Phosphabenzenes as monodentate π-acceptor ligands for rhodium-catalyzed hydroformylation

Article abstract of DOI:10.1002/1521-3765(20010716)7:14<3106::AID-CHEM3106>3.0.CO;2-Y

A new class of phosphinine/rhodium catalysts for the hydroformylation of terminal and internal alkenes is presented in this study. A series of phosphabenzenes 1-14 has been prepared by condensation of phosphane or tris(trimethylsilyl)phosphane with the corresponding pyrylium salt. Trans[(phosphabenzene)₂RhCl(CO)] complexes 21-25 have been prepared and studied spectroscopically and by X-ray crystal-structure analysis, The hydroformylation of oct-1-ene has been used to identify optimal catalyst preformation and reaction conditions. Hydroformylation studies with 15 monophosphabenzenes have been performed. The catalytic performance is dominated by steric influences, with the phosphabenzene 8/rhodium system being the most active catalyst. Turnover frequencies of up to 45370 h^-1 for the hydroformylation of oct-1-ene have been determined. In further studies, hydroformylation activity toward more highly substituted alkenes was investigated and compared with the standard industrial triphenylphosphane/rhodium catalyst. The reactivity differences between the phosphabenzene and the triphenylphosphane catalyst increase on going to the more highly substituted alkenes. Even tetra substituted alkenes reacted with the phosphabenzene catalyst, whereas the triphenylphosphane system failed to give any product. In situ pressure NMR experiments have been performed to identify the resting state of the catalyst. A monophosphabenzene complex [(phosphinine 8)Ir(CO)₃H] could be detected as the predominant catalyst resting state.

Full text of DOI:10.1002/1521-3765(20010716)7:14<3106::AID-CHEM3106>3.0.CO;2-Y

FULL PAPER
Phosphabenzenes as Monodentate p-Acceptor Ligands for
Rhodium-Catalyzed Hydroformylation
Bernhard Breit,*[a] Roland Winde,[a] Thomas Mackewitz,*[b]
Rocco Paciello,[b] and Klaus Harms[c]
Dedicated to Professor Manfred Regitz on the occasion of his 65th birthday
Abstract: A new class of phosphinine/
rhodium catalysts for the hydroformyla-
tion of terminal and internal alkenes is
presented in this study. A series of
phosphabenzenes 1 ± 14 has been pre-
pared by condensation of phosphane or
tris(trimethylsilyl)phosphane with the
corresponding pyrylium salt. Trans-
zenes have been performed. The cata-
lytic performance is dominated by steric
influences, with the phosphabenzene 8/
rhodium system being the most active
catalyst. Turnover frequencies of up to
phosphane/rhodium catalyst. The reac-
tivity differences between the phospha-
benzene and the triphenylphosphane
catalyst increase on going to the more
highly substituted alkenes. Even tetra-
substituted alkenes reacted with the
phosphabenzene catalyst, whereas the
triphenylphosphane system failed to
give any product. In situ pressure
NMR experiments have been performed
to identify the resting state of the
catalyst. A monophosphabenzene com-
plex [(phosphinine 8)Ir(CO)3H] could
be detected as the predominant catalyst
resting state.
1
45370 h for the hydroformylation of
oct-1-ene have been determined. In
further studies, hydroformylation activ-
ity toward more highly substituted al-
kenes was investigated and compared
with the standard industrial triphenyl-
[(phosphabenzene)2RhCl(CO)]
com-
plexes 21 ± 25 have been prepared and
studied spectroscopically and by X-ray
crystal-structure analysis. The hydrofor-
mylation of oct-1-ene has been used to
identify optimal catalyst preformation
and reaction conditions. Hydroformyla-
tion studies with 15 monophosphaben-
Keywords: homogeneous catalysis ´
hydroformylation
´
phosphaben-
zenes ´ phosphanes ´ rhodium
formylation catalysts,[3] carefully designed ligand modifica-
tions have led to improved catalysts.[2, 4] For example, Caseyꢀs
concept of the natural bite angle stimulated the development
of chelating diphosphine- and diphosphite-modified catalysts
that allow the regioselective hydroformylation of terminal
olefins in favor of the usually desired linear isomer.[5]
Advances in the field of stereochemical control have been
made by employing enforced substrate direction with the help
of substrate-bound catalyst directing groups.[6] This concept
has been applied to cyclic[7] and acyclic[8] olefinic building
blocks. Additionally, chiral ligands such as BINAPHOS and
others have been developed which allow for high asymmetric
induction during the hydroformylation of certain prochiral
alkenic substrates.[2, 9]
Introduction
Hydroformylation of olefins is one of the most important
homogeneously catalyzed industrial processes.[1] In the course
of this reaction a new carbon carbon bond is formed and,
simultaneously, the synthetically valuable aldehyde function
is introduced. If the chemo-, regio-, and stereoselectivity of
such an appealing transformation could be controlled, new
perspectives for both fine-chemical production and organic
synthesis would arise. Major advances have been made
toward this goal during the last decades.[2] Since the discovery
of rhodium(i) phosphine complexes as chemoselective hydro-
[a] Prof. Dr. B. Breit, Dr. R. Winde
Organisch-Chemisches Institut
However, one of the long-standing and fundamental
problems of hydroformylation chemistryÐthe low pressure
hydroformylation of internal and more highly substituted
olefinsÐstill awaits a practical solution. Thus, it is known that
the reactivity of hydroformylation catalysts decreases expo-
nentially with the degree of olefin substitution.[10] Although
rhodium phosphine catalysts display a significantly higher
reactivity toward the hydroformylation of terminal olefins
compared with the cobalt phosphine systems, their activity
Ruprecht-Karls-Universität Heidelberg
Im Neuenheimer Feld 270, 69120 Heidelberg (Germany)
Fax : (‡49)6221-54-4205
E-mail: bernhard.breit@urz.uni-heidelberg.de
[b] Dr. T. Mackewitz, Dr. R. Paciello
BASF AG, Ammoniaklaboratorium, ZAG
67056 Ludwigshafen (Germany)
[c] Dr. K. Harms
Fachbereich Chemie der Philipps-Universität Marburg
Hans-Meerweinstrasse, 35043 Marburg (Germany)
3106
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3106 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

3106±3121
toward hydroformylation of internal and higher 1,1-disubsti-
tuted olefins is still rather low.[10] In more detailed studies on
the effect of modifying ligands with respect to the activity of
the rhodium catalyst, some trends have been unraveled that
may give a hint to a solution of this problem.[4a] Thus, strong
and small s-donor ligands may retard or even inhibit the
reaction, while switching to stronger, p-acceptor ligands, such
as phosphites, results in more active hydroformylation
catalysts. In particular, the discovery of bulky phosphites as
modifying ligands for rhodium catalysts has been an impor-
tant advance in this field.[11] High catalyst activity for both
hydroformylation of internal and terminal olefins has been
observed. In situ NMR and IR spectroscopy has revealed that
the enormous catalyst activity may be attributed to a mono-
phosphite rhodium complex as the catalytically active spe-
cies.[12] Recently, similar phosphonite systems have been
prepared with comparable ligand properties.[13] Unfortunate-
ly, a technical application of phosphites has been hampered by
their inherent lability toward hydrolysis and a tendency to
undergo degradation reactions.[14] Hence, the development of
new classes of p-acceptor ligands remains an important task.
In this context, we became interested in a hitherto unexplored
class of potential p-acceptor ligandsÐthe phosphaben-
zenes.[15] Although the first preparation of these systems by
Märkl dates back to 1966,[15] no application as modifying
ligands in homogenous catalysis was described until 1996.[16]
Of particular interest is the electronic situation in phospha-
benzenes compared with their nitrogen counterpartÐthe
pyridine nucleus. Photoelectron spectroscopy[17] and ab initio
calculations[18] show that the HOMO of a phosphabenzene
has p-character with a large coefficient at the phosphorus
atom. Note that the HOMO of pyridine is the lone pair at
nitrogen. The LUMO of a phosphabenzene also has p-
character with a large coefficient at the phosphorus atom. Its
energy is lower than the corresponding pyridine LUMO.
Hence, phosphabenzenes possess, at least qualitatively, an
ideal frontier molecular-orbital situation for an efficient
overlap with filled metal d-orbitals and, hence, the ability to
act as p-acceptor ligands.
Here we wish to report in detail the preparation of
phosphabenzenes and their coordination properties toward
rhodium(i) centers, as well as their use as modifying ligands for
the rhodium-catalyzed hydroformylation of terminal and
internal olefins.[19]
Scheme 1. Phosphabenzenes synthesized and investigated as ligands in
rhodium-catalyzed hydroformylation.
Results
Scheme 2. i) 2 ± 2.5equiv PTMS3, CH3CN, reflux, 5 h (15 ± 51%); ii) PH3,
30 bar, n-butanol, 1108C, 4 h (28 ± 84%).
Preparation of phosphabenzenes: Phosphabenzene ligands
1 ± 14 (Scheme 1) have been prepared employing a modified
version of Märklꢀs original procedure.[20] Thus pyrylium
tetrafluoroborates were treated in acetonitrile with excess
tris(trimethylsilyl)phosphine at reflux temperature to give the
phosphabenzenes 1 ± 14 in moderate to fair yields, after
chromatographic workup. Alternatively, pyrylium salts may
be treated directly with phosphane gas (30 bar) at 1108C to
give the corresponding phosphabenzenes, usually in better
yields (Scheme 2).[21]
The required pyrylium salts were obtained by tetrafluoro-
boric acid-mediated condensation of an appropriate chalcone
derivative with a corresponding acetophenone system.[22] The
chalcones were either prepared separately through aldol
condensation of an arylmethylketone derivative with an
aromatic aldehyde under basic conditions (catalytic amounts
of activated barium hydroxide)[23] or generated in situ from
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3107 $ 17.50+.50/0
3107

B. Breit, T. Mackewitz et al.
FULL PAPER
the same components in the presence of tetrafluoroboric
acid[22] (see Experimental Section). 2,6-Dialkyl-substituted
pyrylium salts were prepared according to published methods
(Scheme 3).[24]
ing phosphabenzene yielded the desired mononuclear trans-
[(h1-phosphinine)2RhCl(CO)] complexes 22 ± 25 in quantita-
tive yields (Scheme 6, Table 1).
Scheme 3. i) 10 mol% Ba(OH)2, ethanol, 20 ± 808C; ii) HBF4 (52%
ethereal solution), 1,2-dichloroethane, reflux, 4 h (22 ± 62%).
Pyrylium salt 16 was obtained in a one-pot operation upon
treatment of chalcone 17 with tetrafluoroboric acid. Hence,
the formation of 16 may be attributed to an interesting
domino reaction sequence consisting of an intramolecular
aromatic substitution to give 18 followed by condensation
with enone 17 to furnish the polycyclic pyrylium tetrafluoro-
borate 16 (Scheme 4).
Scheme 6. Preparation of trans-[(phosphinine)2RhCl(CO)] complexes
21 ± 25. i) [Rh(CO)2Cl]2, CH2Cl2, RT, 30 min, (quant.). 21: R1 ˆ Me, R2 ˆ
Ph, 22: R1 ˆ Ph, R2 ˆ Ph, 23: R1 ˆ 2-Naphth, R2 ˆ Ph, 24: R1 ˆ Ph, R2 ˆ
C6H4(p-OMe), 25: R1 ˆ Ph, R2 ˆ C6H4(p-CF3).
Table 1. trans-[(Phosphinine)2RhCl(CO)] complexes 21 ± 25 prepared ac-
cording to Scheme 6.
1
phosphinine
R1
R2
complex
nÄ(CO) [cm
]
1
3
4
9
10
Me
Ph
2-Naphth
Ph
Ph
Ph
Ph
Ph
21
22
23
24
25
1998[26]
1999
1998
2002
2003
C6H4(p-OMe)
C6H4(p-CF3)
The h1-coordination, as well as the symmetric nature of the
complexes 22 ± 25, is reflected in their 31P NMR spectra. Thus,
the phosphorus atoms absorb at d ˆ 166.1 ± 173.8 as doublets
1
with a J(P,Rh) coupling constant near the expected value of
175 Hz.[19b, 27] Compared with the free ligand, the 31P NMR
signal is shifted about 20 ppm to higher field, which is typical
for a h1-coordinated phosphinine.[19b] The presence of the
carbonyl ligand can be seen by 13C NMR spectroscopy. Thus,
in a typical chemical-shift range of 180.7 ± 181.9,[28] the
carbonyl carbon resonates as a doublet of triplets with typical
coupling constants of 65 Hz [1J(C,Rh)] and 22 Hz [2J(P,C)].[29]
Additionally, the triplet splitting indicates the presence of two
magnetically equivalent phosphabenzene nuclei. Final struc-
tural proof was obtained by an X-ray crystal-structure analysis
of complex 22. Its structure in the solid state is depicted in
Figure 1.
Scheme 4. i) HBF4 (52% ethereal solution), 1,2-dichloroethane, reflux, 4 h
(28%).
The tert-butyl substituted phosphabenzene 15 was obtained
on treatment of a-pyrone 19 with phosphaalkyne 20 in a
cycloaddition-reversion sequence as introduced by Regitz
et al. (Scheme 5).[25]
Scheme 5. i) Benzene, 1408C, 3 d, Schlenk pressure vessel (54%).
Rhodium complexes: In order to study the ligand properties
of phosphabenzenes in comparison with other known ligand
classes, trans-[L2RhCl(CO)] complexes were prepared. Thus,
treating [Rh(CO)2Cl]2 with four equivalents of a correspond-
Figure 1. Ortep plot of the structure of trans-[(3)2RhCl(CO)] (22) in the
solid state.
3108
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3108 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
Table 2. Selected structural data for trans-[(3)2RhCl(CO)] (22).
The P Rh lengths of 2.27 and 2.28 Š (Table 2) are note-
worthy. They compare well with the analogous trispyrrol
phosphine complex (d(Rh P) ˆ 2.28 Š)[30] as well as with the
analogous bulky phosphite complex (d(Rh P) ˆ 2.29 Š, see
Table 3),[31] but are significantly shorter than that found for
the corresponding triphenylphosphine complex of 2.32 pm.[32]
Since trispyrrolphosphine is known to be a strong p-acceptor
ligand,[30] the similarity of the Rh P bond lengths may reflect
the back-bonding capability of the phosphabenzene nucleus.
The least square planes of the phosphabenzene rings are tilted
toward the Rh-P-CO-P-Cl least square plane by about 55.18
and 57.18. Likewise, the ortho-phenyl substituents at the
phosphabenzene nucleus are twisted out of the phosphaben-
zene plane by 40.2 ± 53.28.
Bond Lengths [Š]
Rh C(100)
Rh P(1')
Rh P(1)
Rh Cl(1)
C(100) O(100)
P(1) C(2)
P(1) C(6)
C(2) C(3)
C(3) C(4)
1.824(5)
2.2724(12)
2.2805(12)
2.3465(14)
1.104(7)
1.716(4)
1.723(5)
1.381(6)
1.414(6)
C(4) C(5)
C(5) C(6)
P(1') C(6')
P(1') C(2')
C(2') C(3')
C(3') C(4')
C(4') C(5')
C(5') C(6')
1.397(7)
1.397(6)
1.715(5)
1.722(5)
1.390(6)
1.392(7)
1.396(7)
1.394(6)
Bond Angles [8]
C(100)-Rh-P(1')
C(100)-Rh-P(1)
P(1')-Rh-P(1)
C(100)-Rh-Cl(1)
P(1')-Rh-Cl(1)
P(1)-Rh-Cl(1)
O(100)-C(100)-Rh
C(2)-P(1)-C(6
C(2)-P(1)-Rh
C(6)-P(1)-Rh
C(3)-C(2)-P(1
C(7)-C(2)-P(1)
C(2)-C(3)-C(4)
89.27(16)
91.28(16)
179.03(4)
178.65(18)
90.08(5)
89.39(5)
179.7(5)
105.8(2)
129.50(16)
124.71(17)
120.7(3)
119.8(3)
126.1(4)
C(5)-C(4)-C(3)
C(6)-C(5)-C(4)
C(5)-C(6)-P(1)
C(19)-C(6)-P(1)
C(6')-P(1')-C(2')
C(6')-P(1')-Rh
C(2')-P(1')-Rh
C(3')-C(2')-P(1')
C(7')-C(2')-P(1')
C(2')-C(3')-C(4')
C(3')-C(4')-C(5')
C(6')-C(5')-C(4')
C(5')-C(6')-P(1')
121.1(4)
125.0(4)
121.2(4)
117.6(3)
105.8(2)
124.39(16)
129.78(16)
120.2(3)
120.1(3)
125.9(4)
122.0(4)
124.4(4)
121.5(4)
The maximum angle that is formed from the van der Waals
radii of L and rhodium is 172.58. Note that this is not the
classical cone angle as defined by Tolman,[33] since above and
below the phosphabenzene plane the steric demand is much
smaller. However, semiempirical calculations (PM3) allowed
the rotation barrier about the P Rh bond to be to estimated
1 [34]
as less than 4 kcalmol . Thus, fast rotation of the phos-
phabenzene nuclei at the reaction temperatures employed is
possible and, as a consequence, the steric demand of a
phosphabenzene may be significantly larger than a fixed
model would predict.
