenyl (96%), chlo-
rodiphenylphosphane (95%), n-butyllithium (2.5 M
solution in hexane) and (acetylacetonato)dicarbonylr-
hodium(I) (98%) were obtained from Aldrich. 1-Bro-
mo-2-isopropylbenzene (97%) and 1-bromo-2-cycloh-
exylbenzene (97%) were from Lancaster and p-xylene
2.1. Synthesis of ligands
2.1.1. (2-Cyclohexylphenyl)diphenylphosphane (1)
A solution of n-butyllithium (8.4 ml, 20.9 mmol)
was transferred dropwise via a canula to a freshly
prepared solution of 1-bromo-2-cyclohexylbenzene

72
H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
Table 1
1
13C{ H}-NMR spectra (100 MHz, CDCl3) of studied ligands 1–5 (see Scheme 1)
1
2
3
4
5
Co-4 (ppm)
Co-3 (ppm)
26.11
(s)
26.77
(s)
–
–
–
–
–
–
–
–
–
–
–
–
128.66
(s)
127.54
(s)
Co-2 (ppm)
4JCP (Hz)
Co-1 (ppm)
3JCP (Hz)
Cm-1 (ppm)
4JCP (Hz)
C5 (ppm)
34.10
(s)
41.82
24.3 (d)
–
23.87
(s)
31.24
24.7 (d)
–
–
–
–
–
129.69
3.0 (d)
141.68
6.4 (d)
–
24.38
22.1 (d)
–
20.69
20.8 (d)
21.05
(s)
129.96
(s)
–
–
–
–
125.86
(s)
125.97
(s)
128.59
(s)
127.15
(s)
C3 (ppm)
3JCP (Hz)
C9∗ (ppm)
3JCP (Hz)
C10∗ (ppm)
126.07
4.8 (d)
128.39
6.9 (d)
128.55
(s)
125.41
4.7 (d)
128.45
6.5 (d)
128.58
(s)
129.71
5.4 (d)
128.41
5.6 (d)
127.53
(s)
130.02
(s)
128.49
7.1 (d)
128.71
(s)
130.08
4.7 (d)
128.34
7.1 (d)
128.43
(s)
C4 (ppm)
4JCP (Hz)
C6 (ppm)
5JCP (Hz)
C8∗ (ppm)
2JCP (Hz)
C1 (ppm)
1JCP (Hz)
C7∗ (ppm)
1JCP (Hz)
C2 (ppm)
2JCP (Hz)
C7 (ppm)
4JCP (Hz)
C8 (ppm)
3JCP (Hz)
C9 (ppm)
2JCP (Hz)
C10 (ppm)
3JCP (Hz)
129.02
(s)
133.35
(s)
133.95
19.8 (d)
134.84
10.4 (d)
137.21
10.4 (d)
152.16
23.4 (d)
–
–
–
–
–
–
–
–
129.24
(s)
133.35
(s)
133.94
20.4 (d)
134.43
11.1 (d)
137.21
11.0 (d)
153.33
23.8 (d)
–
–
–
–
–
–
–
–
131.15
1.4 (d)
124.88
1.3 (d)
131.53
18.3 (d)
136.64
10.8 (d)
136.49
13.9 (d)
145.67
19.1 (d)
125.96
1.4 (d)
128.15
19.8 (d)
128.81
17.4 (d)
132.84
3.1 (d)
129.53
(s)
133.23
(s)
133.97
19.6 (d)
136.12
8.1 (d)
135.28
(s)
139.09
24.8 (d)
–
–
–
–
–
–
–
–
127.34
(s)
134.08
(s)
133.88
20.0 (d)
135.86
14.4 (d)
137.64
11.7 (d)
148.25
28.6 (d)
–
–
–
–
–
–
–
–
Table 2
1
31P{ H}-NMR and FT-IR spectroscopic data of free and coordinated phosphanes
Ligand (PR3)
Free PR3
Rh(acac)(CO)(PR3)
31P{ H}-NMR
31P{ H}-NMR
FT-IR
υ(CO) (cm−1
)
1
1
δP (ppm)
δP (ppm)
1JP–Rh (Hz)
1
2
3
4
5
−13.6
−13.8
−16.5
−9.5
–
44.8
41.4
–
–
175
171
–
1977
1974
1978
1976
1975
−11.9
–
–

H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
73
(5.0 g, 20.9 mmol) in diethyl ether (50 ml) at −10 →
0 ◦C (salted ice bath). The bright mixture was stirred
for 2 h at this temperature, after which a solution of
chlorodiphenylphosphane (4.6 g, 3.8 ml, 20.9 mmol)
in diethyl ether (30 ml) was added slowly. The ad-
dition colored the reaction mixture first light yellow
and finally white. Stirring was continued for a further
2 h at −10 → 0 ◦C. After slow warming to room tem-
perature, some solid material precipitated (inorganic
salts). Solid and liquid layers were separated by fil-
tration and solvent was removed in vacuo. The crude
yellow and oily product was obtained from the solvent
phase and recrystallized from ethanol. The yield of the
white pure product was 5.2 g, 15.0 mmol, 72%. Exact
mass (Micromass LCT, ESI+): 345.1746 (M + H)+
phane (4.4 g, 3.6 ml, 20.0 mmol) afforded a crude
brown oily product. The oily product was recrystal-
lized several times from ethanol. The yield of the
resulting white pure product was 1.6 g, 4.8 mmol,
24%. Exact mass (Micromass LCT, ESI+): 327.1293
(M + H)+ (calcd. for C23H20P, 327.1303). H-NMR
1
(400 MHz, CDCl3, see Scheme 1 for numbering) δH:
2.54 (s, 3H, Ho-1), 7.20–7.28 (m, 7H, H7, H9∗, H10∗),
7.33–7.40 (m, 6H, H3, H6, H8∗), 7.79 (d, JHH
=
3
8.0 Hz, 1H, H5), 7.86 (d, JHH = 8.4 Hz, 1H, H4),
3
4
8.34 (dd, 3JHH = 8.6 Hz, JHP = 4.0 Hz, 1H, H8).