Table 3. Comparison of the IR stretching frequency of CO and Rh P bond
lengths for selected trans-[L2RhCl(CO)] complexes.
Ligand
P(pyrrol)3 P[O(2-t-BuC6H4]3 phosphinine 3 PPh3
The IR stretching frequency of the CO ligand in trans-
[L2RhCl(CO)] is an ideal probe for determining the electronic
properties of a phosphabenzene ligand. In analogy to the
[LNi(CO)3] complexes of Tolman,[33] a higher CO stretching
frequency for a trans-[L2RhCl(CO)] complex corresponds to a
diminished electron density at the rhodium center and, hence,
to a stronger back-bonding of the ligand, L.[35] Since, the
rhodium complexes provide easy-to-interpret IR spectra with
a single carbonyl band, these systems have found wide
application as probes for the electronic properties of donor
ligands.[35] The extensive literature available[36, 37] allows a
rapid judgment of the electronic properties of a new ligand.
Hence, for the phosphabenzene complex 22 the CO band was
detected at 1999 cm 1 which is about 21 cm 1 higher in energy
than that found for the corresponding triphenylphosphine
complex,[30] but 14 cm 1 lower in energy than that found in the
complex with an tri(ortho-tert-butylphenyl)phosphite li-
gand[31] (see Table 1). This suggests that the electronic
properties of phosphabenzenes are located in a range between
phosphines and phosphites, with somewhat more similarity to
the phosphite ligands.
1
[a]
nÄ [cm
]
2024[30]
2013[31]
1999[b]
2.2724(12)[b]
1978[30]
2.323(6)[32]
d(Rh P) [Š] 2.282(4)[30] 2.2856(7)[31]
[a] in CH2Cl2; [b] this work.
trifluoromethyl (25) and a methoxy substituent (24) caused a
shift in the same direction. Hence, the methoxy group must
exert its electronic influence, which is of course small,
inductively through the s-system. This would be in accordance
with the observed out-of-plane geometry of all aryl substitu-
ents (see X-ray plot for complex 22 in Figure 1).
In situ high pressure NMR experiments: For the rhodium
triphenylphosphine system, it is known that under hydro-
formylation conditions, two phosphines are coordinated and
an equilibrium between an equatorial ± equatorial and an
axial ± equatorial species in a ratio of 85:15 exists.[38] For the
bulky phosphite rhodium catalyst, in situ high pressure NMR
and IR experiments indicate that a monophosphite rhodium
complex is the prevalent species.[12] High pressure 31P and
1H NMR experiments were carried out in order to get
information on the phosphabenzene rhodium system.
In an initial experiment, [Rh(COD)2OTf] was dissolved in
[D2]dichloromethane, treated with two equivalents of the
phosphabenzene ligand 8 and pressurized with 40 bars of
syngas (CO/H2 1:1) for 3 h at 908C. After cooling to room
temperature, the 31P NMR spectrum displayed a broad signal
at d ˆ 154.5, which split at 908C to a doublet with a typical
1J(P,Rh) coupling constant of 149 Hz. This result clearly shows
that phosphabenzenes coordinate to rhodium under syngas
pressure conditions. However, proton NMR spectroscopy
revealed that the rhodium species belonging to this 31P NMR
resonance is not the trigonal-bipyramidal hydride complex
IR spectra of the complexes were recorded in order to
probe the electronic influence of substituents at the C2-, 4-,
and 6-positions of the phosphabenzene nuclei on the stretch-
ing frequency of the CO ligand of the corresponding trans-
[L2RhCl(CO)] complexes (see Table 1). In all cases, only
marginal changes in the CO IR band were observed. Thus,
going from an ortho/ortho' dialkyl to an ortho/ortho' diaryl-
substituted system did not have an effect. In addition, the
exchange of phenyl substituents to more extended p-systems,
such as the 2-naphthyl substituents, did not effect the CO
stretching frequency. However, in the case of para substitution
of the 4-aryl substituents, a shift to slightly larger wave
1
numbers (1999 ! 2003 cm ) was found. Interestingly, both a
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3109 $ 17.50+.50/0
3109

B. Breit, T. Mackewitz et al.
FULL PAPER
typically seen in hydroformylation catalysis (no hydride signal
could be detected).
In order to characterize this hydride complex and to clarify
the number of ligands attached to the metal center in such a
species, we switched to the more stable iridium complexes.
Hence, [Ir(CO)2acac] (acac ˆ acetylacetonate) and two equiv-
alents of phosphabenzene 8 were dissolved in [D2]dichloro-
methane in a high pressure NMR tube and treated at 908C
with 40 bar syngas pressure. NMR spectra were recorded at
908C. In addition to the free phosphabenzene ligand (ca.
50%), the phosphorus NMR spectra show two phosphorus
signals at d ˆ 143.9 and 144.3. The proton NMR spectrum
shows a metal hydride signal at d ˆ 11.83 as a doublet with a
2J(H,P) coupling of 23.2 Hz. This doublet collapses to a singlet
in the 31P-decoupled 1H NMR experiment. Interestingly,
1H{31P} COSY spectra showed one cross peak for both 31P
signals at d ˆ 143 with the metal hydride signal at d ˆ 11.83.
These data are consistent with the presence of an h1-
monophosphabenzene hydrido-iridium complex. The 31P
NMR spectrum suggests that this species presumably exists
as two resolved rotamers at 908C. The splitting of the
hydride signal to a doublet, as well as the presence and
quantity of free phosphabenzene ligand, is consistent with the
coordination of only one phosphabenzene to the metal center
under syngas pressure. The small size of the 2J(H,P) coupling
constant suggests an equatorial position for the phosphaben-
zene nucleus.[39] Hence, iridium complex 26 is the predom-
inant species and, thus, may also represent the resting state of
the corresponding rhodium catalyst (Figure 2).
Scheme 7. Possible reaction pathways and products in the course of a
hydroformylation reaction with oct-1-ene or oct-2-ene.
preformation conditions for the rhodium/phosphabenzene
catalyst system. 2,4,6-Triphenylphosphabenzene (3) was cho-
sen as the standard ligand. Pretreatment of a mixture of
[Rh(CO)2acac] and 20 equiv. of phosphabenzene 3 in toluene
for 30 min at 758C with 10 bar CO/H2, followed by addition of
oct-1-ene through a pressure vessel gave the kinetic data
depicted in Figure 3b.
When compared with the control experiment in the absence
of phosphabenzene ligand (see Figure 3a), it becomes clear
that deactivation of the rhodium had occurred. This may be
attributed to the formation of unreactive rhodium clusters; a
phenomenon which has been observed previously.[40] In the
second experiment, see Figure 3c, phosphabenzene 3 was
allowed to react with [Rh(CO)2acac] in toluene for 45 min at
RT to ensure coordination to the rhodium center. After
addition of oct-1-ene and transfer to an autoclave, the
reaction mixture was heated to 758C and subsequently
pressurized with 10 bar H2/CO (1:1). The kinetic data
obtained clearly show that this procedure provided an active
hydroformylation catalyst. Interestingly, a strong tendency
toward isomerization was observed. One might speculate that
not all of the rhodium precursor was transferred into the
active rhodium phosphabenzene catalyst at the beginning of
the kinetic experiment. However, it was assumed that at the
end of a complete hydroformylation experiment all of the
rhodium(i) precursor should exist as an active hydroformyla-
tion catalyst. Hence, we let the catalyst do a complete
hydroformylation run of oct-1-ene for 15 h. After this time,
GC analysis showed complete consumption of the starting
material. The product mixture consisted exclusively of C9
aldehydes. More importantly, 31P NMR analysis of the
mixture showed that no phosphabenzene degradation had
occurred during that first hydroformylation run; this indicated
the stability of phosphabenzene ligands under hydroformyla-
tion conditions.[41] Next, a new hydroformylation run could be
started cleanly by addition of a second portion of oct-1-ene.
GC analysis provided the kinetic data depicted in Figure 3d,
which show clean hydroformylation catalysis. Hence, this
preformation procedure became our standard and was sub-
sequently applied to all further hydroformylation experiments
with oct-1-ene.
Figure 2. Proposed (Rh) and experimentally observed (Ir) mono-phos-
phabenzene tris-carbonyl d9-metal complexes 26 under 40 bar syngas
pressure.
The coordination behavior of phosphabenzenes mirrors
that of the bulky phosphite ligands.[12] Hence, one might
expect a related catalytic activity for the phosphabenzene
class of ligands.
Hydroformylation experiments: oct-1-ene: In order to eval-
uate the phosphabenzene rhodium system as a catalyst for the
hydroformylation reaction, oct-1-ene was chosen as a first test
substrate. The possible reaction pathways and reaction
products under hydroformylation conditions are depicted in
Scheme 7. Thus, in a single hydroformylation experiment, one
can learn about the regioselectivity of the hydroformylation
of the terminal alkene and the potential for isomerizing
terminal to internal alkenes, as well as the ability to hydro-
formylate internal alkenes.
At the beginning of a kinetic experiment, one has to ensure
that all of the rhodium precursor has been transferred into the
active hydroformylation catalyst; otherwise the observed
kinetic data would reflect the activation phase for the catalyst.
Hence, we began with an investigation to identify the ideal
For all subsequently reported hydroformylation experi-
ments, a complete set of kinetic data was recorded. In all
3110
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3110 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
experiments, complete consumption of octenes was reached
after a reaction period of about 4 h.
As shown in Figure 3d, the hydroformylation reaction with
oct-1-ene is zeroth order under our conditions, at least up to a
conversion of 30%. That is, turn over frequencies in this range
are independent of conversion. This allows comparison of the
catalytic activity by determining the turnover frequencies
(TOF) within the low conversion range (see Tables 4 ± 8,
below). Importantly, at low conversions, the internal olefins
formed by isomerization of the terminal olefin have not yet
been attacked. Hence, determination of turnover frequencies
for olefin isomerization is possible as well.
Influence of the ligand/rhodium ratio: The phosphabenzene 3/
rhodium ratio necessary to ensure that all the rhodium centers
are converted in situ into the active complexes and that all the
effects observed could be ascribed to the varied parameters
was determined. Ligand/rhodium ratios of 2:1, 20:1 and 91:1
were explored. At low phosphabenzene concentrations (Ta-
ble 4, entry 1), low hydroformylation but significant isomer-
Table 4. Influence of ligand/rhodium ratio on the hydroformylation of oct-
1-ene with [Rh(CO)2acac]/3, at 908C, 10 bar (CO/H2, 1:1) in toluene (c0 ˆ
2.27m) (oct-1-ene/Rh, 4166:1).[a]
L/Rh
Conv.
[%]
TOFald
TOFiso
TOFald
TOFiso
/
Isomax
[%]
n/isomax
1
1
[h
]
[h
]
1
2
3
2
20
91
19.4
21.6
14.6
955
2801
2442
721
610
54
1.3
4.6
45.2
26.5
9.8
2.7
2.7
2.0
1.9
[a] TOFald ˆ turnover frequency of aldehyde formation; TOFiso ˆ turnover
frequency of alkene isomerization.
ization activity was observed. This indicates that not all
rhodium species have been transferred into catalytically active
phosphabenzene/rhodium complexes. Similar effects are known
for bulky phosphite/rhodium systems.[42] Increasing the ligand
concentration (20 equiv, Table 4, entry 2) provided a catalyst
system with high hydroformylation activity and a lower tendency
for isomerization. The ratio of TOFald/TOFiso increased to 4.6:1;
this resulted in a maximum value of 9.8% for internal alkene
formation. Increasing the ligand concentration further (91 equiv,
Table 4, entry 3) slightly decreased the hydroformylation rate,
but almost suppressed isomerization activity. These results
indicate that a ligand/Rh ratio of 20:1 is sufficient to transfer
all of the rhodium into catalytically active species. Even
though at higher L/Rh ratios the isomerization rate could be
decreased further, the lower ligand loading (20:1) was chosen
as the standard for all further experiments.
Figure 3. Kinetics of hydroformylation experiments of oct-1-ene a) with
[Rh(CO)2acac] and b ± d) with [Rh(CO)2acac]/3 after different catalyst
~
&
^
preformation procedures; ˆ oct-1-ene, ˆ internal octene isomers,
ˆ
n-nonanal,  ˆ 2-methyloctanal, * ˆ 2-ethylheptanal. Initial oct-1-ene
concentration (c0 ˆ 2.27m) in toluene; for a) oct-1-ene/Rh ˆ 4166:1, for
b ± c) oct-1-ene/phosphinine 3/Rh ˆ 4166:20:1. a) Control experiment with-
out phosphabenzene ligand; [Rh(CO)2acac] at 758C, 10 bar (H2/CO 1:1),
toluene (c0 ˆ 2.27m). b) [Rh(CO)2acac]/3 in toluene were allowed to react
for 30 min at 758C, 10 bar (H2/CO 1:1) followed by addition of oct-1-ene
through a pressure chamber. c) [Rh(CO)2acac]/3 in toluene were allowed
to react for 30 min at RT followed by addition of oct-1-ene. The mixture
was transferred to a preheated autoclave (758C) and the reaction started by
addition of 10 bar H2/CO (1:1). d) Second run experiment: 908C, 10 bar H2/
CO (1:1). For details see Experimental Section.
Influence of reaction temperature: An increase of the
reaction temperature from 708C to 908C increased the rates
of both hydroformylation and isomerization. However, on
going further to 1108C, a dramatic drop for the rate of
hydroformylation was observed. Conversely, the isomeriza-
tion rate largely increased. Hence, 908C became the reaction
temperature of choice, since a good compromise between
rapid hydroformylation and modest isomerization could be
achieved (Table 5).
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3111 $ 17.50+.50/0
3111

B. Breit, T. Mackewitz et al.
FULL PAPER
Table 5. Influence of temperature on the hydroformylation of oct-1-ene
with [Rh(CO)2acac]/3, at 10 bar (CO/H2 1:1) in toluene (c0 ˆ 2.27m) (oct-1-
ene/3/Rh, 4166:20:1).
creased from 1.3% for ligand 1 up to a maximum of 20.3%
for 8. However, ligand 15 with a tert-butyl/phenyl substitution
pattern in ortho/ortho'-position did not provide an active
hydroformylation catalyst.
In order to probe whether an electronic modification of the
phosphabenzene nucleus would have any effect on the
catalytic properties of the resulting rhodium/phosphabenzene
catalyst, ligands 9 ± 12 were tested.
T
[8C]
Conv.
[%]
TOFald
TOFiso
TOFald
TOFiso
/
Isomax
[%]
n/isomax
1
1
[h
]
[h
]
1
2
3
70
90
110
19.6
21.6
27.1
1557
2801
1075
48
610
3316
32.4
4.6
0.3
3.9
9.8
53.0
2.2
2.0
3.3
Experiments with ligands 3, 10 and 11 (Table 8, entries 1 ±
3) show only a marginal influence of the CF3 or MeO group in
the para-position to the 4-aryl substituent at the phospha-
Influence of reaction pressure: To study the influence of the
syngas pressure, experiments at 5, 10, 20 and 40 bars were
performed (see Table 6). A strong increase of hydroformyla-
tion and isomerization rate was observed. While the hydro-
formylation rate showed an almost fivefold increase, the
Table 8. Influence of ligand structure (electronics) on the hydroformyla-
tion of oct-1-ene with [Rh(CO)2acac]/L, at 908C, 10 bar (CO/H2 1:1) in
toluene (c0 ˆ 2.27m) (oct-1-ene/L/Rh 4166:20:1).
Ligand
Conv.
[%]
TOFald
TOFiso
TOFald
TOFiso
/
Isomax
[%]
n/isomax
Table 6. Influence of syngas pressure (CO/H2, 1:1) on the hydroformyla-
tion of oct-1-ene with [Rh(CO)2acac]/3, at 908C in toluene (c0 ˆ 2.27m)
(oct-1-ene/3/Rh, 4166:20:1).
1
1
[h
]
[h
]
1
2
3
4
5
6
7
3
9
21.6
30.6
21.8
27.1
32.8
27.8
26.4
2801
2443
2655
1951
4583
2181
1995
779
110
863
200
801
180
175
3.6
22.2
3.1
9.8
5.7
9.8
3.2
12.6
4.6
10.7
5.5
5.5
2.0
1.8
2.0
2.0
1.9
1.9
1.9
p
[bar]
Conv.
(%)
TOFald
TOFiso
TOFald
TOFiso
/
Isomax (%)
n/isomax
1
1
10
11
12
13
14
[h
]
[h
]
1
2
3
4
5
10
20
40
41.1
36.5
31.6
40.1
1111
2437
4371
5141
550
610
883
2.0
4.0
4.9
3.3
13.2
9.8
23.5
14.1
1.9
2.0
2.4
2.1
12.1
11.4
1550
isomerization rate increased only by a factor of three on going
from five to 40 bar total pressure (CO/H2 1:1). This behavior
differs from that of the bulky phosphite/rhodium catalysts,
which show only a marginal rate increase at higher syngas
pressures.[42] The lowest isomerization was found at 10 bar
with a maximum amount of 9.8% for the formation of internal
alkenes. Hence, 10 bar syngas pressure was selected as the
optimal reaction pressure for all further studies.
benzene nucleus. This is consistent with the small change in
the CO stretching frequency found for the rhodium com-
plexes 22, 24, and 25, which indicates only minor electronic
effects on the rhodium center. Introducing two CF3 groups in
the 3- and 5-positions of an ortho-aryl substituent (ligand 11,
Table 8, entry 4) lowered the catalytic activity. However,
adding one further CF3 group at the 4-position of the para aryl
group (ligand 12, Table 8, entry 5) leads to a more active
hydroformylation catalyst.
In a further attempt, we explored whether the introduction
of a weakly coordinating donor into the phosphabenzene
system might influence the isomerization propensity. One
may assume that such an additional hemilabile coordinative
function might suppress the formation of coordinatively
unsaturated rhodium intermediates, and thus reduce olefin
isomerization. Hence, methoxy substituted ligands 13 and 14
were tested (Table 8, entries 6 and 7). Compared with the
standard triphenylphosphabenzene ligand 3 system, the
hydroformylation activity for both the 2- and 3-methoxy
derivatives is only slightly diminished. However, isomeriza-
tion is significantly reduced. In both cases a maximum of
5.5% internal olefin formation was detected.
The above experiments identified the rhodium catalyst
obtained from phosphabenzene 8 as the most active for
hydroformylation of oct-1-ene. Figure 4 shows the concen-
tration of starting material and products as a function of time
under a) standard conditions (908C, 10 bar syngas) and
b) enforced reaction conditions (1308C, 40 bar syngas). Inter-
estingly, for the standard conditions, a conversion of oct-1-ene
of 78% was reached after 30 min. In addition to aldehydes a
maximum of 20.3% internal octenes had formed at this point
of time. After 30 minutes, a decrease in the concentration of
internal olefins was accompanied by an increase in 2-ethyl-
Influence of ligand structure: In order to determine how steric
effects at the phosphabenzene ligands influence the hydro-
formylation reaction, a series of monophosphabenzene li-
gands was tested under standardized conditions (see Table 7).
The results depicted in the table show that increasing the
steric demand of the ortho/ortho'-substituents at the phos-
phabenzene leads to more active catalysts. Going from
dimethyl derivative 1 (Table 7, entry 1) to ligand 8 (Table 7,
entry 8) led to a fivefold increase of the hydroformylation
rate. Likewise isomerization toward internal alkenes in-
Table 7. Influence of ligand structure (sterics) on the hydroformylation of
oct-1-ene with [Rh(CO)2acac]/L, at 908C, 10 bar (CO/H2 1:1) in toluene
(c0 ˆ 2.27m) (oct-1-ene/L/Rh 4166:20:1).
Ligand
Conv.
[%]
TOFald
TOFiso
TOFald
TOFiso
/
Isomax
[%]
n/isomax
1
1
[h
]
[h
]
1
2
3
4
5
6
7
8
1
2
3
4
5
7
6
8
28.2
24.6
21.6
35.0
31.4
32.5
33.5
11.5
1111
1711
2801
2866
1975
2522
3612
5751
18
61.7
3.8
3.6
23.7
21.1
10.4
1.8
1.3
18.0
9.8
3.0
3.7
5.2
15.3
20.3
1.8
2.4
2.0
1.9
2.0
1.5
2.2
3.3
451
779
121
94
242
1964
2071
2.8
3112
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3112 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
Scheme 8. Hydroformylation of cyclohexene. i) Rh/L, CO/H2, 40 bar, Tˆ
908C, toluene. TOF [h 1] L ˆ PPh3: 109, L ˆ 8: 1959.
Figure 4. Kinetics of the hydroformylation of oct-1-ene with [Rh(CO)2-
acac]/8, in toluene (c0 ˆ 2.27m) (oct-1-ene/8/Rh, 4166:20:1) at a) 908C,
&
^
10 bar (CO/H2 1:1) and b) 1308C, 40 bar (CO/H2 1:1); ˆ oct-1-ene,
ˆ
~
internal octene isomers, ˆ n-nonanal,  ˆ 2-methyl octanal, * ˆ 2-ethyl-
heptanal.
heptanal and 2-propylhexanal concentration. Hence, internal
olefins were attacked and hydroformylated; this indicated
that phosphabenzene/rhodium catalysts should be suited to
hydroformylate internal and more highly substituted olefins.
Increasing the reaction temperature to 1308C and the
syngas pressure to 40 bar provided an extremely active
hydroformylation catalyst (Figure 4b). Here, a maximum
Figure 5. a) Kinetics of the hydroformylation of cyclohexene with phos-
phinine 8/[Rh(CO)2acac] at 908C, 40 bar in toluene (c0 ˆ 2.63m) for 4 h
~
&
(cyclohexene/L/Rh, 4166:20:1), c(cyclohexene), c(cyclohexane carbal-
dehyde); b) Logarithmic plot of cyclohexene concentration against time
(khydroform. ˆ 7.3  10 3 s 1).
1
turnover frequency of 45370 h was observed for aldehyde
dium system was about 18 times slower with an initial TOF of
formation. Complete consumption of starting material was
reached after only 30 min. Additionally, complete consump-
tion of the internal octenes formed by prior alkene isomer-
ization was observed after 60 min.
Since phosphabenzene 8 evidently provided the most active
rhodium catalyst for the hydroformylation of oct-1-ene, it was
decided to employ this system in all subsequent studies on
hydroformylation of higher-substituted alkenes.
1
109 h .
Surprisingly, the TOF for hydroformylation of cyclohexene
with the phosphine/rhodium catalyst is only moderately lower
than the TOF upon hydroformylation of oct-1-ene. This is
rather different behavior from the bulky monodentate
phosphites. Their catalytic activity drops by about a factor
of 80 on going from oct-1-ene to cyclohexene as the substrate.
Interestingly, the reaction is first order with respect to
substrate (Figure 5b); this indicates that olefin coordination
has become the rate determining step.
Hydroformylation of more highly substituted olefins
Cyclohexene: Cyclohexene was selected in order to determine
the potential of phosphabenzene rhodium catalysts for hydro-
formylation of internal olefins, since it allows observation of
hydroformylation formally undisturbed by olefin isomeriza-
tion.
Oct-2-ene: A more complex situation is given for oct-2-ene as
an internal olefin substrate. In addition to hydroformylation
of the internal olefin, it is also of interest to see whether olefin
isomerization occurs simultaneously. An n-selective isomer-
izing hydroformylation might give rise to the formation of
linear aldehydes. Formation of oct-1-ene and subsequent
hydroformylation, for example, could then provide n-nonanal.
Such an isomerizing hydroformylation could be of industrial
Hydroformylation was performed at 908C and 40 bar
syngas pressure. For the phosphabenzene rhodium system,
1
an initial turnover frequency of 1959 h was detected (see
Scheme 8, Figure 5). The standard triphenylphosphane/rho-
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3113 $ 17.50+.50/0
3113

B. Breit, T. Mackewitz et al.
FULL PAPER
Table 9. Results of the hydroformylation of oct-2-ene (cis/trans, 77:23) with [Rh(CO)2acac]/8 and TPP at 908C, 10 bar (CO/H2 1:1) in toluene (c0 ˆ 7.13m)
after 4 h (oct-2-ene/L/Rh, 7222:20:1).
Ligand
oct-2-ene
[mol%]
oct-3/4-ene
[mol%]
nonanal
[mol%]
2-methyloctanal
[mol%]
2-ethylheptanal
[mol%]
2-propylhexanal
[mol%]
8
0
31
1
5
24
3
43
42
18
17
14
2
PPh3
significance with respect to ªraffinate 2º, a product of the
steam cracking process, which is a mixture of butenes.[4b]
The complete set of potential reaction pathways for an
isomerizing hydroformylation of oct-2-ene is depicted in
Scheme 7. The experiment was carried out at 908C and 10 bar
syngas pressure. After 4 h, the phosphabenzene catalyst had
transferred all the starting material into aldehyde products
(see Table 9). Interestingly, 24% of n-nonanal was formed.
the standard rhodium/triphenylphosphine catalyst as well as
with the phosphabenzene/rhodium system proceeded smooth-
ly to furnish the corresponding lactol as a mixture of anomers
in greater than 99% selectivity. Notably, the phosphabenzene/
rhodium system was about 2.5 times faster.
Tri- and tetrasubstituted alkenes: Trisubstituted alkenes such
as a-pinene are among the least reactive alkenes with respect
to the hydroformylation reaction. While the triphenylphos-
phine/rhodium system failed to provide any aldehyde product
under the reaction conditions employed, the rhodium/phos-
phabenzene catalyst transformed a-pinene smoothly into the
corresponding aldehyde (Scheme 10). Aldehyde selectivity
was greater than 99% and combined with good regio- as well
as diastereoselectivity.
1
The initial TOF for aldehyde formation was 6933 h . If one
assumes an intrinsic linearity for the rhodium/8 catalyst of
72%, as observed in the course of hydroformylation of oct-1-
ene (see above), one can conclude that 33% of the oct-2-ene
was isomerized to oct-1-ene (or its corresponding rhodium
complex) prior to hydroformylation. For the triphenylphos-
phine catalyst, 36% of octenes are left unconsumed after 4 h.
Additionally, only 5% of nonanal could be detected. Thus, the
phosphabenzene/rhodium catalyst is substantially more active
toward hydroformylation of internal olefins and, additionally,
shows a significant tendency for producing terminal aldehydes
from internal olefins.
1,1-Disubstituted alkenes: Isobutene was selected as a repre-
sentative example for the class of 1,1-disubstituted alkenes
(Scheme 9). Identical hydroformylation experiments with the
Scheme 10. Hydroformylation of tri- and tetrasubstituted alkenes. i) Rh/8,
CO/H2, 60 bar, Tˆ 808C, toluene. TOF ˆ 56 h 1. ii) Rh/L, CO/H2, 60 bar,
Tˆ 1008C, toluene. TOF [h 1] L ˆ PPh3: <1, L ˆ 8: 118.
The hydroformylation of tetrasubstituted alkenes, such as
tetramethylethylene, is more difficult.[10] Hydroformylation
may be achieved under very high pressures and temperatures
with an unmodified cobalt catalyst.[43] However, to the best of
our knowledge, no efficient hydroformylations in the low to
medium pressure range with ligand-modified rhodium cata-
lysts are known. Interestingly, the rhodium/phosphabenzene 8
catalyst is even able to consume tetramethylethylene as a
substrate for hydroformylation under the rather mild reaction
conditions of 1008C and 60 bar syngas pressure (Scheme 10).
The exclusive reaction product was 3,4-dimethylpentanal,
which arises through successive alkene isomerization and
regioselective hydroformylation. A remarkable turnover
Scheme 9. Hydroformylation of 1,1-disubstituted alkenes. i) Rh/L, CO/H2,
30 bar, Tˆ 808C, toluene. TOF [h 1] L ˆ PPh3: 332, L ˆ 8: 3132. ii) Rh/L,
CO/H2, 20 bar, Tˆ 908C, toluene. TOF [h 1] L ˆ PPh3: 1317, L ˆ 8: 3291.
rhodium/phosphabenzene 8 and the standard rhodium/tri-
phenylphosphine catalyst were performed. The reaction
mixture was analyzed after a reaction period of 4 h. Once
again, the phosphabenzene/rhodium catalyst was about 100
times faster than the standard rhodium/triphenylphosphine
system. The aldehyde selectivity and regioselectivity were in
both cases greater than 99%.
1
frequency of 118 h was observed.
Discussion
Methallyl alcohol was chosen to study a functionalized 1,1-
disubstituted alkene (Scheme 9). Hydroformylation with both
The above results clearly show that phosphabenzenes are an
emerging class of phosphorus-based ligands with a great
3114
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3114 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
potential for use in homogenous catalysis. They can easily be
prepared by condensation of pyrylium salts with PH3 or a
synthetic equivalent thereof. Although both h1 and h6
coordination modes have been observed in transition metal
complexes of phosphinines,[15] the preferred binding mode of
the phosphabenzenes investigated in this study is the h1
coordination toward a rhodium(i) center. The electronic
ligand properties of phosphabenzenes could be estimated
from the CO-stretching frequency of the corresponding trans-
[(phosphinine)2RhCl(CO)] complexes. Electronically, the
phosphinines resemble phosphite ligands; however, the p-
acceptor property is somewhat reduced. Nevertheless, phos-
phabenzene rhodium(i) complexes have been identified as
extremely active and robust hydroformylation catalysts.
Hydroformylation experiments have shown that for this class
of catalysts, a careful catalyst-preformation procedure is
necessary in order to transfer all the rhodium centers into
catalytically active species. This may be due to the reduced
donor capability of the phosphinine, which may allow the
formation of nonactive rhodium-carbonyl clusters under
syngas pressure. Extensive hydroformylation experiments
under several reaction conditions and with different phosphi-
nine ligands have revealed some interesting trends. Thus, the
optimal ligand to rhodium ratio seems to be in the range of
20:1. This ratio seems efficient for transferring all rhodium
centers into catalytically active phosphinine complexes. Addi-
tionally, this ligand to rhodium ratio allows a high rate for
hydroformylation to be combined with a moderate isomer-
ization activity.
studies have been carried out in order to understand why
the rate determining alkene coordination/insertion is accel-
erated for the phosphinine/rhodium systems. Although ex-
periments with rhodium complexes did not allow the detec-
tion of the expected trigonal bipyramidal hydrido complex 26
(Figure 2), such a species could be identified in the case of the
more stable iridium(i) system. NMR experiments indicate that
only one phosphinine ligand is coordinated to the transition
metal center. This situation is similar to that of the bulky
phosphite rhodium catalysts.[39] Hence, as in the case of the
bulky phosphite ligands, the reasons for the high catalyst
activity may be attributed to the formation of a monoligand
rhodium species. Such a species should have a larger
accessible space compared with the [HRh(CO)2(PPh3)2]
system. Hence, the energy barrier for coordination and
hydrometallation of a sterically more demanding alkene
should be reduced; this is reflected well in the experimental
results. Furthermore, in the case of oct-1-ene hydroformyla-
tion, the larger accessible space at the rhodium center may
allow a more facile reaction to branched species; this may
explain the rather modest linearities observed experimentally.
Additionally, one may speculate that in a monophosphinine
rhodium species, the dissociation of a CO ligand should be
easier for electronic reasons. This would induce low coordi-
nation numbers and could explain the high tendency of
phosphinine rhodium catalysts toward alkene isomerization.
Conclusion
Studies on the influence of sterics and electronics in the
phosphinine ligand systems have shown a clear trend. Steric
effects dominate the catalytic properties of the phosphinine
rhodium systems: the larger the ortho/ortho' substituents of
the phosphinine nucleus, the higher the hydroformylation
activity of the corresponding rhodium catalyst. Concomitant
with an increase of the hydroformylation rate is an increase in
the isomerization rate, with ligand 8 showing maximum
This study has introduced phosphabenzene/rhodium catalysts
as powerful catalysts for the hydroformylation of terminal
and, in particular, internal olefins. The catalyst activity is
based on the formation of a monoligand metal species, which
accounts for the higher rates of hydroformylation and
isomerization. Further studies will have to focus on the design
of chelating phosphinine ligands to provide active and
regioselective hydroformylation catalysts. Experiments to
evaluate the full potential of phosphinines as ligands in
homogenous late transition metal catalysis are underway.
1
activity. Thus, turnover frequencies of up to 45370 h for the
hydroformylation of oct-1-ene could be reached.
The hydroformylation of more highly substituted alkene
substrates with the most active phosphinine 8/rhodium
catalyst were studied in further investigations. The results
have been compared with the standard industrial rhodium/
triphenylphosphane catalyst. Thus, for all classes of alkenes,
including 1,2-disubstituted, 1,1-disubstituted, and trisubstitut-
ed alkenes, the phosphinine catalyst is significantly more
active. As expected, the reactivity difference between the two
catalyst systems becomes larger with increasing alkene
substitution. The most striking example is the hydroformyla-
tion of a tetrasubstituted alkene, tetramethylethylene. Where-
as the triphenylphosphane catalyst did not give any detectable
product at all, the phosphinine catalyst consumed this
Experimental Section
General: Reactions were performed in flame-dried glassware either under
argon (purity >99.998%) or under nitrogen. The solvents were dried by
standard procedures, distilled, and stored under nitrogen. All temperatures
quoted are uncorrected. 1H, 13C, and 31P NMR spectra were taken on a
BrukerARX-200, BrukerAC-300, BrukerDRX-400, and BrukerAMX-
500 with TMS, chloroform, or benzene as internal standards. For 31P NMR
spectra, 85% H3PO4 was used as the external standard. IR spectroscopy:
FTIR-510 (Nicolet); mass spectrometry: MATCH7 (Varian), ZAB2F
(Vakuum Generators), and JMS700 (JOEL). Melting points were meas-
ured on apparatus by Dr. Tottoli (Büchi). Elemental analyses: CHN rapid
analyzer (Heraeus), Vario EL (Analysensysteme GmbH). Flash chroma-
tography: Silica gel Si60, Merck, Darmstadt, 40 ± 63 mm. Analytical gas
chromatography was performed at ªSichromateº (Siemens) and HP5880A
with integrator (Hewlett Packard) employing a column of type OV1,
25 m  0.25 mm (Hewlett Packard). Hydroformylation reactions were
performed in a 100 mL stainless-steel autoclave equipped with a sampling
device and gas transfer stirrer (Premex, Switzerland). The stirrer engine
was a ªEurodigi-viskº (Ika). A thermostat, type CC301 (Huber), filled with
1
substrate with a turnover frequency of 118 h .