2.1.4. (2,5-Dimethylphenyl)diphenylphosphane (4)
This compound was prepared similarly to (2-cyclo-
hexylphenyl)diphenyl phosphane but with chlorodiph-
enylphosphane replaced by dichloro(2,5-dimethylph-
enyl)phosphane. Dichloro(2,5-dimethylphenyl)phos-
phane was synthesized by Friedel–Crafts reaction
of phosphorus trichloride with p-xylene in the pres-
ence of aluminum chloride as catalyst [31]. Reac-
tions of bromobenzene (4.7 g, 3.2 ml, 30.0 mmol),
n-butyllithium (12.0 ml, 30.0 mmol) and dichloro(2,5-
dimethylphenyl)phosphane (3.1 g, 15.0 mmol) af-
forded a solid light yellow product. The solid product
was recrystallized from ethanol. The yield of the
resulting white pure product was 4.0 g, 13.8 mmol,
92%. Exact mass (Micromass LCT, ESI+): 291.1325
1
(calcd. for C24H26P, 345.1772). H-NMR (400 MHz,
CDCl3, see Scheme 1 for numbering) δH: 1.15–1.25
(m, 2 H, Ho-4), 1.35 (m, 4H, Ho-2), 1.60–1.75 (m,
4H, Ho-3), 3.29 (m, 1H, Ho-1), 6.83 (m, 1H, H6),
7.06 (m, 1H, H5), 7.26–7.34 (m, 12H, H3, H4, H8∗,
H9∗ and H10∗). 13C{1H}- and 31P{1H}-NMR shifts
are presented in Tables 1 and 2.
2.1.2. (2-Isopropylphenyl)diphenylphosphane (2)
With the procedure described above for (2-cyclohe-
xylphenyl)diphenylphosphane, reactions of 1-bromo-2-
isopropylbenzene (6.0 g, 30.0 mmol), n-butyllithium
(12.0 ml, 30.0 mmol) and chlorodiphenylphosphane
(6.6 g, 5.6 ml, 30.0 mmol) afforded a crude oily prod-
uct. A white solid product was obtained by wash-
ing the oily product with hexane/ethanol mixture.
The product was recrystallized from ethanol. The
yield of the white pure product was 2.1 g, 7.0 mmol,
23%. Exact mass (Micromass LCT, ESI+): 305.1465
1
(M + H)+ (calcd. for C20H20P, 291.1303). H-NMR
(400 MHz, CDCl3, see Scheme 1 for numbering)
δH: 2.16 (s, 3H, Hm-1), 2.34 (s, 3H, Ho-1), 6.57 (d,
3JHP = 4.8 Hz, 1H, H6), 7.06 (dd, JHH = 7.7 Hz,
3
5JHP = 1.3 Hz, 1H, H4), 7.10 (dd, JHH = 7.6 Hz,
3
4JHP = 4.6 Hz, 1H, H3), 7.28 (m, 4H, H8∗), 7.34 (m,
6H, H9∗ and H10∗). Preparation of 4 has been de-
scribed in the literature, however, by different method,
with lower yield and without interpretation of NMR
shifts [32].
1
(M + H)+ (calcd. for C21H22P, 305.1459). H-NMR
(400 MHz, CDCl3, see Scheme 1 for numbering)
3
δH: 1.13 (d, JHH = 6.8 Hz, 6H, Ho-2), 3.70 (sep,
3JHH = 6.8 Hz, 1H, Ho-1), 6.81 (m, 1H, H6), 7.07 (m,
1H, H5), 7.25–7.28 (m, 4H, H8∗), 7.31–7.35 (m, 8H,
H3, H4, H9∗ and H10∗). Although ligand 2 is a known
compound, its synthesis has not been described earlier
[30].
2.1.5. (2-Phenylphenyl)diphenylphosphane (5)
With the procedure described for (2-cyclohexylphe-
nyl)diphenylphosphane, reactions of 2-bromobiphenyl
(2.8 g, 2.1 ml, 12.0 mmol), n-butyllithium (4.8 ml,
12.0 mmol) and chlorodiphenylphosphane (2.6 g,
12.0 mmol) afforded a crude oily product. The white
solid product was obtained by washing the crude
product with hexane/ethanol mixture. The product
was recrystallized from ethanol. The yield of the
white pure product was 2.8 g, 8.3 mmol, 69%. Exact
2.1.3. (2-Methylnaphthyl)diphenylphosphane (3)
With the procedure described for (2-cyclohexylph-
enyl)diphenylphosphane, reactions of 1-bromo-2-me-
thylnaphthalene (4.4 g, 3.1 ml, 20.0 mmol), n-butylli-
thium (8.0 ml, 20.0 mmol) and chlorodiphenylphos-

74
H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
mass (Micromass LCT, ESI+): 339.1294 (M + H)+
2.4. Hydroformylation of propene and 1-hexene
1
(calcd. for C24H20P, 339.1303). H-NMR (400 MHz,
CDCl3, see Scheme 1 for numbering) δH: 7.06 (m,
Propene hydroformylation: Experiments were car-
ried out in a 250 ml autoclave (Berghof) equipped
with a sampling system and a 230 ml Teflon liner. The
experiments were done in semi-batch mode so that
synthesis gas pressure was kept constant during the
experiment. The rhodium precursor was Rh(NO3)3
(Fluka). In a typical experiment, the autoclave was
charged with the rhodium precursor (0.02 mmol calcu-
lated as rhodium), acetone (310 mmol, Merck, >99%),
internal standards decane (7 mmol, Fluka, >98%) and
hexane (12 mmol, Riedel de Häen, >99%) and the
phosphane. The ligand to rhodium ratio was 10:1 on
molar basis. The system was flushed with nitrogen and
pressurized with propene (2 bar, Aga, 99.8%), heated
to the reaction temperature (100 ◦C) with continuous
stirring, and then pressurized to the reaction pressure
(10 bar) with a 1:1 molar ratio of H2 and CO (MG,
99.997%). At least nine samples were taken for anal-
ysis in each experiment: one of the fresh reaction, one
immediately after pressurizing with H2 and CO, which
was considered as the starting point of the reaction,
six during the experiment and one after the reaction.
1-Hexene hydroformylation: Experiments were con-
ducted in a 100 ml autoclave (Berghof) with 60 ml
Teflon liner. The experiments were done in batch mode
with rhodium precursor Rh(acac)(CO)2. The reactor
was charged under a nitrogen purge with substrate,
rhodium catalyst, ligand (L/Rh 50) and internal stan-
dard cyclohexane. The autoclave was then sealed and
pressurized with a 1:1 mixture of H2 and CO (MG,
99.997%) to 25 bar and heated to 80 ◦C. After 4 h, the
autoclave was cooled and brought to normal atmo-
spheric pressure.
1H, H6), 7.16–40 (m, 18H, H3, H4, H5, H8∗, H9∗,
o-3
H10∗, Ho-2, H
and Ho-4). Preparation of 5 has
been described in the literature, but by a different
method, and with lower yield [33].
2.2. Rh(acac)(CO)(PR3) complexes
(Acetylacetonato)dicarbonylrhodium(I)
(50 mg,
0.19 mmol) and the phosphane ligand (0.19 mmol)
were dissolved in minimum amount of tetrahydrofuran
in separate flasks. The solutions containing equimolar
amounts of Rh(acac)(CO)2 and phosphane (PR3) were
combined and stirred at room temperature overnight.
After that the solution was evaporated to dryness
in vacuum. Spectroscopic data of the compounds
in Table 2 are characteristic of Rh(acac)(CO)(PR3)
species. However, attempts to crystallize the com-
pound for further studies were not successful.
2.3. Computational details
We used Gaussian 94 [34] and Sybyl 6.03 [35]
programs in modeling. No special geometry around
the phosphorus was used in ligand modeling. Ini-
tial guesses of the geometries of Rh(acac)(CO)(PR3)
species were based on the solid state structure of
Rh(acac)(CO)2 [36].
Ligand: Structures were optimized on the HF level
of theory using the 3-21G∗ basis set. Optimization
was carried out only to release the excess strain
and thus no conformational analysis was performed.