For reactions of more highly substituted alkenes, the rate
determining step has been shown to be early in the catalytic
cycle and may be either the alkene coordination or inser-
tion.[42] This is also true for the phosphinine 8/rhodium
catalyst, as reflected in the clean, first-order kinetics upon
hydroformylation of cyclohexene. In situ pressure NMR
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3115 $ 17.50+.50/0
3115

B. Breit, T. Mackewitz et al.
FULL PAPER
synthetic oil served as the heating device. In all hydroformylation
experiments, the autoclave was connected to a gas reservoir that kept the
CO/H2 gas pressure constant throughout the reaction. Gases: Premixed
carbon monoxide 1.8 (98%), hydrogen 3.0 (99.9%) in a ratio of 1:1
(Messer-Griesheim).
was recrystallized from hot ether (50 mL) to give the pyrylium salt as a
beige solid. Yield: 10.1 g, 26%; m.p. 1788C; H NMR (300 MHz, CDCl3):
1
d ˆ 1.47 (d, 3J(H,H) ˆ 7.0 Hz, 12H; CH3), 3.56 (sept, 3J(H,H) ˆ 7.0 Hz, 2H;
CH), 7.56 (t, 3J(H,H) ˆ 7 Hz, 2H; ArH), 7.66 (tt, 3J(H,H) ˆ 7 Hz,
4J(H,H) ˆ 2 Hz, 1H; ArH), 8.11 (dd, 3J(H,H) ˆ 7 Hz, 4J(H,H) ˆ 2 Hz,
2H; ArH), 8.11 (s, 2H; C4-H); 13C NMR (CDCl3, 75.47 MHz): d ˆ 20.1
(4C), 34.5 (2C), 116.5 (C4, 2C), 129.8 (2C), 130.1 (2C), 132.2, 135.2, 167.8
(C5), 184.8 (C3, 2C); elemental analysis calcd (%) for C17H21BF4O (328.2):
C 62.22, H 6.45; found C 62.25, H 6.65.
The following compounds were prepared by literature procedures:
chalkone,[44] 2,4,6-triphenylpyrylium tetrafluoroborate,[22] tris(trimethyl-
silyl)phosphine,[45] and 2,2-dimethylpropylidine phosphane,[45]
Benzylidene-2-acetonaphton: Activated barium hydroxide[23] (1.0 g,
5.8 mmol) and benzaldehyde (9.34 g, 88.0 mmol) were added successively
to a solution of 2-acetylnaphthalin (10.0 g, 58.8 mmol) in ethanol (80 mL).
The reaction mixture was heated under reflux for 2.5 h. After cooling to
RT, a precipitate formed which was collected by filtration, then washed
with water (3 Â 5 mL) and ethanol (2 Â 5 mL). The residue was dried in
vacuo to give benzylidene-2-acetophenone as pale yellow crystals. Yield:
14.4 g, 95%, m.p. 1038C (ref. [46], 1058C); 1H NMR (200 MHz, CDCl3):
d ˆ 7.43 ± 7.47 (m, 3H), 7.56 ± 7.62 (m, 2H), 7.65 ± 7.74 (m, 3H), 7.86 ± 8.02
(m, 4H), 8.12 (dd, J(H,H) ˆ 8.74, 1.76 Hz, 1H), 8.55 (s, 1H); 13C NMR
(50.33 MHz, CDCl3): d ˆ 122.0, 124.4, 126.7, 127.8, 128.3 (2C), 128.5 (2
signals), 128.9 (2C), 129.5, 129.9, 130.5, 132.5, 134.9, 135.4, 135.5, 144.7,
190.2. Analytical data correspond to those reported previously.[46]
6-Methyl-2,4-diphenylpyrylium tetrafluoroborate: In analogy to a known
procedure,[49] the pyrylium salt was obtained from borontrifluoride etherate
(22.2 g, 327 mmol), acetophenone (16.0 g, 133 mmol), and acetic anhydride
(14.1 g, 138 mmol) as yellow crystals. Yield: 7.9 g, 35%; m.p. 2488C;
1H NMR (200 MHz, TFA): d ˆ 3.19 (s, 3H; CH3), 7.78 ± 7.97 (m, 6H; ArH),
8.17 ± 8.27 (m, 3H; ArH), 8.36 (d, J(H,H) ˆ 1.2 Hz, 1H; ArH), 8.40 (s, 1H),
8.72 (s, 1H); 13C NMR (TFA, 50.33 MHz): d ˆ 22.8, 116.9, 120.7, 130.8,
131.0 (2C), 131.7 (2C), 133.0 (2C), 133.2 (2C), 135.0, 138.9, 139.0, 171.0,
176.4, 179.8; elemental analysis calcd (%) for C18H15BF4O (334.11): C
64.71, H 4.53; found C 64.74, H 4.53.
General procedure for the preparation of 2,4,6 aryl-substituted pyrylium
salts:[22] Tetrafluoroboric acid (52% ethereal solution, 2.2 equiv) was added
at 708C within 2 min to a solution of the corresponding chalcone derivative
(2.2 equiv) and acetyl derivative (1 equiv) in 1,2-dichloroethane. The
reaction mixture was heated under reflux for 4 h, then allowed to cool to
RT; this was followed by the addition of a threefold amount of ether. An oil
formed which was separated and recrystallized from hot ethanol.
4-Trifluormethylbenzylidene acetophenone: 4-Trifluoromethyl benzalde-
hyde (5.0 g, 28.7 mmol) was added at RT to a solution of acetophenone
(2.87 g, 23.9 mmol) and activated barium hydroxide (200 mg, 1.2 mmol) in
ethanol (20 mL). After the solution had been stirred magnetically for 5 min
at RT, a yellow precipitate formed, which was collected by filtration then
washed with water (2 Â 2 mL) and ethanol (2 Â 1 mL), and dried in vacuo
to give the product as yellow crystals. Yield: 6.2 g, 94%; m.p. 1248C
(ref. [47], 1258C); 1H NMR (300 MHz, CDCl3): d ˆ 7.41 ± 7.77 (m, 9H), 8.03
(dd, J(H,H) ˆ 6.7, 1.8 Hz, 2H); 13C NMR (CDCl3, 50.33 MHz): d ˆ 124.2,
2-(2-Naphthyl)-4,6-diphenylpyrylium tetrafluoroborate: The pyrylium salt
was obtained from chalcone (5.00 g, 24.0 mmol), 2-acetonaphthone (2.04 g,
12.0 mmol), and tetrafluoroboric acid (52% in ether, 4.68 g, 28.8 mmol) in
1,2-dichloroethane (20 mL) as orange crystals. Yield: 7.61 g, 57%; m.p.
2548C; 1H NMR (300 MHz, TFA/CDCl3): d ˆ 7.34 ± 7.42 (m, 1H), 7.50 ± 7.75
(m, 8H), 7.80 ± 8.20 (m, 7H), 8.25 (s, 1H), 8.47 (s, 1H), 8.69 (s, 1H);
13C NMR (TFA/CDCl3, 75.47 MHz): d ˆ 113.8, 114.3, 122.0, 125.2, 127.9,
128.2 (two signals), 128.7 (2C), 129.0 (2C), 129.3 (two signals), 130.0 (2C),
130.3, 130.7 (2C), 131.3, 132.0, 132.7, 135.7 (two signals), 136.1, 165.5, 170.1,
3
125.8 (q, J(C,F) ˆ 3.7 Hz, 2C), 128.6 (2 signals, each 2C), 128.7, 129.3 (q,
1J(C,F) ˆ 285 Hz), 131.5, 133.1 (2C), 137.7, 138.2, 189.9; the signal for C-CF3
could not be detected. 19F NMR (CDCl3, 188 MHz): d ˆ 63.3; elemental
analysis calcd (%) for C16H11F3O (276.3): C 69.57, H 4.01; found C 69.44, H
4.12.
‡
170.3; FAB /HR (C25H15F6O): calcd 359.1436; found 359.1453.
4-Methoxybenzylidene acetophenone: 4-Methoxybenzaldehyde (12.5 g,
91.6 mmol) was added at RT to a solution of acetophenone (10.0 g,
83.2 mmol) and activated barium hydroxide (300 mg, 1.8 mmol) in ethanol
(40 mL). The reaction mixture was heated under reflux for 6 h. After
cooling to RT, the solvent was reduced in vacuo to a volume of 30 mL.
Pouring this mixture into vigorously stirred petroleum ether (50 mL)
provided a precipitate, which was collected by filtration. Drying in vacuo
provided the product as a pale yellow powder. Yield: 9.5 g, 48%; m.p. 708C
(ref. [48], 728C); 1H NMR (300 MHz, CDCl3): d ˆ 3.81 (s, 3H; OCH3), 6.90
(dt, 3J(H,H) ˆ 8.8 Hz, 4J(H,H) ˆ 2.9 Hz, 2H), 7.41 (d, 3J(H,H) ˆ 15.6 Hz,
1H), 7.44 ± 7.60 (m, 5H; Ph), 7.78 (d, 3J(H,H) ˆ 15.6 Hz, 1H), 8.01 (dt,
3J(H,H) ˆ 6.9 Hz, 4J(H,H) ˆ 2.1 Hz, 2H); 13C NMR (CDCl3, 50.33 MHz):
d ˆ 55.3, 114.3 (2C), 119.7, 127.5, 128.3 (2C), 128.4 (2C), 130.1 (2C), 132.4,
138.4, 144.5, 161.6, 190.4; elemental analysis calcd (%) for C16H14O2 (238.3):
C 80.65, H 5.92; found C 80.46, H 6.11.
2,6-Di-(2-naphthyl)-4-phenylpyrylium tetrafluoroborate: The pyrylium salt
was obtained as orange crystals from 1-(2-naphthyl)-3-phenylpropenone
(7.00 g, 27.1 mmol), 2-acetonaphthone (2.31 g, 13.6 mmol), and tetrafluoro-
boric acid (52% in ether, 4.85 g, 29.8 mmol) in 1,2-dichloroethane (50 mL).
Yield: 1.97 g, 30%; m.p. 2548C; 1H NMR (300 MHz, [D6]acetone):
d ˆ 7.77 ± 7.95 (m, 7H), 8.18 (d, J(H,H) ˆ 8.0 Hz, 2H), 8.34 (d, J(H,H) ˆ
8.1 Hz, 2H), 8.36 (d, J(H,H) ˆ 8.7 Hz, 2H), 8.64 (dt, J(H,H) ˆ 6.2 Hz,
1.4 Hz, 2H), 8.74 (dd, J(H,H) ˆ 8.0 Hz, 2.0 Hz, 2H), 9.36 (s, 2H), 9.43 (d,
J(H,H) ˆ 2.0 Hz, 2H); 13C NMR (TFA/CDCl3, 75.47 MHz): d ˆ 115.1 (C2,
2C), 123.2 (2C), 126.8 (2C), 129.9 (2C), 130.1, 130.4 (2C), 130.7 (2C), 131.8
(2C), 132.1 (2C), 132.3 (2C), 132.5 (2C), 132.7 (2C), 132.9 (2C), 134.8,
138.0, 138.5, 167.0 (C-3), 171.9 (C-1, 2C).
4-[4-(Trifluormethyl)phenyl]-2,6-diphenylpyrylium tetrafluoroborate: The
pyrylium salt was obtained as yellow crystals from 1-phenyl-3-(4-trifluoro-
methylphenyl)propenone (3.00 g, 10.9 mmol), 2-acetophenone (0.65 g,
5.43 mmol), and tetrafluoroboric acid (52% in ether, 1.77 g, 10.9 mmol)
in 1,2-dichloroethane (15 mL). Yield: 0.9 g, 36%; m.p. 2478C; 1H NMR
(300 MHz, TFA/CDCl3): d ˆ 7.40 ± 7.96 (m, all H); 13C NMR (TFA/CDCl3,
75.47 MHz): d ˆ 115.5 (2C), 124.2 (q, 1J(C,F) ˆ 270 Hz, CF3), 126.3 (q,
3J(C,F) ˆ 3.8 Hz, 2C), 128.8, 129.4 (4C), 129.5 (2C), 129.7 (4C), 131.0 (2C),
Pyrylium Salts
2,6-Dimethyl-4-phenylprylium
tetrafluoroborate:[24]
a-Methylstyrene
(28.4 g, 240 mmol) was added to a solution of tetrafluoroboric acid (54%
in diethyl ether, 21.1 g, 240 mmol) in acetic anhydride (200 mL) at 08C and
the reaction mixture was stirred for 2 h. After standing overnight at RT, the
reaction mixture was poured into ether (400 mL). A precipitate formed
which was collected by filtration. Recrystallization from hot water
(100 mL) gave the product as yellow crystals. Yield: 16.5 g, 25%; m.p.
1968C; 1H NMR (200 MHz, trifluoroacetic acid (TFA)/CDCl3): d ˆ 2.96 (s,
6H; CH3), 7.62 ± 7.72 (m, 2H; ArH), 7.75 ± 7.84 (m, 1H; ArH), 7.98 ± 8.05
(m, 2H; ArH), 8.07 (s, 2H; C3-H); 13C NMR (TFA/CDCl3, 50.33 MHz):
d ˆ 20.9 (2C), 118.3 (C3, 2C), 129.7 (2C), 130.7, 131.6 (2C), 136.5, 168.3
(C4), 178.7 (C2, 2C); elemental analysis calcd (%) for C13H13BF4O (199.1):
C 57.39, H 4.82; found C 57.39, H 4.93.
‡
133.7 (q, 2J(C,F) ˆ 33.0 Hz), 135.3 (2C), 167.0, 173.0 (2C); FAB /HR
(C24H16F3O): calcd 377.1153; found 377.1153.
2-[3,5-Bis(trifluoromethyl)phenyl]-4,6-diphenylpyrylium tetrafluoroborate:
The pyrylium salt was obtained as yellow crystals from 3,5-bis(trifluoro-
methyl)acetophenone (5.00 g, 19.5 mmol), chalcone (8.13 g, 39.0 mmol),
and tetrafluoroboric acid (52% in ether, 6.10 g, 39.0 mmol) in 1,2-
dichloroethane (10 mL). Yield: 6.50 g, 62%; m.p. 2528C; 1H NMR
(300 MHz, CDCl3): d ˆ 7.62 ± 7.88 (m, 7H), 8.11 ± 8.28 (m, 5H), 8.62 (d,
J(H,H) ˆ 1.6 Hz, 1H), 8.64 (s, 1H), 8.67 (d, J(H,H) ˆ 1.6 Hz, 1H); 13C NMR
(TFA, 75.47 MHz): d ˆ 117.6, 117.8, 124.6 (q, 1J(C,F) ˆ 272 Hz, 2CF3), 130.1
(2C), 130.4 (2C), 130.7 (2C), 131.1 (2C), 131.5, 132.2, 132.5, 133.0, 134.3,
2,6-Diisopropyl-4-phenylpyrylium tetrafluoroborate: a-methylstyrene
(14.2 g, 120 mmol) was added slowly at 08C to a solution of tetrafluoroboric
acid (54% in diethyl ether, 20.1 g, 120 mmol) in isobutyric anhydride
(100 mL). The reaction mixture was stirred for 1 h at RTand additional 5 h
at reflux. After cooling the mixture to RT, the precipitated solid was
collected by filtration. After washing with cold ether (20 mL) the residue
2
‡
136.7 (q, J(C,F) ˆ 35 Hz, 2C), 138.7, 138.8, 169.8, 170.8, 175.7; FAB /HR
(C25H15F6O): calcd 445.1027; found 445.1009; elemental analysis calcd (%)
for C25H15BF10O (532.18): C 56.41, H 2.84; found C 56.04, H 2.92.
3116
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3116 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
2-[3,5-Bis(trifluormethyl)phenyl]-4-(4-trifluormethylphenyl)-6-phenylpyry-
lium tetrafluoroborate: The pyrylium salt was obtained as yellow crystals
from 1-phenyl-3-(4-trifluoromethylphenyl)propenone (6.90 g, 25.0 mmol),
3,5-bis(trifluoromethyl)acetophenone (3.20 g, 12.5 mmol), and tetrafluoro-
boric acid (52% in ether, 4.22 g, 25.0 mmol) in 1,2-dichloroethane (25 mL).