For the cone angle measurements, the metal(dummy
atom)–phosphorus distance 2.28 Å and van der Waals’
radius of hydrogen 1.2 Å were used.
Complexes: Hartree–Fock and two density func-
tional methods, namely B3LYP and B3PW91, were
used for the geometry optimization of complex struc-
tures. Calculations of Rh(acac)(CO)2, Rh(CO)Cl(SP)
(SP = (2-thiomethylphenyl)diphenylphosphane), and
Rh(CO)Cl(NP) (NP = (2-dimethylaminophenyl)dip-
henylphosphane) were carried out with three different
basis sets: 3-21G∗, 6-31G∗ and 6-31+G∗. Calcula-
tions of Rh(acac)(CO)(PR3) species were carried out
with the combination B3PW91/6-31G∗.
The procedure was the same for the two alkenes: a
disposable inner Teflon liner was used to avoid the ac-
cumulation of rhodium on the reactor walls. Further-
more, the purity of the system was checked with blank
runs before each experiment. The products were ana-
lyzed with a Hewlett-Packard 5890 GC equipped with
a capillary column (HP-1, 1.0 ␮m × 0.32 mm × 60 m)
and a flame-ionization detector. Products were quan-
tified by the internal standard method. In addition, the
aldehydes were identified by GC–MS analysis. Con-
version, selectivity and i/n ratio were calculated on
molar basis. Conversion was calculated with respect
to propene (1-hexene) and selectivity with respect to

H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
75
aldehydes. The i/n ratio of the aldehydes was defined
as the amount of branched product divided by the
amount of linear product. Hydrogenation was not de-
tected in any of the experiments.
repulsion, in the case of o-MeP and 3, the substituent
was placed inside the cone. The calculated cone angles
of the free ligands are given in Scheme 1. As substan-
tial effect of the size of the ligand on the cone angle
was observed by modifying the o-position. Compared
with PPh3 (149◦) the cone angle increased from 184◦
for 1 to 191◦ for 5. An even larger increase from o-MeP
(151◦) to 3 (223◦) occurred when one o-methyl sub-
stituted benzene ring was replaced with an o-methyl
substituted naphthyl ring.
3. Results and discussion
3.1. Synthesis of free ligands
This study is a continuation of our previous study
on ligands o-MeP, o-EtP and 2,4,5-MeP. The present
ligands, 1–5, likewise represent modifications of
the o-position of the aromatic, PPh3-like, structure,
are bulkier than those of our previous study [4,29].
The ligands were prepared by a modified literature
method from o-substituted brominated aromatics by
lithiation with n-butyllithium and further reaction
with arylchlorophosphanes [19]. The yields were rea-
sonably good (69–92%) except for 2 and 3, which
required several washing and recrystallization steps
and appeared in lower yield (23 and 24%).
3.4. Syntheses of complexes
Reaction between equimolar amounts of Rh(acac)-
(CO)2 and phosphanes (PR3) afforded rhodium com-
plexes with the general formula Rh(acac)(CO)(PR3).
Spectroscopic (31P{1H}-NMR, FT-IR) characteri-
zations of the Rh(I) complexes are summarized in
Table 2.
3.5. Ab initio calculations of complexes
Molecular modeling methods were also applied to
the structures of complexes. The stereoelectronic na-
ture of the coordinated ligands was expected to pro-
vide valuable information for the structure activity
search, especially because the crystallization of the
Rh(acac)(CO)(PR3) species failed.
3.2. Characterization of free ligands
The structural characterization of the ligands was
done by NMR methods. In 13C{1H}-NMR spectra,
the chemical shifts for phenyl carbons (Scheme 1,
marked 7∗, 8∗, 9∗, 10∗), o-substituted aromatic ring
carbons 1, 2 and 6, substituent carbons o-1 and m-1,
and naphthyl ring carbons 1 and 2 could be assigned
for all ligand structures (Table 1). The unconfirmed
1H-NMR shifts were assigned on the basis of the reli-
able information from two-dimensional NMR spectra:
(H, C)-correlated HSQC, (H, H)-correlated COSY-90
and long-range (C, H)-correlated COLOC. Even then,
however, the multiplicity could not be identified for
all of the aromatic spin systems and cyclohexyl sub-
stituent because of signal overlap. Two-dimensional
spectra were necessary in the assignment of 13C chem-
ical shifts of the carbons 3–5 (Scheme 1) and also the
carbons of the naphthyl ring.
The Rh(acac)(CO)2 derivatives were chosen as
model compounds for the search for the appropriate
method and level of calculation. As in the case of
free ligand optimization, Hartree–Fock method was
applied to the complexes. Two density functional
methods, B3LYP and B3PW91, were studied as well.
Results are shown in Table 3. As expected, the inade-
quacy of the HF method to handle electron correlation
was evident. In particular, HF failed to predict the
inter-atomic distances by overestimating the Rh–C
and Rh–O distances. Two density functional meth-
ods gave better agreement with the solid-state struc-
tures. The agreement got better moving from 3-21G∗
to 6-31G∗, but failed to improve with the use of
6-31+G∗. This was considered to be related to the rel-
ativistic effects found on second-row transition metals.
The evaluation of the two density functional meth-
ods was, therefore, continued with 6-31G∗. To take
into account the effect of the coordinated phosphane
ligand, two additional complexes, Rh(CO)Cl(SP)
3.3. Ab initio calculations of free ligands
The optimization procedure led to a geometry where
with just one exception, the o-substituents were lo-
cated outside the cone. Owing to the much lower steric

76
H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
Table 3
Standard deviation in bond distances and angles between the
solid-state structure and gas-phase structure
Basis set
HF
B3LYP
B3PW91
Rh(acac)(CO)2
3-21G∗
0.040/1.991
0.058/1.862
0.022/0.874
0.021/0.831
0.023/0.841
0.019/0.803
0.015/0.724
0.017/0.706
6-31G∗
6-31+G∗
RhCOCl(SP)a
6-31G∗
0.021/1.52
0.013/1.60
0.012/1.36
0.011/1.58
RhCOCl(NP)b
6-31G∗
a SP = (2-thiomethylphenyl)diphenylphosphane.
b NP = (2-dimethylaminophenyl)diphenylphosphane.
(SP = (2-thiomethylphenyl)diphenylphosphane) and
Rh(CO)Cl(NP) (NP = (2-dimethylaminophenyl)dip-
henylphosphane), were selected for study. Both DFT
methods gave reasonable accuracy, with B3PW91
method taking a slight edge over B3PW91 accompa-
nied with the 6-31G∗ basis set. Thus, B3PW91/6-31G∗
was the choice for the computational treatment for
Rh(acac)(CO)(PR3) species.
Because of the convergence failures with the opti-
mization procedure, only three complexes were suc-
cessfully optimized: Rh(acac)(CO)(3), Rh(acac)(CO)-
(5) and Rh(acac)(CO)(o-MeP) (Scheme 2). The
third was considered to be an adequate model for
Rh(acac)(CO)(4) and Rh(acac)(CO)(2) as well.
Cone angle measurements of the coordinated lig-
ands differed significantly from those of the free lig-
and (see Scheme 2). This is most clearly seen with
ligand 3. In free ligand state it is energetically favor-
able to bend the naphthyl ring away from the rest of
the structure, leaving the methyl group inside the cone.
In coordinated state the naphthyl ring rotates as far
as possible from the metal center, leaving the methyl
group directed towards the rhodium. Such behavior is
a good example of the situation where the free ligand
structure does not give even a qualitative picture of the
coordinated state and there is a huge difference in the
cone angle values for coordinate and free state (159
and 223◦, respectively). Correspondingly the cone an-
gle values for the free and bound ligand 5 were 191◦
versus 177◦.
Scheme 2. Gas-phase structures of free and coordinated ligands
and their calculated cone angles.
Further evidence of this was provided by X-ray crys-
tallography of trans-Rh(CO)Cl(o-MeP)2 complex
[29]. It is reasonable to believe that a similar situa-
tion is prevails for 2 and 4. Similarly the coordinated
5 must model the coordinated ligand 1 adequately
enough.
3.6. Hydroformylation of propene
Results from the hydroformylation of propene are
displayed in Table 4. The disadvantages of the un-
modified rhodium catalyst are that the initial activity
and conversion of the propenes hydroformylation re-
action are very low, even though the regioselectivity to
produce isobutanal was 57% under the conditions em-
ployed. The widely used ligand PPh3 was, therefore,
chosen as a good reference for the prepared ligands.
The methyl group is also directed towards the
metal center in the case of coordinated o-MeP ligand.

H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
77
Table 4
Results for hydroformylation of propene by rhodium catalysts with o-substituted aromatic phosphane ligands
Ligand
θ (◦)
δP (ppm)
X (%)
Initial rate (mol/(molRh s))
Sisobutanal (%)
i/n
1
2
3
4
184
187
223
159
191
151
159
149
–
−13.6
−13.8
−16.5
−9.5
−11.9
−10.7
−11.8
−3.3
–
67
63
46
70
45
81
50
97
26
24
17
23
11
35
20
45
–
46.0
47.0
45.0
43.0
48.5
44.0
45.0
36.0
57.0
0.85
0.89
0.82
0.75
0.94
0.79
0.82
0.56
1.3
5
o-MeP
2,4,5-MeP
PPh3
No ligand
10 (3 h)
Reaction conditions: 10 bar (CO/H2 = 1), T = 100 ◦C, X = conversion = 2 h, propene/Rh = 512, L/Rh = 10. Initial rate for aldehyde
formation. o-MeP = (2-methylphenyl)diphenylphosphane and 2,4,5-MeP = (2,4,5-trimethylphenyl)diphenylphosphane.