Yield: 1.60 g, 22%; m.p. 2428C; 1H NMR (300 MHz, TFA/CDCl3): d ˆ
7.67 ± 7.76 (m, 3H), 7.80 ± 7.90 (m, 3H), 8.18 ± 8.30 (m, 5H), 8.62 ± 8.73 (m,
3H); 13C NMR (TFA/CDCl3, 75.47 MHz): d ˆ 117.1, 117.4, 127.2 (m, 2C),
127.9, 128.2 (2C), 128.6 (2C), 129.0, 129.8, 130.5, 130.6, 134.1 (q, 2J(C,F) ˆ
33 Hz), 135.6, 136.5, 137.1, 166.8, 168.0, 173.7 (signals for CF3 groups could
not be detected because of low intensity); 19F NMR (188 MHz, CDCl3):
orange-yellow solid was collected by filtration, then washed with water and
toluene, and dried in vacuo. Recrystallization from hot methanol/dichloro-
methane provided the pyrylium salt as a bright yellow solid. Yield: 180 g,
32%; m.p. 2018C; 1H NMR (200 MHz, CDCl3): d ˆ 2.43 (s, 6H; CH3), 2.57
(s, 6H; CH3), 7.25 (s, 3H), 7.30 (s, 1H), 7.58 ± 7.73 (m, 3H), 7.82 (d, J(H,H) ˆ
8.0 Hz, 2H), 8.08 (dd, J(H,H) ˆ 8.0 Hz, 1.8 Hz, 2H), 8.20 (d, J(H,H) ˆ
9.4 Hz, 2H); 13C NMR (50.34 MHz, CDCl3): d ˆ 21.0, 21.6, 118.0, 126.3,
128.4, 129.3, 130.3, 131.3, 132.8, 133.5, 135.1, 138.4, 145.7, 165.8, 173.8;
elemental analysis calcd (%) for C27H25BF4O (452.29): calcd C 71.70, H
5.57; found C 71.69, H 5.74.
General procedure for the preparation of phosphabenzenes
‡
d ˆ 64.0, 64.2, 64.3; FAB /HR (C26H14F9O): calcd 513.0901; found
From pyrylium salts and tris(trimethylsilyl)phosphane (Method A):[19]
Tris(trimethylsilyl)phosphane (2 ± 2.5equiv) was added dropwise at RT to
a magnetically stirred solution of the corresponding pyrylium salt (1equiv)
in acetonitrile. The resulting black reaction mixture was heated under
reflux for 5 h. After cooling to RT, the solvent was removed in vacuo. The
residue was dissolved in CH2Cl2 and an appropriate amount of silica gel was
added (0.5 gmmol 1). Evaporation of the solvent was followed by flash
chromatography with petroleum ether/ethyl acetate (19:1) to give the
corresponding phosphabenzene as a colorless to pale yellow solid or oil.
513.0921.
2,6-Diphenyl-4-(4-methoxyphenyl)-pyrylium tetrafluoroborate: The pyryli-
um salt was obtained as yellow crystals from 1-phenyl-3-(4-methoxyphe-
nyl)propenone (9.50 g, 41.9 mmol), acetophenone (2.52 g, 21.0 mmol), and
tetrafluoroboric acid (52% in ether, 6.90 g, 42.0 mmol) in 1,2-dichloro-
ethane (70 mL). Yield: 3.30 g, 37%; m.p. 2338C; 1H NMR (200 MHz, TFA/
CDCl3): d ˆ 4.06 (s, 3H; OCH3), 7.28 (dd, J(H,H) ˆ 4.0 Hz, 1.7 Hz, 2H),
7.32 ± 7.90 (m, 6H), 8.21 ± 8.37 (m, 6H), 8.48 (s, 2H); 13C NMR (TFA/
CDCl3, 50.33 MHz): d ˆ 56.2, 112.6 (2C), 116.8 (2C), 123.1, 128.2 (4C),
128.8 (2C), 130.7 (4C), 132.3 (2C), 136.1 (2C), 165.3, 167.7, 170.8 (2C);
elemental analysis calcd (%) for C24H19BF4O2 (426.21): C 67.63, H 4.49;
found C 67.28, H 4.32.
From pyrylium salts and phosphane (Method B):[21] A solution of the
corresponding pyrylium salt in a suitable solvent was placed in an autoclave
(material: HC). In some cases hydrogen bromide was added as a catalyst.
(It was later found that the addition of the acid catalyst is not essential.)
The autoclave was pressurized with phosphane gas (5 bar) at room
temperature, then the reaction mixture was heated to 1108C with vigorous
mechanical stirring. When the reaction temperature was reached, the
phosphane pressure was increased to 30 bar. After 4 h the autoclave was
cooled to room temperature, depressurized, and phosphane traces were
removed by flushing with nitrogen. The reaction mixture was concentrated,
then filtered. The remaining solid was washed with the original solvent,
dissolved in toluene and subsequently washed with several portions of
water until the water phase was neutral. Toluene was removed in vacuo,
and the residue was washed with pentane to yield pure phosphabenzene. If
necessary further purification could be achieved by recrystallization.
2-(2-Methoxyphenyl)-4,6-diphenyl-pyrylium tetrafluoroborate: The pyryli-
um salt was obtained as yellow crystals from 2-methoxyacetophenone
(8.00 g, 53.2 mmol), chalcone (22.2 g, 106 mmol), and tetrafluoroboric acid
(52% in ether, 18.0 g, 106 mmol) in 1,2-dichloroethane (20 mL). Yield:
13.2 g, 58%; m.p. 1688C; 1H NMR (300 MHz, CDCl3): d ˆ 4.04 (s, 3H;
OCH3), 7.20 (t, J(H,H) ˆ 7.6 Hz, 1H), 7.25 (d, J(H,H) ˆ 8.5 Hz, 1H), 7.50 ±
7.60 (m, 7H), 8.06 (dd, J(H,H) ˆ 8.0 Hz, 1.5 Hz, 1H), 8.11 ± 8.16 (m, 2H),
8.22 ± 8.28 (m, 2H), 8.48 (d, J(H,H) ˆ 1.5 Hz, 1H), 8.66 (d, J(H,H) ˆ 1.5 Hz,
1H); 13C NMR (CDCl3, 75.47 MHz): d ˆ 56.5, 112.9, 113.9, 116.9, 118.0,
122.2, 128.0, 128.3 (2C), 128.4, 129.5 (2C), 130.0 (2C), 130.5 (2C), 132.2,
135.0, 135.1, 137.2, 159.7, 165.3, 168.4, 170.1; elemental analysis calcd (%)
for C24H19BF4O2 (426.20): C 67.63, H 4.49; found C 67.38, H 4.67.
2,6-Dimethyl-4-phenylphosphinine (1):[50] Phosphinine 1 was obtained as a
colorless solid according to method A from the corresponding pyrylium salt
(5 g, 18.4 mmol) and P(TMS)3 (9.20 g, 36.8 mmol) in acetonitrile (50 mL).
2-(3-Methoxyphenyl)-4,6-diphenyl-pyrylium tetrafluoroborate: The pyryli-
um salt was obtained as yellow crystals from 3-methoxyacetophenone
(15.0 g, 99.9 mmol), chalcone (41.6 g, 200 mmol), and tetrafluoroboric acid
(52% in ether, 33.7 g, 200 mmol) in 1,2-dichloroethane (40 mL). Yield:
41.6 g, 57%; m.p. 1818C; 1H NMR (300 MHz, TFA/CDCl3): d ˆ 3.96 (s,
3H; OCH3), 7.60 ± 7.90 (m, 10H), 8.10 ± 8.16 (m, 2H), 8.24 ± 8.28 (m, 2H),
8.48 (d, J(H,H) ˆ 1.5 Hz, 1H), 8.51 (d, J(H,H) ˆ 1.5 Hz, 1H); 13C NMR
(TFA/CDCl3, 75.47 MHz): d ˆ 55.9, 114.7, 114.9, 121.0, 121.4, 128.3, 129.2
(2C), 129.4(2C), 129.6(2C), 129.7(2C), 130.4, 130.5, 131.7, 132.4, 136.0,
136.1, 160.7, 167.0, 170.9, 171.3.
1
Yield: 876 mg, 24%; m.p. 618C; H NMR (200 MHz, CDCl3): d ˆ 2.74 (d,
3J(H,P) ˆ 15.1 Hz, 6H; CH3), 7.32 ± 7.44 (m, 3H), 7.53 (d, J ˆ 7.7 Hz, 2H),
7.71 (d, J ˆ 6.7 Hz, 2H); 13C NMR (75.47 MHz, CDCl3): d ˆ 24.6 (d,
2J(C,P) ˆ 35.8 Hz, 2CH3), 127.6 (two signals, 2C each), 128.8, 132.1 (d,
4
3
2J(C,P) ˆ 13.2 Hz, C3, 2C), 142.3 (d, J(C,P) ˆ 3.2 Hz), 143.3 (d, J(C,P) ˆ
15.3 Hz, C4), 168.4 (d, 1J(C,P) ˆ 49.6 Hz, C2, 2C); 31P NMR (81 MHz,
CDCl3): d ˆ 194.3; elemental analysis calcd (%) for C13H13P (200.20): C
77.98, H 6.55; found C 77.74, H 6.63.
6-Methyl-2,4-diphenylphosphinine (2):[20] Phosphinine 2 was obtained as a
colorless solid according to method A from the corresponding pyrylium salt
(5.01 g, 15.0 mmol) and P(TMS)3 (11.30 g, 45.0 mmol) in acetonitrile
(60 mL). Yield: 1.0 g, 25%; m.p. 768C; 1H NMR (200 MHz, CDCl3): d ˆ
2.83 (d, 2J(H,P) ˆ 15.5 Hz, 3H; CH3), 7.45 (m, 6H), 7.65 (d, J ˆ 7.3 Hz, 2H),
7.70 (d, J ˆ 8.0 Hz, 2H), 7.88 (d, J ˆ 7.0 Hz, 1H), 8.06 (d, J ˆ 5.7 Hz, 1H);
13C NMR (100.62 MHz, CDCl3): d ˆ 24.8 (d, 2J(C,P) ˆ 37.2 Hz, CH3), 127.5
(2C), 127.7 (two signals), 127.8 (2C), 128.8 (2C), 128.9 (2C), 130.9 (d,
2J(C,P) ˆ 12.7 Hz, C3/5), 133.0 (d, 2J(C,P) ˆ 12.7 Hz, C3/5), 142.3, 143.5 (d,
3J(C,P) ˆ 23.7 Hz), 143.7 (d, 3J(C,P) ˆ 14.9 Hz, C4), 168.6 (d, 1J(C,P) ˆ
50.6 Hz, C2/6), 171.6 (d, 1J(C,P) ˆ 51.2 Hz, C2/6); 31P NMR (81 MHz,
CDCl3): d ˆ 189.7; elemental analysis calcd (%) for C18H15P (262.27): C
82.43, H 5.77; found C 82.36, H 5.61.
8-(2-Naphthyl)-10,11-diphenyl-11H-benzo[4,5]indeno[1,2-b]pyrylium tet-
rafluoroborate 16: Tetrafluoroboric acid (52% ethereal solution, 3.1 g,
19.4 mmol) was added within 2 min to a solution of 1-(2-naphtyl)-3-
phenylpropenone (5.00 g, 19.4 mmol) in 1,2-dichloroethane (15 mL) at
708C. The reaction mixture was heated to reflux for 4 h, then allowed to
cool to RT and poured into ether (45 mL). An oil precipitated which was
separated and recrystallized from hot ethanol to give the pyrylium salt 16 as
yellow crystals. Yield: 3.20 g, 28%; m.p. 2248C; 1H NMR (300 MHz, TFA/
CDCl3): d ˆ 5.89 (s, 1H; CHPh), 7.38 ± 7.46 (m, 4H), 7.55 (s, 1H), 7.57 (s,
2H), 7.60 ± 7.80 (m, 3H), 7.88 (d, J ˆ 8.1 Hz, 1H), 7.91 ± 7.99 (m, 5H), 8.00 ±
8.10 (m, 2H), 8.12 (d, J(H,H) ˆ 1.1 Hz, 2H), 8.24 (dd, J(H,H) ˆ 12.8 Hz,
J ˆ 8.7 Hz, 1H), 8.35 (s, 2H), 8.77 (s, 1H); 13C NMR (75.47 MHz, TFA/
CDCl3): d ˆ 51.0 (CHPh), 115.7, 117.4, 121.8, 125.2, 125.4 (two signals, 2C
each), 126.2, 128.0, 128.1 (two signals, 2C each), 128.2, 128.3, 128.4, 128.6,
129.0, 129.2, 129.5, 129.9, 130.1, 130.6, 131.5, 131.7, 132.6, 132.9, 133.15,
2,4,6-Triphenylphosphinine (3):[51, 21] According to method A, phosphinine
3 was obtained as a pale yellow solid from the corresponding pyrylium salt
(5.00 g, 12.6 mmol) and P(TMS)3 (6.32 g, 25.2 mmol) in acetonitrile
(30 mL); yield: 1.5 g, 35%. According to method B, apart from a reaction
time was 6 h, phosphinine 3 was obtained as a colorless solid from the
corresponding pyrylium salt (20 g, 50 mmol) in n-butanol (150 mL) after
recrystallization from diethyl ether/methanol. Yield: 12.6 g, 77%; m.p.
1738C; 1H NMR (300 MHz, CDCl3): d ˆ 7.38 ± 7.46 (m, 8H), 7.64 ± 7.73 (m,
7H), 8.21 (d, 3J(H,P) ˆ 6.0 Hz, 2H; H at C3/5); 13C NMR (75.47 MHz,
CDCl3): d ˆ 127.6 (2C), 127.7 (2C), 127.9 (4C), 128.9 (2C), 129.0 (4C),
131.9 (d, 2J(C,P) ˆ 12.1 Hz, C3/5, 2C), 142.1, 142.2, 143.3 (d, 2J(C,P)ˆ
‡
133.25, 136.3, 137.8, 138.5, 154.7, 161.4, 167.4, 174.5; FAB /HR (C38H25O):
calcd 497.1905; found 497.1903.
2,6-Bis(2,4-dimethylphenyl)-4-phenylpyrylium tetrafluoroborate:[21] Tetra-
fluoroboric acid (54% ethereal solution, 606 g, 3.7 mmol) was added slowly
at 808C to
a solution of benzaldehyde (271 g, 2.6 mmol) and 2,4-
dimethylacetophenone (542 g, 3.7 mol) in 1,2-dichloroethane. After stirring
for a further 4 h at this temperature, the reaction mixture was allowed to
cool to RT. The volatile components were removed in vacuo, and the
residue was treated with a toluene/water mixture (1:1). The precipitated
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3117 $ 17.50+.50/0
3117

B. Breit, T. Mackewitz et al.
FULL PAPER
3
1
24.1 Hz, 2C), 144.1 (d, J(C,P) ˆ 13.7 Hz, C4), 171.7 (d, J(C,P) ˆ 51.9 Hz,
C2/6, 2C); 31P NMR (81 MHz, CDCl3): d ˆ 185.1; elemental analysis calcd
(%) for C23H17P (324.34): C 85.17, H 5.28; found C 85.24, H 5.10.
CDCl3): d ˆ 55.9 (C6), 119.3 (d, 3J(C,P) ˆ 9.2 Hz, C9), 124.5 (C15), 125.6
(C20), 125.9, 126.0 (C34), 126.1, 126.2, 126.4, 126.6, 127.3 (C24), 127.6, 127.7,
127.9 (C19), 128.2 (C18), 128.4 (C23), 128.5, 128.6 (C22), 129.0, 129.1, 129.9,
132.9 (C28), 133.7 (C33), 134.1 (C16), 134.5 (d, 2J(C,P) ˆ 12.2 Hz, C2),
139.2 (C17), 141.0 (d, 2J(C,P) ˆ 24.9 Hz, C25), 141.7 (C21), 142.6 (d,
3J(C,P) ˆ 14.2 Hz, C3), 142.7 (C7), 143.4 (d, 2J(C,P) ˆ 11.2 Hz, C8), 151.7
(d, 3J(C,P) ˆ 12.7 Hz, C4), 168.4 (d, 1J(C,P) ˆ 48.3 Hz, C1), 169.2 (d,
1J(C,P) ˆ 47.8 Hz, C5); 31P NMR (81 MHz, CDCl3): d ˆ 162.7; HRMS
(C38H25P): calcd 512.1694; found 512.1694.
2,6-Bis-(2,4-dimethylphenyl)-4-phenylphosphinine (8):[21] According to
method B, phosphinine 8 was obtained as a colorless solid from the
corresponding pyrylium salt (20 g, 44 mmol) in n-butanol (150 mL) after
recrystallization from diethyl ether/methanol. Yield: 8.9 g, 53%; m.p.
1718C; 1H NMR (200 MHz, CD2Cl2): d ˆ 2.34 (s, 6H; CH3), 2.37 (s, 6H;
CH3), 7.04 ± 7.16 (m, 4H), 7.18 ± 7.28 (m, 2H), 7.30 ± 7.50 (m, 3H), 7.65 ± 7.68
(m, 2H), 7.93 (d, 3J(H,P) ˆ 6.2 Hz, 2H; C3/5-H); 13C NMR (50.23 MHz,
CDCl3): d ˆ 20.9 (d, J(C,P) ˆ 2.7 Hz), 21.1, 126.5, 127.7 (d, J(C,P) ˆ 5.5 Hz),
128.9, 130.2, 130.4, 131.3, 132.7, 132.8 (d, 2J(C,P) ˆ 11.0 Hz, C3/5), 134.8 (d,
J(C,P) ˆ 4.6 Hz), 137.4, 140.4 (d, 2J(C,P) ˆ 23.0 Hz), 142.1 (d, 3J(C,P) ˆ
13.8 Hz, C4), 172.5 (d, 1J(C,P) ˆ 53.9 Hz, C2/6); 31P NMR (81 MHz,
CDCl3): d ˆ 194.3; elemental analysis calcd (%) for C27H25P (380.46): C
85.24, H 6.62; found C 84.98, H 6.59.