Our o-substituted ligands gave considerably higher i/n
ratios than triphenylphosphane. Relative to PPh3, the
selectivity to isobutanal increased from 43 to 48.5%
for all ligands studied, and the initial activity decreased
from 11 to 35 mol/(molRh s). Still, the n-butanal con-
tinued to be the main product.
ical shifts are attributable mostly from large range
of cone angles caused by steric crowding of the
o-substituted phenylphosphanes. The cone angles in-
crease with downfield 31P{1H}-NMR chemical shift.
Ligands with 31P{1H}-NMR shift between −9.5
and −11.8 ppm are o-methyl-substituted phenylphos-
phanes (15–159◦) and the bulky o-cyclohexyl- or
o-isopropyl-substituent containing phenylphosphanes
(184–187◦) have 31P{1H}-NMR shift around −13.6
to −13.8 ppm. Comparison of the structurally simi-
lar o-alkyl-substituted ligands (1–3 and o-MeP) with
PPh3 shows previously reported trend between the
31P{1H}-NMR shift and the selectivity to isobutanal
and the initial activity [4]. The selectivities increase
and the initial activities decrease as the 31P{1H}-NMR
shift of the free ligand decreases.
The decrease in the initial activity is more marked
than the increase in the selectivity. Relative to PPh3,
the most notable difference in the initial activity is
observed for the phenyl substituted ligand 5 and the
smallest for o-MeP. Even rigid naphthylphosphane 3
had less suppressing effect on the initial activity than
did 5. This may be a reflection of the smaller cone
angle of the coordinated ligand 3 than of the coordi-
nated 5. Both the ligands containing m- and p-methyl
substituents (4 and 2,4,5-MeP) and the ligands con-
taining bulkier o-alkyl-substituent (1 and 2) decreased
the initial activity more than did o-MeP, even though
the selectivity to form isobutanal was more or less the
same with the five ligands. On the average, o-MeP
would seem to represent an optimum between activity
and selectivity.
The regioselectivity of the hydroformylation reac-
tion is determined by the stereoelectronic factors of the
catalytically active complex. According to previous
studies, the more basic the phosphane the less active
it is and the higher the i/n ratio [12,37–39]. Carbonyl
stretching frequencies of metal complexes are used as
a measure of the basicity of phosphanes [6]. In our
work, however, the only narrow range of the carbonyl
IR stretching frequencies for Rh(acac)(CO)(PR3)
complexes (Table 2) did not reveal any clear correla-
tion with the propene hydroformylation results. Even
though the regioselectivities of the rhodium catalysts
with studied ligands were similar, there could be
seen slight increase in i/n ratios as the cone angle of
o-substituted phenylphosphane was between 184 and
191◦ compared to o-methyl-substituted phenylphos-
phanes (151–159◦). The cone angles seems to give
also a rough ordering with respect to conversion. The
bulkier the phosphane, the lower the conversion.
The electronegativity of substituents on phosphorus
and the angles between them are the two most impor-
tant variables determining 31P{1H}-NMR chemical
shift, and hence the shift characterize both the elec-
tronic nature and steric nature of the phosphorus
atom [5]. Evidently, the ligands are electronically
very similar and differences in 31P{1H}-NMR chem-

78
H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
Table 5
Results for hydroformylation of 1-hexene by rhodium catalysts with o-substituted aromatic phosphane ligands
Ligand
θ (◦)
δP (ppm)
X (%)
S2-hex (%)
S3-hex (%)
S2-mh (%)
S1-hep (%)
i/n
1
2
3
4a
5
PPh3
184
187
223
159
191
149
−13.6
−13.8
−16.5
−9.5
90
59
99
2
99
82
2
0
4
0
3
2
4
5
10
0
9
3
30
32
26
0
29
24
63
64
60
100
59
71
0.48
0.50
0.43
–
0.49
0.34
−11.9
−3.3
Reaction conditions: 25 bar (CO/H2 = 1), T = 80 ◦C, X = conversion = 4 h, 1-hexene 15.5 mmol, toluene 10 ml, Rh(acac)(CO)2
1.9×10−2 mmol, L/Rh = 50. S2-hex = total amount of 2-hexene; S3-hex = total amount of 3-hexene; S2-mh = selectivity to 2-methylhexanal;
S1-hep = selectivity to 1-heptanal.
a At 20 bar and 100 ◦C, and 1-hexene:Rh = 10,000: X = 98%, S2-hex = 61%, S3-hex = 18%, S2-ep = 2%, S2-mh = 8%, S1-hep = 11%
and i/n = 0.91. S2-ep = selectivity to 2-ethylpentanal.
3.7. Hydroformylation of 1-hexene
capable of competing in binding with (against) CO,
and the formed HRh(CO)2(PPh3)2 species gives rise
to lower i/n ratio. In contrast to this, the sterically
hindered o-modified phosphanes favors the formation
of HRh(CO)3(PR3) species. As a result, CO dissoci-
ation is more likely to occur, and the formed 16e−
complexes with the large space available are favor-
able for producing branched alkylrhodium species
and enhanced i/n ratios [8–10,12]. However, sterically
demanding o-alkyl-substituted phenylphosphanes
have lower ␲-acceptor ability than o-alkyl-substituted
phenylphosphites even then they both have large cone
angle [8]. Due to that the catalytically active rhodium
center become less electrophilic and probably hence,
the reactivity will decrease.
Results from the 1-hexene hydroformylation tests
are compiled in Table 5. All the catalyst systems af-
forded relatively high chemoselectivity to aldehydes,
with 1-heptanal and 2-methylhexanal the major prod-
ucts. Only small amounts of 2- and 3-hexene were
formed, and thus 2-ethylpentanal was not formed at all.
The low conversion with the o–m-methyl-substituted
ligand 4 (2%) is in accordance with our earlier re-
sults for 2,4,5-MeP, which likewise did not form
an active catalyst at 80 ◦C, but restored the activity
when the temperature was raised to 100 ◦C (Table 5)
[29].
The modified ligands gave considerably higher i/n
ratios than triphenylphosphane. The effect was the
same as with propene but less pronounced. Likewise,
there were similarities between the two alkenes in the
branched aldehyde formation. However, the effect of
the modified ligands on branching was less in the case
of 1-hexene than propene. Relative to PPh3, the selec-
tivity of the modified catalyst to 2-methylhexanal in-
creased from 2 to 8%. The catalyst systems containing
1, 2 and 5 and gave the highest i/n ratios.
Despite of the different experimental setups, hy-
droformylation of the two alkenes with the ligand-
modified rhodium catalysts showed similar trends,
which were also in agreement with findings in our
earlier studies with o-MeP, o-EtP and 2,4,5-MeP
ligands [4]. Relative to triphenylphosphane, steric
bulk due to the o-modification affected the selectiv-
ity to the branched aldehyde formation positively.
Evidently, without any substituents, PPh3 (149◦) is
4. Conclusion
We have shown that, in the rhodium-catalyzed hy-
droformylation, a wider group of o-alkyl-substituted
aromatic phosphane ligands enhance the selectivity
to branched aldehydes. The effect of the modified
ligands on branching is less pronounced in the case
of 1-hexene compared to propene. The steric stress of
o-alkyl-substituent has significant influence both on
the formation of active rhodium species and catalytic
reaction. The regioselectivity to isobutanal increase
and the initial activity decrease as the cone angle
increase with downfield 31P{1H}-NMR chemical
shift of closely related free ligand. Most probably
the formation of HRh(CO)3(PR3) species are due to
the increased i/n ratios. However, the simultaneous

H. Riihimäki et al. / Journal of Molecular Catalysis A: Chemical 200 (2003) 69–79
79
optimization of reactivity and selectivity constitutes a
major challenge for the future.
[16] J.M. Smith, B.C. Taverner, N.J. Coville, J. Organomet. Chem.
530 (1997) 131.
[17] D. White, N.J. Colville, Adv. Organomet. Chem. 36 (1994)
99.
[18] J. Bruckmann, C. Krüger, F. Lutz, Z. Naturforsch. 50b (1995)
351.
Acknowledgements
[19] G. Dyer, D.W. Meek, J. Am. Chem. Soc. 89 (1967) 3983.
[20] E.G. Hope, T. Kemmitt, W. Levason, J. Chem. Soc., Perkin
Trans. II (1987) 487.
[21] L. Horner, G. Simons, Phosphorus Sulfur 14 (1983) 189.
[22] L. Horner, G. Simons, Phosphorus Sulfur 15 (1983) 165.
[23] R.H. Laitinen, V. Heikkinen, M. Haukka, A.M.P. Koskinen,
J. Pursiainen, J. Organomet. Chem. 598 (2000) 235.
[24] P. Suomalainen, S. Jääskeläinen, M. Haukka, R.H. Laitinen, J.
Pursiainen, T.A. Pakkanen, Eur. J. Inorg. Chem. (12) (2000)
2607.
We gratefully acknowledge the financial support
from Neste Oxo AB and the National Technology
Agency of Finland (Tekes). We wish to thank Ms.