4-(4-Methoxyphenyl)-2,6-diphenylphosphinine 9:[21, 51] According to meth-
od A, phosphinine 9 was obtained as a yellow solid from the corresponding
pyrylium salt (3.30 g, 7.74 mmol) and P(TMS)3 (3.88 g, 15.5 mmol) in
acetonitrile (10 mL); yield: 540 mg, 20%. According to method B,
phosphinine 9 was obtained as a pale yellow solid from the corresponding
pyrylium salt (13 g, 31 mmol) in n-butanol (150 mL). Yield: 4.3 g, 39%;
m.p. 1048C; 1H NMR (200 MHz, C6D6): d ˆ 3.44 (s, 3H; CH3), 6.97 (d, 3J ˆ
8.8 Hz, 2H), 7.34 ± 7.26 (m, 6H), 7.50 (d, J ˆ 8.5 Hz, 2H), 7.82 (dt, J ˆ 6.5 Hz,
1.5 Hz, 4H), 8.26 (d, 3J(H,P) ˆ 5.8 Hz, 2H; C3/5-H); 13C NMR (50.33 MHz,
C6D6): d ˆ 55.3, 115.2 (2C), 127.9 (2C), 128.4 (4C), 128.9 (4C), 129.6 (2C),
132.3 (d, 2J(C,P) ˆ 12.4 Hz, C3/5, 2C), 135.4 (d, 4J(C,P) ˆ 3.3 Hz), 144.5,
144.5 (d, 2J(C,P) ˆ 24.3 Hz, 2C), 160.6 (C11), 172.8 (d, 1J(C,P) ˆ 52 Hz, C2/
6, 2C); 31P NMR (81 MHz, C6D6): d ˆ 181.9; elemental analysis calcd (%)
for C24H19OP (354.36): C 81.34 H 5.40; found C 81.38, H 5.60.
2-(2-Naphthyl)-4,6-diphenylphosphinine (4):[21] According to method A,
phosphinine 4 was obtained as an orange solid from the corresponding
pyrylium salt (6.50 g, 14.5 mmol) and P(TMS)3 (4.38 g, 17.5 mmol) in
acetonitrile (55 mL); yield: 1.8 g, 33%. According to method B, apart from
a phosphane pressure of 20 bar at reaction temperature, phosphinine 4 was
obtained as a yellow solid from the corresponding pyrylium salt (10 g,
22 mmol) and 1 mL hydrogen bromide in acetic acid (30 wt%) in ethanol
(150 mL). Yield: 10 g, 84%; 1H NMR (200 MHz, CDCl3): d ˆ 7.41 ± 7.60 (m,
8H), 7.72 ± 7.81 (m, 4H), 7.88 ± 8.00 (m, 4H), 8.22 (dd, 3J ˆ 6.0 Hz,
4J ˆ 1.2 Hz, 1H; H4), 8.22 (s, 1H; H8), 8.32 (dd, 3J ˆ 6 Hz, 4J ˆ 1.2 Hz,
1H; H2); 13C NMR (100.61 MHz, CDCl3): d ˆ 126.0, 126.1, 126.2, 126.3,
126.5, 127.7 (2C), 127.8 (2C), 128.0 (2C), 128.4, 128.7, 128.9 (2C),
2
2
129.0 (2C), 131.8 (d, J(C,P) ˆ 12.1 Hz, C3/5), 131.9 (d, J(C,P) ˆ 12.1 Hz,
C3/5), 132.9, 133.7, 140.7 (d, 2J(C,P) ˆ 24.0 Hz), 142.2, 143.4 (d, 2J(C,P) ˆ
24.0 Hz), 144.2 (d, J(C,P) ˆ 13.6 Hz, C4), 171.6 (d, 1J(C,P) ˆ 51.5 Hz, C2/
3
6), 171.8 (d, 1J(C,P) ˆ 51.5 Hz, C2/6); 31P NMR (162 MHz, CDCl3): d ˆ
184.5.
2,6-Di-(2-naphthyl)-4-phenylphosphinine (5):[21] According to method A,
phosphinine 5 was obtained as a red-orange solid from the corresponding
pyrylium salt (4.00 g, 8.06 mmol) and P(TMS)3 (4.44 g, 17.7 mmol) in
acetonitrile (80 mL); yield: 450 mg, 15%. According to method B,
phosphinine 5 was obtained as a yellow solid from the corresponding
pyrylium salt (2.5 g, 5.0 mmol) and hydrogen bromide in acetic acid (1 mL,
30 wt%) in n-butanol (150 mL) after recrystallization from toluene/
pentane. Yield: 1.0 g, 47%; m.p. 1718C; 1H NMR (300 MHz, CD2Cl2):
d ˆ 7.57 (m, 7H), 7.75 (m, 2H), 7.80 ± 7.99 (m, 8H), 8.24 (s, 2H; H5), 8.29 (d,
J ˆ 6.0 Hz, 2H); 13C NMR (125.77 MHz, CDCl3): d ˆ 126.0 (d, J(C,P) ˆ
12.6 Hz, 2C), 126.3 (2C), 126.4, 126.5 (2C), 127.7 (2C), 127.9 (2C), 128.0
(2C), 128.4 (2C), 128.7 (2C), 129.0 (2C), 132.0 (d, 2J(C,P) ˆ 11.3 Hz, C3/5,
2
2C), 133.0 (2C), 133.7 (2C), 140.6 (d, J(C,P) ˆ 23.9 Hz, 2C), 142.2, 144.3
(d, 3J(C,P) ˆ 13.8 Hz, C4), 171.6 (d, 1J(C,P) ˆ 51.5 Hz, C-2/6, 2C); 31P NMR
(81 MHz, CDCl3): d ˆ 185.5; HRMS (C31H21P): calcd 424.1381; found
424.1380.
2,6-Diisopropyl-4-phenylphosphinine (6): According to method A, phos-
phinine 6 was obtained as a pale yellow viscous oil from the corresponding
pyrylium salt (10.00 g, 30.5 mmol) and P(TMS)3 (19.0 g, 76.0 mmol) in
4-[4-(Trifluormethyl)phenyl]-2,6-diphenylphosphinine (10): According to
method A, phosphinine 10 was obtained as a yellow solid from the
corresponding pyrylium salt (2.00 g, 4.3 mmol) and P(TMS)3 (2.7 g,
10.7 mmol) in acetonitrile (60 mL). Yield: 680 mg, 51%; 1H NMR
(200 MHz, C6D6): d ˆ 7.35 ± 7.54 (m, 8H), 7.70 (d, J ˆ 8.0 Hz, 2H), 7.91
(dt, J ˆ 8.0 Hz, 1.5 Hz, 4H), 8.17 (d, 3J(P,H) ˆ 6.0 Hz, 2H; C3/5-H);
13C NMR (125.77 MHz, CDCl3): d ˆ 124.2 (q, 1J(C,F) ˆ 273 Hz, CF3),
125.9 (q, 3J(C,F) ˆ 3.8 Hz, 2C), 127.7 (d, 3J(C,P) ˆ 12.5 Hz, 4C), 128.10
(2C), 128.15 (2C), 129.0 (4C), 130.0 (q, 2J(C,F) ˆ 32.7 Hz), 131.5 (d,
2J(C,P) ˆ 11.3 Hz, C3/5, 2C), 142.5 (d, 3J(C,P) ˆ 13.7 Hz, C4), 143.1 (d,
2J(C,P) ˆ 23.9 Hz, 2C), 145.7, 172.1 (d, 1J(C,P) ˆ 52.8 Hz, C2/6, 2C); 31P
NMR (81 MHz, CDCl3): d ˆ 189.6; 19F NMR (C6D6, 188 MHz): d ˆ 62.4;
elemental analysis calcd (%) for C24H16F3P (392.34): C 73.47 H 4.11; found
C 73.20, H 3.85.
1
acetonitrile (25 mL). Yield: 2.3 g, 30%; H NMR (300 MHz, CDCl3): d ˆ
3
3
1.41 (d, J(H,H) ˆ 7 Hz, 12H; CH3), 3.31 (sept, J(H,H) ˆ 7 Hz, 2H; CH),
7.36 ± 7.60 (m, 5H; Ph), 7.83 (d, 3J(H,P) ˆ 7 Hz, 2H; H4); 13C NMR
(125.77 MHz, CDCl3): d ˆ 26.0 (2C), 38.0 (d, 2J(C,P) ˆ 29.0 Hz, 2C), 127.5,
127.7 (d, J(C,P) ˆ 1.9 Hz, 2C), 128.8 (2C), 130.1 (d, 3J(C,P) ˆ 12.2 Hz, C3/5,
2C), 142.9 (d, 4J(C,P) ˆ 3.0 Hz), 143.1 (d, 3J(C,P) ˆ 16 Hz, C4), 181.6 (d,
1J(C,P) ˆ 52.8 Hz, C1/6, 2C); 31P NMR (81 MHz, CDCl3): d ˆ 187.4;
elemental analysis calcd (%) for C17H21P (256.31): C 79.66, H 8.26; found
C 79.53, H 8.50.
8-Naphthalin-2-yl-10,11-diphenyl-11H-7-phosphabenzo[a]fluoren (7): Ac-
cording to method A, phosphinine 7 was obtained from the corresponding
pyrylium salt (2.90 g, 4.96 mmol) and P(TMS)3 (2.9 g, 11.7 mmol) in
2-[3,5-Bis(trifluormethyl)phenyl]-4,6-diphenylphosphinine (11): Accord-
ing to method A, phosphinine 11 was obtained as a yellow solid from the
corresponding pyrylium salt (4.00 g, 7.5 mmol) and P(TMS)3 (4.70 g,
18.8 mmol) in acetonitrile (20 mL). Yield: 700 mg, 20%; 1H NMR
(200 MHz, CDCl3): d ˆ 7.42 ± 7.55 (m, 6H), 7.68 ± 7.76 (m, 4H), 7.94 (s,
1H; H1), 8.16 (s, 3H), 8.27 (d, 3J ˆ 6.2 Hz, 1H; H9); 13C NMR
(125.77 MHz, CDCl3): d ˆ 121.5 (m), 123.3 (q, 1J(C,F) ˆ 273 Hz, 2CF3),
127.67 (d, 3J(C,P) ˆ 11.3 Hz, 2C), 127.72 (m, 2C), 127.8 (2C), 128.4 (2C),
129.1 (2C), 129.2 (2C), 132.1 (d, 2J(C,P) ˆ 12.5 Hz, C3/5), 132.3 (q,
2J(C,F) ˆ 32.7 Hz, 2C), 132.8 (d, 2J(C,P) ˆ 12.5 Hz, C3/5), 141.6, 142.7 (d,
2J(C,P) ˆ 23.8 Hz), 144.9 (d, 3J(C,P) ˆ 13.8 Hz, C8), 145.4 (d, 2J(C,P) ˆ
26.4 Hz), 168.1 (d, 1J(C,P) ˆ 52.8 Hz, C2/6), 172.2 (d, 1J(C,P) ˆ 52.8 Hz,
C2/6); 31P NMR (202 MHz, CDCl3): d ˆ 186.2; HRMS (C25H15F6P): calcd
460.0815 found 460.0812.
acetonitrile (20 mL, 1.15 g, 45%) as an orange solid. Assignment of signals
is based on 1J(C,H), 2,3J(C,H) and 3J(H,H) correlation, and 13C DEPT
NMR experiments. 1H NMR (300 MHz, CDCl3): d ˆ 5.65 (s, 1H; H6), 6.46
(d, 3J(H,H) ˆ 6.8 Hz, 2H; H18), 6.80 ± 6.96 (m, 3H; H19, H20), 7.13 (d,
3J(H,H) ˆ 8.0 Hz, 2H; H22), 7.20 ± 7.39 (m, 5H; H23, H-24, H13‡1ArH),
2-[3,5-Bis(trifluormethyl)phenyl]-4-[4-(trifluormethyl)phenyl]-6-phenyl-
phosphinine (12): According to method A, phosphinine 12 was obtained as
a yellow solid from the corresponding pyrylium salt (1.50 g, 2.5 mmol) and
P(TMS)3 (1.60 g, 6.25 mmol) in acetonitrile (20 mL). Yield: 270 mg, 20%;
1H NMR (300 MHz, CDCl3): d ˆ 7.38 (d, J ˆ 6.0 Hz, 3H), 7.62 (d, J ˆ
6.2 Hz, 1H), 7.68 (s, 5H), 7.84 (s, 1H; H1), 8.04 (s, 3H), 8.12 (d, Jˆ
3
7.42 ± 7.50 (m, 2H; H30/32‡1 ArH), 7.67 (d, J(H,H) ˆ 8.2 Hz, 1H; H15),
7.82 ± 7.94 (m, 7H; H2, H10, H30/32, H26, ‡ 3 ArH), 8.19 (s, 1H; H34), 8.30
3
4
(dd, J(P,H) ˆ 8.4 Hz, J(H,H) ˆ 1.5 Hz, 1H; H9); 13C NMR (125.77 MHz,
3118
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3118 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
1
6.0 Hz, 1H); 13C NMR (75.47 MHz, CDCl3): d ˆ 121.8, 123.3 (q, J(C,F) ˆ
273 Hz, 2CF3), 126.2 (q, 3J(C,F) ˆ 3.7 Hz, 2C), 127.6 (2C), 127.7 (2C), 128.2
(2C), 128.6 (2C), 129.2, 130.4 (q, 2J(C,P) ˆ 32.6 Hz, C15), 131.8 (d,
2J(C,P) ˆ 12.3 Hz, C3/5), 132.4 (q, 2J(C,F) ˆ 33.5 Hz, 2C), 132.5 (d,
2J(C,P) ˆ 11.4 Hz, C3/5), 142.4 (d, 2J(C,P) ˆ 24.4 Hz), 143.3 (d, 3J(C,P) ˆ
2J(P,C) ˆ 22.1 Hz, CO); 31P NMR (162 MHz, CDCl3): d ˆ 168.9
1
‡
(d,
1J(Rh,P) ˆ 174 Hz);
IR(CH2Cl2):
nÄ ˆ 1999cm
,
FAB /HR
(C47H34ClOP2Rh): calcd. 814.0828; found 814.0828.
X-ray crystal structure analysis of rhodium complex 22: [C46H34ClOP2Rh] ´
Å
0.5CH2Cl2, Mw ˆ 857.51, triclinic, space group P1, a ˆ 1096.6(1), b ˆ
1
13.7 Hz, C4), 145.0 (m), 145.3, 168.5 (d, J(C,P) ˆ 53.0 Hz, C1/6), 172.6 (d,
1386.8(1), c ˆ 1430.6(1) pm, a ˆ 79.09(1), b ˆ 74.83(1), g ˆ 74.74(1)8, U ˆ
1J(C,P) ˆ 52.0 Hz, C1/6); 31P NMR (81 MHz, CDCl3): d ˆ 191.5; 19F NMR
(188 MHz, CDCl3): d ˆ 62.8, 63.2; ; HRMS (C26H14F9P): calcd 528.0689
found 528.0692.
3
2008(3) Š3, Dc ˆ 1.418 gcm
for Z ˆ 2, F(000) ˆ 874, m(MoKa) ˆ
0.674mm 1, l ˆ 0.71073 Š, Tˆ 193(2) K, crystal size 0.3  0.3  0.2 mm.
A total of 8182 reflections was collected (Enraf Nonius CAD4) by using W/
2q scans in the range 2.3 < q < 268, 7851 were unique (Rint ˆ 0.0493). The
structure was solved by direct methods. Full-matrix least-squares refine-
ment was based on F2, with all non-hydrogen atoms anisotropic and with
hydrogens included in calculated positions with isotropic temperature
factors 1.2 times that of the Ueq of the atom to which they were bonded. The
refinement converged at R1 ˆ 0.0716 (for 6686 reflections with I > 2s(I))
and wR2 ˆ 0.2300 (all data) [w ˆ (s2(Fo)2 ‡ (0.1918P)2 ‡ 0.6691P) 1 where
P ˆ (Fo2 ‡ 2Fc2†/3]; final GOF ˆ 1.056; largest peak and hole in the final
difference Fourier: 2.26/3.37 eŠ 3. The programs used were SHELXS-97[53]
and SHELXL-97.[54] Crystallographic data (excluding structure factors)
have been deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC-156962. Copies of the data can be
obtained free of charge on application to CCDC, 112 Union Road,
Cambridge CB2 1EZ, UK (fax: (‡44)1233-336033; e-mail: deposit@ccdc.
cam.ac.uk).