Päivi Joensuu and Ms. Sari Ek for the mass spectra,
Ms. Merja Harteva for performing part of the propene
hydroformylation tests and Ms. Liisa Saharinen for
assistance in the 1-hexene hydroformylation tests and
preparation of rhodium complexes.
[25] H.K. Reinius, R.H. Laitinen, A.O.I. Krause, J.T. Pursiainen,
Stud. Surf. Sci. Catal. 130 (2000) 551.
[26] P. Suomalainen, H.K. Reinius, H. Riihimäki, R.H. Laitinen,
S. Jääskeläinen, M. Haukka, J.T. Pursiainen, T.A. Pakkanen,
A.O.I. Krause, J. Mol. Catal. A: Chem. 169 (2001) 67.
[27] C.B. Ziegler, R.F. Heck, J. Org. Chem. 43 (1978) 2941.
[28] F. Ancilotti, M. Lamiand, M. Marchionna, J. Mol. Catal. 63
(1990) 15.
References
[1] B. Breit, W. Seiche, Synthesis 1 (2001) 1.
[2] A.M. Trzeciak, J.J. Ziüłkowski, Coord. Chem. Rev. 190–192
(1999) 883.
[29] P. Suomalainen, H. Riihimäki, S. Jääskeläinen, M. Haukka,
J.T. Pursiainen, T.A. Pakkanen, Catal. Lett. 77 (2001) 125.
[30] H.J. Jakobsen, T. Liptaj, T. Bundgaard, S. Sørensen, J. Magn.
Reson. 26 (1977) 71.
[31] B. Buchner, L.B. Lockhart, J. Am. Chem. Soc. 73 (1951) 755.
[32] H. Schindlbauer, Monatsh. Chem. 96 (1965) 2051.
[33] F. Mitterhofer, H. Schindlbauer, Monatsh. Chem. 98 (1967)
206.
[34] M.J. Frish, G.A. Petersson, J.A. Montgomery, K.
Raghavachari, M.A. Al-Lahman, V.G. Zakrzewski, J.V. Ortiz,
J.B. Foresman, J. Ciolowski, B.B. Stefanov, A. Nanayakkara,
M. Challacombe, C.Y. Peng, P.Y. Ayala, W. Chen, M.W.
Wong, J.L. Andres, E.S. Replogle, R. Gomperts, R.L. Martin,
D.J. Fox, J.S. Binkley, D.J. Defrees, J. Baker, J.P. Stewart,
M. Head-Gordon, C. Gonzales, J.A. Pople, Gaussian Inc.,
Pittsburgh, PA, 1995.
[3] M. Beller, B. Cornils, C.D. Frohning, C.W. Kohlpaintner, J.
Mol. Catal. A: Chem. 104 (1995) 17.
[4] H.K. Reinius, P. Suomalainen, H. Riihimäki, E. Karvinen, J.
Pursiainen, A.O.I. Krause, J. Catal. 199 (2001) 302.
[5] C.A. Tolman, Chem. Rev. 77 (1977) 313.
[6] C.A. Tolman, J. Am. Chem. Soc. 92 (1970) 2953.
[7] C.A. Tolman, J. Am. Chem. Soc. 92 (1970) 2956.
[8] P.W.N.M. van Leeuwen, C.F. Roobeek, J. Organomet. Chem.
258 (1983) 343.
[9] T. Jongsma, G. Challa, P.W.N.M. van Leeuwen, J. Organomet.
Chem. 421 (1991) 121.
[10] S.N. Poelsma, P.M. Maitlis, J. Organomet. Chem. 451 (1993)
C15.
[11] A. van Rooy, E.N. Orij, P.C.J. Kamer, P.W.N.M. van Leeuwen,
Organometallics 14 (1995) 34.
[12] B. Cornils, W.A. Herrmann, Applied Homogeneous Catalysis
with Organometallic Compounds, vol. 1, VCH, Weinheim,
1996, Chapter 2, pp. 58–59.
[35] Sybyl 6.03, Tripos Associates, 1699 S. Hanley Road, Suite
303, St. Louis, MO 63144, USA.
[36] F. Huq, A.C. Skapski, J. Cryst. Mol. Struct. 4 (1974) 411.
[37] C.A. Streuli, Anal. Chem. 32 (1960) 985.
[38] C.A. Streuli, Anal. Chem. 31 (1959) 1652.
[39] W.R. Moser, C.J. Papile, D.A. Brannon, R.A. Duwell, S.J.
Weininger, J. Mol. Catal. 41 (1987) 271.
[13] C.J. Elsevier, B. Kowall, H. Kragten, Inorg. Chem. 34 (1995)
4836.
[14] S. Song, E.C. Alyea, Can. J. Chem. 74 (1996) 2304.
[15] T.L. Brown, K.J. Lee, Coord. Chem. Rev. 128 (1993)
89.