2-(2-Methoxyphenyl)-4,6-diphenylphosphinine (13):[21] According to meth-
od A, phosphinine 13 was obtained as a yellow solid from the correspond-
ing pyrylium salt (10.00 g, 23.5 mmol) and P(TMS)3 (14.70 g, 58.7 mmol) in
acetonitrile (30 mL); yield: 3.95 g, 48%. According to method B, phosphi-
nine 13 was obtained as a pale yellow solid from the corresponding
pyrylium salt (20 g, 47 mmol) in n-butanol (150 mL). Yield: 4.7 g, 28%;
1
m.p. 788C; H NMR (300 MHz, CDCl3): d ˆ 3.83 (s, 3H; CH3), 7.00 ± 7.11
(m, 2H), 7.34 ± 7.52 (m, 8H), 7.64 ± 7.76 (m, 4H), 8.14 ± 8.20 (m, 2H; H at C3/
5); 13C NMR (75.47 MHz, CDCl3): d ˆ 55.8, 111.4, 121.0, 127.7, 127.76,
127.84, 128.1, 128.2 (2C), 128.6 (2C), 128.86 (2C), 128.92 (2C), 129.3, 131.4
(d, 2J(C,P) ˆ 11.7 Hz, C3/5), 133.9 (d, 2J(C,P) ˆ 12.1 Hz, C3/5), 142.3, 142.9
3
2
3
(d, J(C,P) ˆ 13.6 Hz, C4), 143.6 (d, J(C,P) ˆ 24.9 Hz), 155.9 (d, J(C,P) ˆ
1
1
4.2 Hz), 168.4 (d, J(C,P) ˆ 52.1 Hz, C2/6), 171.4 (d, J(C,P) ˆ 53.3 Hz, C2/
6); 31P NMR (81 MHz, CDCl3): d ˆ 191.8; elemental analysis calcd (%) for
C24H19OP (354.36): C 81.34 H 5.40; found C 81.10, H 5.44.
2-(3-Methoxyphenyl)-4,6-diphenylphosphinine (14):[21] According to meth-
od A, phosphinine 14 was obtained as a pale yellow solid from the
corresponding pyrylium salt (10.00 g, 23.5 mmol) and P(TMS)3 (11.80 g,
47 mmol) in acetonitrile (100 mL); yield: 2.95 g, 36%. According to
method B, phosphinine 14 was obtained as a pale yellow solid from the
corresponding pyrylium salt (13 g, 31 mmol) in n-butanol (150 mL). Yield:
4.3 g, 39%; m.p. 1218C; 1H NMR (500 MHz, CDCl3): d ˆ 3.87 (s, 3H; CH3),
6.96 (dd, J ˆ 7.9 Hz, 2.4 Hz, 1H), 7.32 (d, J ˆ 7.7 Hz, 1H), 7.37 ± 7.42 (m,
3H), 7.46 ± 7.50 (m, 4H), 7.68 (d, J ˆ 7.3 Hz, 2H), 7.73 (d, J ˆ 7.9 Hz, 2H),
8.17 (d, J ˆ 5.9 Hz, 2H); 13C NMR (75.47 MHz, CDCl3): d ˆ 55.3, 113.2 (d,
3J(C,P) ˆ 12.8 Hz), 113.6, 120.2 (d, 3J(C,P) ˆ 12.8 Hz), 127.6 (2C), 127.8
(two signals), 128.0 (2C), 128.9 (2C), 129.0 (2C), 129.9, 131.7 (d, 2J(C,P) ˆ
11.3 Hz, C3/5), 131.8 (d, 2J(C,P) ˆ 11.3 Hz, C3/5), 142.1 (d, 4J(C,P) ˆ 3 Hz),
143.4 (d, 2J(C,P) ˆ 24.1 Hz), 144.1 (d, 3J(C,P) ˆ 13.7 Hz, C4), 144.8 (d,
2J(C,P) ˆ 24.9 Hz), 160.0, 171.6 (d, 1J(C,P) ˆ 52.8 Hz, C2/6), 171.8 (d,
1J(C,P) ˆ 52.1 Hz, C2/6); 31P NMR (81 MHz, CDCl3): d ˆ 185.2; elemental
analysis calcd (%) for C24H19OP (354.36): C 81.34 H 5.40; found C 81.40, H
5.46.
trans-[Rh{h1-(C31H21P)}2Cl(CO)] (23): 1H NMR (400 MHz, CDCl3): d ˆ
7.40 ± 7.52 (m, 18H), 7.71 (d, J ˆ 8.4 Hz, 4H), 7.79 (t, J ˆ 6.8 Hz, 8H), 8.08
(d, J ˆ 8.4 Hz, 4H), 8.34 ± 8.40 (m, 8H); ); 13C NMR (100.61 MHz, CDCl3):
d ˆ 126.9, 127.0, 128.0, 128.06 (2C), 128.12 (2C), 128.6, 128.7, 128.8, 129.5,
133.5 (d, J ˆ 19 Hz), 137.2, 138.7 (pt, J ˆ 6.5 Hz), 141.6, 142.7 (t, J ˆ
11.8 Hz), 162.6 (m), 181.4 (dt, 1J(C,Rh) ˆ 65.1 Hz, 2J(C,P) ˆ 22.0 Hz,
CO); 31P NMR (81 MHz, CDCl3): d ˆ 166.8 (d, 1J(Rh,P) ˆ 174.2 Hz);
1
IR(CH2Cl2): nÄ ˆ 1998cm
.
trans-[Rh{h1-(C24H19OP)}2Cl(CO)] (24): 1H NMR (400 MHz, CDCl3): d ˆ
3.83 (s, 6H; H9), 7.00 (d, J ˆ 8.8 Hz, 4H), 7.34 ± 7.41 (m, 12H), 7.57 (d, J ˆ
8.8 Hz, 4H), 7.89 (d, J ˆ 7.0 Hz, 8H), 8.16 (t, J ˆ 10.2 Hz, 4H); ); 13C NMR
(100.61 MHz, CDCl3): d ˆ 55.4, 114.5, 128.3, 128.4, 128.8, 129.5 (pt, J ˆ
5.1 Hz), 133.7, 135.5, 140.8 (pt, J ˆ 6.9 Hz), 141.7 (pt, J ˆ 11.6 Hz), 159.9,
161.8 (dpt, J ˆ 14.4 Hz, 2.9 Hz), 180.9 (dt, 1J(Rh,C) ˆ 64.5 Hz, 2J(C,P) ˆ
21.6, CO); 31P NMR (81 MHz, CDCl3): d ˆ 166.1 (d, 1J(Rh,P) ˆ 174.0 Hz);
1
IR(CH2Cl2): nÄ ˆ 2002cm
.
trans-[Rh{h1-(C24H16F3P)}2Cl(CO)] (25): 1H NMR (200 MHz, CDCl3): d ˆ
7.15 ± 8.29 (m, all H); 13C NMR (100.61 MHz, CDCl3): d ˆ 122.7, 125.4,
126.0, 128.0, 128.4, 128.7, 129.4, 135.7, 136.2, 140.4, 144.7, 162.2 (pt, J ˆ
9.6 Hz, C1), 180.7 (dt, 1J(Rh,C) ˆ 65.1 Hz, 2J(C,P) ˆ 20.7 Hz, CO), 31P
2-tert-Butyl-4,6-diphenylphosphinine (15): A solution of 4,6-diphenyl-a-
pyrone 19[52] (497 mg, 2.0 mmol) and 2,2-dimethylpropylidinphosphane[45]
(220 mg, 2.2 mmol) in benzene (3 mL) was heated in a Schlenk pressure
vessel to 1408C for 3 d. After cooling the solution to RT, all volatile
components were removed in vacuo, and the residue was purified by flash
chromatography with petroleum ether/ethylacetate (19:1) to give phosphi-
NMR (81 MHz, CDCl3): d ˆ 173.8 (d, 1J(Rh,P) ˆ 175.0 Hz); IR(CH2Cl2):
1
nÄ ˆ 2003cm
.
Experiments to determine the nature of the catalytically active species:
Under an atmosphere of dry and oxygen-free nitrogen (glove box)
[Rh(COD)2OTf] (12.3 mg, 26.3 mmol) and phosphabenzene 8 (20 mg,
52.6 mmol) were dissolved in CD2Cl2 (1 mL) and transferred into a sapphire
NMR pressure vessel. A pressure of 40 bar CO/H2 (1:1) was applied, and
the sample was heated for 1 h at 908C. After cooling to RT, the sample thus
obtained was investigated by 1H and 31P NMR spectroscopy. 1H NMR
(400 MHz, 258C, CD2Cl2): A RhH signal could not be detected; 31P NMR
(161 MHz, CD2Cl2): d ˆ 154.5 (brs). 908C: 31P NMR (161 MHz, CD2Cl2):
d ˆ 146.75 (d, 1J(P,Rh) ˆ 140 Hz, major species), 156.63 (d, 1J(P,Rh) ˆ
149 Hz, minor species), ratio major/minor: about 6:1.
1
nine 15 as a yellow oil. Yield: 330 mg, 54%; H NMR (300 MHz, CDCl3):
d ˆ 1.46 (d, 4J(H,P) ˆ 1.5 Hz, 9H; C(CH3)3), 7.10 ± 7.28 (m, 6H), 7.48 (m,
2H), 7.68 (m, 2H), 8.03 (dd, J ˆ 5.8 Hz, 1.1 Hz, 1H), 8.11 (dd, J ˆ 6.4 Hz,
1.1 Hz, 1H); 13C NMR (75.47 MHz, CDCl3): d ˆ 32.7 (d, 3J(C,P) ˆ 12.2 Hz,
C(CH3)3), 38.7 (d, 2J(C,P) ˆ 21.1 Hz, C(CH3)3), 127.8 (2C), 128.6 (2C),
128.8 (2C), 129.4 (2C), 129.5 (d, 2J(C,P) ˆ 12.4 Hz), 130.1 (d, 2J(C,P) ˆ
12.4 Hz), 130.8, 131.0, 143.0 (d, 3J(C,P) ˆ 29.4 Hz), 143.4 (d, 3J(C,P) ˆ
14.0 Hz), 144.2, 170.6 (d, 1J(C,P) ˆ 52.0 Hz, C6), 185.3 (d, 1J(C,P) ˆ
59.3 Hz, C2); 31P NMR (81 MHz, CDCl3): d ˆ 187.8.
General procedure for the preparation of rhodium(i)-phosphinine com-
plexes: The corresponding phosphinine (300 mmol) was added slowly to a
solution of rhodium dicarbonyl chloride dimer (26.5 mg, 68.2 mmol) in
CH2Cl2 (2 mL). A vigorous gas evolution was observed. After stirring the
mixture magnetically for a further hour at RT, the solvent was removed in
vacuo to give the rhodium complexes in quantitative yields.
trans-[Rh{h1-(C23H17P)}2Cl(CO)] (22): Crystals suitable for X-ray analysis
were obtained from a concentrated solution of the complex in CH2Cl2 at
RT. 1H NMR (400 MHz, CDCl3): d ˆ 7.39 ± 7.48 (m, 14H), 7.53 (m, 4H), 7.67
(d, J ˆ 7.9 Hz, 4H), 7.94 (d, J ˆ 7.4 Hz, 8H), 8.25 (pseudo t, J ˆ 10.2 Hz,
4H); 13C NMR (100.61 MHz, CDCl3): d ˆ 127.5, 128.0, 128.1, 128.2, 128.9,
129.4 (pt, J ˆ 5.1 Hz), 135.7, 140.5 (pt, J ˆ 6.8 Hz), 141.1 (C4), 141.9 (pt, J ˆ
11.8 Hz, C3), 161.7 (ptd, J ˆ 11.8 Hz, 3.8 Hz), 181.9 (dt, 1J(C,Rh) ˆ 64.0 Hz,
Under an atmosphere of dry and oxygen-free nitrogen (glove box),
[Ir(CO)2acac] (9.1 mg, 26.3 mmol) and phosphabenzene
8 (20 mg,
52.6 mmol) were dissolved in CD2Cl2 (1 mL) and transferred into a sapphire
NMR pressure vessel. A pressure of 40 bar CO/H2 (1:1) was applied, and
the sample was heated for 1 h at 908C. After cooling to 908C the sample
thus obtained was investigated by 1H and 31P NMR spectroscopy. 1H NMR
(500 MHz, CD2Cl2): d ˆ 11.83 (d, 2J(H,P) ˆ 23.2 Hz, P-Ir-H); 31P NMR
(202.46 MHz, CD2Cl2): d ˆ 189.7 (free phosphinine), 144.3 (s), 143.9(s),
ratio: about 2:1:1.
General procedure for hydroformylation of oct-1-ene: [Rh(CO)2acac]
(3.1 mg, 12.0 mmol) and the appropriate amount of phosphinine (usually
240 mmol) were dissolved in toluene (5 mL) and stirred magnetically for
30 min. Subsequently, oct-1-ene (2.86 g, 25 mmol) was added, and the
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3119 $ 17.50+.50/0
3119

B. Breit, T. Mackewitz et al.
FULL PAPER
resulting solution was transferred into the autoclave. Additional toluene
(5 mL) was added, and the autoclave was heated to 908C while the mixture
was stirred (1000 min 1). When the mixture reached this temperature, a
CO/H2 (1:1) gas pressure of 10 bar was applied. After 20 h, complete
consumption of the starting material was determined by GC analysis and
the concentrations (c0) of all ingredients were quantified. The gas pressure
was released, followed by immediate and rapid addition of oct-1-ene
(5.71 g, 50 mmol). The kinetic experiment was started immediately by
readjustment of the gas pressure (CO/H2). Samples were taken from the
autoclave through a sample valve at the times indicated and analyzed
immediately by GC. Substraction of the initial concentration (c0) provided
the kinetic data of the hydroformylation experiment.
nine 8 (0.091 g, 0.024 mmol), and isobutene (3.61 g, 50 mmol) in toluene
(15.0 mL) to react at 908C and 20 bar CO/H2 (1:1) for 120 min according to
variant A. Product yield: 3-methyl-g-lactol ˆ 94%; aldehyde selectivity ˆ
99%; regioselectivity ˆ 99%.
b) Rhodium/triphenylphosphine catalyst: a methallylic alcohol conversion
of 88% was obtained by reacting [Rh(CO)2acac] (3.1 mg, 0.012 mmol),
triphenylphosphane (0.063 g, 0.024 mmol) and isobutene (3.61 g, 50 mmol)
in toluene (15.0 mL) at 908C and 20 bar CO/H2 (1:1) for 300 min according
to variant A. Product yield: 3-methyl-g-lactol ˆ 88%; aldehyde
selectivity ˆ 99%; regioselectivity ˆ 99%).
Hydroformylation of ( )-a-pinene: an a-pinene conversion of 13% was
obtained from reacting [Rh(CO)2acac] (4.9 mg, 0.019 mmol), phosphinine
8 (0.144 g, 0.38 mmol) and ( )-a-pinene (17.9 g, 132 mmol) in toluene
(22 mL) at 808C and 60 bar CO/H2 (1:1) for 16 h according to variant A.
Aldehyde selectivity ˆ 99%; regio- and diastereoselectivity ˆ 90%. The
product was purified through distillation and subsequent flash chromatog-
raphy on silica with petroleum ether/ethyl acetate (19:1) to furnish ( )-
endo-3-formylpinane as a colorless oil; yield: 2.085 g, 9.5%. Spectroscopic
and analytical data correspond to those reported previously.[55]
Hydroformylation of cyclohexene: [Rh(CO)2acac] (3.1 mg, 12.0 mmol) and
the corresponding ligand (20 equiv, 240 mmol) were dissolved in toluene
and stirred for 30 min at room temperature. Cyclohexene (4.10 g, 50 mmol)
was added, then the reaction mixture was transferred into the autoclave
and heated to 908C within 15 min. When the mixture reached this
temperature, a syngas pressure of 40 bar was applied. The reaction mixture
was analyzed by GC with internal standard and correction factors.
General procedures for the hydroformylation of more highly substituted
olefins: Variant A: [Rh(CO)2acac] and the ligand were dissolved in toluene,
and the mixture transferred into the autoclave. A pressure of 3 ± 5 bar CO/
H2 (1:1) was applied, and the autoclave was heated within 30 min to the
corresponding reaction temperature while the mixture was stirred
(1000 min 1). The olefin was added to the reaction mixture through a
pressure vessel, then the syngas pressure was adjusted to the desired
reaction pressure. After the indicated reaction time, the autoclave was
cooled to room temperature and depressurized, then the reaction mixture
was analyzed by GC with internal standard and correction factors.
Hydroformylation of 2,3-dimethylbut-2-ene
a) Rhodium/phosphinine catalyst: 2,3-dimethylbut-2-ene (25.3 g,
300.6 mmol) in toluene (25 g) was 29.2% converted by treating it with
[Rh(CO)2acac] (8 mg, 0.031 mmol) and phosphinine
0.69 mmol) at 1008C and 60 bar CO/H2 (1:1) for 24 h according to variant
A. Yield: 3,4-dimethylpentanal ˆ 29.2%; aldehyde selectivity ˆ 99%; re-
gioselectivity ˆ 99%.
b) Rhodium/triphenylphosphine catalyst: 2,3-dimethylbut-2-ene (24.1 g,
286.4 mmol) in toluene (25 g) was 0.5% converted by treating it with
8
(0.2628 g,
[Rh(CO)2acac] (7.2 mg, 0.028 mmol) and phosphinine
8 (161.8 mg,
Variant B: [Rh(CO)2acac] and the ligand were dissolved in toluene, and
transferred into the autoclave. A pressure of 3 ± 5 bar CO/H2 (1:1) was
applied, and the autoclave was heated within 30 min to the corresponding
reaction temperature while the mixture was stirred (1000 min 1). The
autoclave was cooled to room temperature and depressurized. The olefin
was added through a pressure vessel to the reaction mixture, which was
then heated to the reaction temperature within 30 min. When this
temperature was reached, the desired syngas pressure was applied. After
the indicated reaction time, the autoclave was cooled to room temperature
and depressurized, and the reaction mixture analyzed by GC with internal
standard and correction factors.
0.617 mmol) at 1008C and 60 bar CO/H2 (1:1) for 24 h according to variant
A. Yield: 3,4-dimethylpentanal ˆ 0.5%; aldehyde selectivity ˆ 99%; re-
gioselectivity ˆ 99%).
Hydroformylation of 3-methylbut-3-en-1-ol (isoprenol): isoprenol (6.2 g,
70 mmol) in toluene (6.0 g) was 73% converted by treating it with
[Rh(CO)2acac] (1.8 mg, 0.007 mmol) and phosphinine
8 (48.1 mg,
0.127 mmol) at 808C and 40 bar CO/H2 (1:1) for 24 h according to variant
A. Yield: 4-methyl-3,4,5,6-tetrahydro-2H-pyran-2-ol ˆ 73%; aldehyde
selectivity ˆ 99%; regioselectivity ˆ 99%).
Hydroformylation of oct-2-ene
Acknowledgments
a) Rhodium/phosphinine catalyst: Oct-2-ene (Z/E 80:20) (24.8 g,
221 mmol) in toluene (24.9 g) was completely converted by treating it
with [Rh(CO)2acac] (7.9 mg, 0.03 mmol) and phosphinine 8 (0.2334 g,
0.61 mmol) at 908C and 10 bar CO/H2 (1:1) for 4 h according to variant A.
Product yields: 3-/4-octene (1%), n-nonanal (24%), 2-methyloctanal
(43%), 2-ethylhepatanal (18%), and 2-propylhexanal (14%). Aldehyde
selectivity (99%).
This work was supported by the Fonds der Chemischen Industrie and the
Alfried Krupp Award for young university teachers of the Krupp
Foundation, as well as by BASF AG.
[1] K. Weissermel, H.-J. Arpe, Industrielle Organische Chemie, VCH,
Weinheim, 1988, pp. 133 ± 148.
[2] B. Breit, W. Seiche, Synthesis 2001, 1 ± 36.
[3] D. Evans, J. A. Osborn, G. Wilkinson, J. Chem. Soc. A 1968, 3133 ±
3142.
b) Rhodium/triphenylphosphine catalyst: Oct-2-ene in toluene (12.0 g,
108 mmol) was 69% converted by treating it with [Rh(CO)2acac] (3.9 mg,
0.015 mmol) and triphenylphosphane (0.075 g, 0.29 mmol) at 908C and
10 bar CO/H2 (1:1) for 4 h according to variant A. Product yields: 3-/4-
octene (5%), n-nonanal (3%), 2-methyloctanal (42%), 2-ethylhepatanal
(17%), and 2-propylhexanal (2%). Aldehyde selectivity (95%).
[4] Recent reviews: a) J. A. Moulijn, P. W. N. M. van Leeuwen, R. A.
van Santen, Studies in Surface Science and Catalysis, Vol. 79: Catal-
ysisÐAn Integrated Approach to Homogeneous, Heterogeneous and
Industrial Catalysis, Elsevier, Amsterdam, 1993; b) C. D. Frohning,
C. W. Kohlpaintner in Applied Homogeneous Catalysis with Organo-
metallic Compounds, (Eds.: B. Cornils, W. A. Herrmann), VCH,
Weinheim, 1996, Chapter 2, pp. 29 ± 104; c) M. Beller, B. Cornils,
C. D. Frohning, C. W. Kohlpaintner, J. Mol. Cat. A 1995, 104, 17 ± 85.
[5] C. P. Casey, G. T. Whiteker, Isr. J. Chem. 1990, 30, 299 ± 304.
[6] B. Breit, Chem. Eur. J. 2000, 6, 1519 ± 1524.
[7] a) S. D. Burke, J. E. Cobb, Teterahedron Lett. 1986, 27, 4237 ± 4240;
b) W. R. Jackson, P. Perlmutter, E. E. Tasdelen, J. Chem. Soc. Chem.
Commun. 1990, 763 ± 764.
[8] a) B. Breit, Angew. Chem. 1996, 108, 3021 ± 3023; Angew. Chem. Int.
Ed. Engl. 1996, 35, 2835 ± 2837; b) B. Breit, Liebigs Ann. 1997, 1841 ±
1851; B. Breit, M. Dauber, K. Harms, Chem. Eur. J. 1999, 5, 2819 ±
2827; c) B. Breit, J. Chem. Soc. Chem. Commun. 1997, 591 ± 592; d) B.
Hydroformylation of isobutene
a) Rhodium/phosphinine catalyst: Isobutene (15.7 g, 280 mmol) in toluene
(15.0 g) was 85% converted by treating it with [Rh(CO)2acac] (5 mg,
0.019 mmol) and phosphinine 8 (0.1326 g, 0.38 mmol) at 808C and 30 bar
CO/H2 (1:1) for 4 h according to variant B. Product yield: Isovaleraldehyde
85%; aldehyde selectivity ˆ 99%; regioselectivity ˆ 99%.
b) Rhodium/triphenylphosphine catalyst: isobutene (15.7 g, 280 mmol) in
toluene (15.0 g) was 9% converted by treating it with [Rh(CO)2acac]
(5 mg, 0.019 mmol) and triphenylphosphane (0.099 g, 0.38 mmol) at 808C
and 30 bar CO/H2 (1:1) for 4 h according to variant B. Product yield:
isovaleraldehyde ˆ 8%.
Hydroformylation of methallyl alcohol
a) Rhodium/phosphinine catalyst: A methallyl alcohol conversion of 94%
was obtained by allowing [Rh(CO)2acac] (3.1 mg, 0.012 mmol), phosphi-
3120
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3120 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 14

Rhodium-Catalyzed Hydroformylation
3106±3121
Breit, Eur. J. Org. Chem. 1998, 1123 ± 1134; e) B. Breit, Tetrahedron
Lett. 1998, 39, 5163 ± 5166; f) B. Breit, S. K. Zahn, Angew. Chem. 1999,
111, 1022 ± 1024; Angew. Chem. Int. Ed. 1999, 38, 969 ± 971.
[29] Compare [Rh(CO)2acac]: For the CO ligand 1J(C,Rh) ˆ 73 Hz; see
ref. [28] p. 550; for another example see ref. [8d].
[30] K. G. Moloy, J. L. Petersen, J. Am. Chem. Soc. 1995, 117, 7696 ± 7710.
Â
Â
[9] a) K. Nozaki, I. Ojima in Catalytic Asymmetric Synthesis 2nd ed. (Ed.:
I. Ojima), Wiley-VCH, New York, 2000, Chapter 7, pp. 429 ± 463; b) F.
Agbossou, J.-F. Carpentier, A. Mortreux, Chem. Rev. 1995, 95, 2485 ±
[31] E. Fernandez, A. Ruiz, C. Claver, S. Castillon, A. Polo, J. F. Piniella, A.
Alvarez-Larena, Organometallics 1998, 17, 2857 ± 2864.
[32] a) M. Li Wu, M. J. Desmond, R. S. Drago, Inorg. Chem. 1979, 18,
679 ± 686; b) K. R. Dunbar, S. C. Haefner, Inorg. Chem. 1992, 31,
3676 ± 3679; c) A. Ceriotti, G. Ciani, A. Sironi, J. Organomet. Chem.
1983, 247, 345 ± 350; d) P. A. Chaloner, C. Claver, P. B. Hitchcock,
A. M. Masdue, A. Ruiz, Acta Crystallogr. 1991, C47, 1307 ± 1308.
[33] C. A. Tolman, Chem. Rev. 1977, 77, 313 ± 348.
Â
2506; c) S. Gladiali, J. C. Bayon, C. Claver, Tetrahedron: Asymmetry
1995, 6, 1453 ± 1474.
Â
[10] B. Heil, L. Marko, Chem. Ber. 1969, 102, 2238 ± 2240.
[11] a) R. L. Pruett, J. A. Smith, J. Org. Chem. 1969, 34, 327 ± 330;
b) P. W. N. M. van Leeuwen, C. F. Roobeek, J. Organomet. Chem.
1983, 258, 343 ± 350.
[12] T. Jongsma, G. Challa, P. W. N. M. van Leeuwen, J. Organomet. Chem.
1991, 421, 121 ± 128.
[34] From PM3 calculations for complex 26 (M ˆ Rh) with the PCSpartan
software package.
[35] a) Y. Ohgomori, S. Yoshida, Y. Watanabe, J. Chem. Soc. Dalton Trans.
1987, 2969 ± 2974; b) D. de Montauzon, R. Poilblanc, J. Organomet.
Chem. 1975, 93, 397 ± 404; c) T. Sakakura, T. Sodeyama, K. Sasaki, K.
Wada, M. Tanaka, J. Am. Chem. Soc. 1990, 112, 7221 ± 7229.
[36] a) D. Evans, J. A. Osborn, G. Wilkinson, Inorg. Synth. 1990, 28, 79 ±
80; b) S. Franks, F. R. Hartley, Inorg. Chim. Acta 1981, 47, 235 ± 248;
c) G. M. Intille, Inorg. Chem. 1972, 11, 695 ± 702; d) A. J. Deeming,
B. L. Shaw, J. Chem. Soc. A 1969, 597 ± 602; e) J. A. McCleverty, G.
Wilkinson, Inorg. Synth. 1966, 8, 214 ± 217.
[13] D. Selent, K.-D. Wiese, D. Röttger, A. Börner, Angew. Chem. 2000,
112, 1694 ± 1696; Angew. Chem. Int. Ed. 2000, 39, 1639 ± 1641.
Â
[14] a) A. M. Trzeciak, J. J. Ziolkowski, Coord. Chem. Rev. 1999, 190 ± 192,
883 ± 900; b) L. A. van der Veen, P. C. J. Kamer, P. W. N. M. van Leeu-
wen, Angew. Chem. 1999, 111, 349 ± 351; Angew. Chem. Int. Ed. 1999,
38, 336 ± 338.
[15] a) G. Märkl in Multiple Bonds and Low Coordination in Phosphorus
Chemistry (Eds.: M. Regitz, O. J. Scherer) Thieme, 1990, pp. 220 ± 257;
b) F. Mathey, Rev. Heteroat. Chem. 1992, 6, 1 ± 24; c) G. Märkl, Angew.
Chem. 1966, 78, 907 ± 908; Angew. Chem. Int. Ed. Engl. 1966, 5, 846 ±
847.
[37] R. P. Hughes in Comprehensive Organometallic Chemistry (Eds.: G.
Wilkinson, F. G. A. Stone, E. W. Abel), Pergamon, Oxford, 1982,
Chapter 35, pp. 296ff.
[16] For the use of h6-phosphabenzene iron complexes as catalysts for
pyridine formation see: F. Knoch, F. Kremer, U. Schmidt, U. Zenneck,
P. Le Floch, F. Mathey, Organometallics 1996, 15, 2713 ± 2719; see also
ref. [19].
[38] J. M. Brown, A. G. Kent, J. Chem. Soc. Perkin Trans. 2 1987, 1597 ±
1607.
[39] M. Kranenburg, Y. E. M. van der Burgt, P. C. J. Kamer, P. W. N. M.
van Leeuwen, Organometallics 1995, 14, 3081 ± 3089.
[17] a) C. Batich, E. Heilbronner, V. Hornung, A. J. Ashe III, D. T. Clark,
U. T. Cobley, D. Kilcast, I. Scanlan, J. Am. Chem. Soc. 1973, 95, 928 ±
930; b) A. J. Ashe III, F. Burger, M. Y. El-Sheik, E. Heilbronner, J. P.
Maier, J.-F. Muller, Helv. Chim. Acta 1976, 59, 1944 ± 1948.
[18] a) W. J. Hehre, R. Ditchfield, R. F. Stewart, J. A. Pople, J. Chem. Phys.
1970, 52, 2769 ± 2773; b) G. Frison, A. Sevin, N. Avarvari, F. Mathey, P.
Le Floch, J. Org. Chem. 1999, 64, 5524 ± 5529.
[40] D. T. Brown, T. Eguchi, B. T. Heaton, J. A. Iggo, R. Whyman, J. Chem.
Soc. Dalton Trans. 1991, 677 ± 683.
[41] Tested for ligands 3 and 8.
[42] A. van Rooy, E. N. Orij, P. C. J. Kamer, P. W. N. M. van Leeuwen,
Organometallics 1995, 14, 34 ± 43.
[43] W. F. Gresham, R. E. Brooks, (DuPont de Nemours & Co.) US
2.497.303, 1950; [Chem. Abstr. 1950, 44, 4492e].
[19] a) B. Breit, J. Chem. Soc. Chem. Commun. 1996, 2071 ± 2072; b) B.
Breit, R. Winde, K. Harms, J. Chem. Soc. Perkin Trans. 1 1997, 2681 ±
2682; c) B. Breit, R. Paciello, B. Geissler, M. Röper (BASF AG) DE
19.621.967, 1997 (EP 906261, WO 97/46507); [Chem. Abstr. 1998, 128,
129.496]; R. Paciello, T. Mackewitz, M. Röper, B. Breit (BASF AG)
DE 19.911.920, 2000 (WO 00/55164); [Chem. Abstr. 2000, 133,
238.120.
[44] Organikum, 19th ed., Deutscher Verlag der Wissenschaften, Leipzig,
1993, p. 469.
[45] G. Becker, H. Schmidt, G. Uhl, W. Uhl, M. Regitz, W. Rösch, U.-J.
Vogelbacher, Inorg. Synth. 1990, 27, 243 ± 253.
[46] R. E. Lutz, T. A. Martin, J. F. Codington, T. M. Amacker, R. K.
Allison, N. H. Leake, R. J. Rowlett, J. D. Smith, J. W. Wilson III, J.
Org. Chem. 1949, 14, 982 ± 1000.
[20] G. Märkl, F. Lieb, A. Merz, Angew. Chem. 1967, 79, 947 ± 948; Angew.
Chem. Int. Ed. Engl. 1967, 6, 944 ± 945.
[21] R. Paciello, E. Zeller, B. Breit, M. Röper (BASF AG) DE 19.743.197,
1999 (EP 1025111, WO 99/16774); [Chem. Abstr. 1999, 130, 237700u];
T. Mackewitz, M. Röper (BASF AG) DE 19.911.922, 2000 (EP
1036796); [Chem. Abstr. 2000, 133, 239110].
[22] K. Dimroth, C. Reichardt, K. Vogel, Org. Synth. Coll. Vol. V, 1135 ±
1137.
[23] A. Garcia-Raso, J. Garcia-Raso, B. Campaner, R. Mestres, J. V.
Sinisterra, Synthesis 1982, 1037 ± 1041.
[24] P. F. G. Praill, A. L. Whitear, J. Chem. Soc. 1961, 3573 ± 3579.
[25] W. Rösch, M. Regitz, Z. Naturforsch. 1986, 41b, 931 ± 933.
[26] S. Voû, Ph.D. thesis, Universität Marburg (Germany), 1998.
[27] S. Berger, S. Braun, H.-O. Kalinowski, NMR-Spektroskopie von
Nichtmetallen, Vol. 3, Thieme, Stuttgart, 1993, pp. 164 ± 166.
[28] Compare [Rh4(CO)12], 13C NMR (CH2Cl2): d ˆ 190.3; see: H.-O.
Kalinowski, S. Berger, S. Braun, 13C-NMR-Spektroskopie, Thieme,
Stuttgart, 1984, p. 383.
[47] C. Reichardt, R. Müller, Liebigs Ann. Chem. 1976, 1937 ± 1952.
[48] N. A. Portnoy, C. J. Morrow, M. S. Chattha, J. C. Williams, A. M.
Aguiar, Tetrahedron Lett. 1971, 1401 ± 1402.
[49] J. A. Durden, D. G. Crosby, J. Org. Chem. 1965, 30, 1684 ± 1687.
[50] F. Lieb, Ph.D. thesis, Universität Würzburg (Germany), 1970.
[51] G. Märkl, F. Lieb, A. Merz, Angew. Chem. 1967, 79, 475; Angew.
Chem. Int. Ed. Engl. 1967, 6, 458 ± 459.
[52] A. R. Katritzky, J. Arrowsmith, Z. bin Bahari, C. Jayaram, T. Siddiqui,
S. Vassilatos, J. Chem. Soc. Perkin Trans. 1 1980, 2851 ± 2855.
[53] G. M. Sheldrick, SHELXS-96, Program for solving crystal structures,
University of Göttingen, 1997.
[54] G. M. Sheldrick, SHELXL-97, Program for refining crystal structures,
University of Göttingen, 1997.
[55] F. Langer, K. Püntener, R. Stürmer, P. Knochel, Tetrahedron:
Asymmetry 1997, 8, 715 ± 738.
Received: January 22, 2001 [F3017]
Chem. Eur. J. 2001, 7, No. 14
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0714-3121 $ 17.50+.50/0
3121