Diastereoisomeric bisphosphite ligands in the hydroformylation of octenes: Rhodium catalysis and HP-NMR investigations

Article abstract of DOI:10.1039/b814120g

Diastereoisomeric hydroformylation catalysts show differences for the catalyst preformation pathway and a strongly reduced n-octene hydroformylation activity for the (S,S,R)-isomer. The Royal Society of Chemistry.

Full text of DOI:10.1039/b814120g

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Diastereoisomeric bisphosphite ligands in the hydroformylation of
octenes: rhodium catalysis and HP-NMR investigationsw
a
a
b
ac
¨
Detlef Selent,* Wolfgang Baumann, Klaus-Diether Wiese and Armin Borner*
Received (in Cambridge, UK) 13th August 2008, Accepted 20th October 2008
First published as an Advance Article on the web 31st October 2008
DOI: 10.1039/b814120g
Diastereoisomeric hydroformylation catalysts show differences
for the catalyst preformation pathway and a strongly reduced n-
octene hydroformylation activity for the (S,S,R)-isomer.
the overall steric demand can vary. The transition metal ligand
complex, when travelling through the catalytic cycle, experiences
structural changes with the ligand adopting energetically favour-
able conformations, and configurations. Consequently, also the
appropriate set of ligand cone angle profiles would be informa-
tive to describe steric demands.6 In a parallel paper we show that
the ratio of diastereomeric ligands bearing a stereogenic P-center
and a rotationally flexible biaryl axis can be affected by achiral
counter ligands in the precatalyst or resting state (e.g. anionic
ligands as are allyl, acetylacetonate, hydride).7 A further change
of this ratio during the formation of catalytic intermediates
discussed as are alkyl and acyl complexes, is likely to be expected.
Herein we will give evidence that regioselectivity in 1-octene
hydroformylation does not necessarily change due to applica-
tion of a distinct diastereomeric bisphosphite (1) based on 2,20-
dihydroxybinaphthyl. However, significant ligand-dependent
differences are observed for precatalyst formation and catalyst
activity, respectively. A pronounced change in regioselectivity is
noted for internal olefins applied as substrates, where kinetic
control of consecutive isomerization-hydroformylation reaction
is one of the requirements for predominant n-regioselectivity.
1-(R,R,R)- and 1-(S,R,S) (Scheme 1) were synthesized by
employment of enantiopure starting material according to lit-
erature.8 1-(S,S,R) was obtained by subsequent reaction of (S)-
2,20-dihydroxybinaphthyl with the BINOL-phosphochloridites
of (S)-2,20-dihydroxybinaphthyl and (R)-2,20-dihydroxy-
binaphthyl in the presence of NEt3. Prior to the final esterification
step, it was necessary to purify intermediate 20-((S)-di-
naphtho[2,1-d:10,20-f][1,3,2]dioxaphosphepin-4-yloxy)-1,10-(S)-
binaphthyl-2-ol by column chromatography. 31P NMR
spectroscopy proved the presence of configurationally
different phosphorus atoms absorbing at 144.8 (d) and
146.5 (d) ppm exhibiting a 20.8 Hz ‘through space’ coupling,
and the presence of 5% of 1-(S,S,S) as an impurity.9
Rhodium-catalyzed hydroformylation of olefins affording alde-
hydes, is one of the most important homogeneously catalyzed
reactions.1 It is used for the production of commodities as well as
for the synthesis of fine chemicals. In dependence on the catalyst
used either linear or branched aldehydes are formed predomi-
nantly. Regioselectivity can be adjusted with the assistance of
ancillary ligands, mainly trivalent phosphorus compounds. In
most cases branched aldehydes are chiral, therefore the goal of
the asymmetric hydroformylation is the achievement of high iso-
regioselectivity and stereoselectivity.2 In contrast, only high
regioselectivity counts for the production of n-aldehydes. Also
for n-regioselective reactions, frequently ligands have been em-
ployed which possess elements of chirality. Of particular impor-
tance are biaryl-based bidentate ligands, which have seen
widespread application. Without appropriate substitution, free
rotation around the biaryl axis is not restricted under reaction
conditions and atropo isomers can not be differentiated. How-
ever, the incorporation of additional chirality elements in the
molecule, like stereogenic P-atoms, may lead to diastereomers,
which are configurationally stable.3 The performance of diaster-
eomeric catalysts is hardly to be predicted, though they are
expected to influence differently the outcome of the hydro-
formylation reaction. Successful concepts describing ligand proper-
ties include the bite angle for the achiral Xanthphos ligand type,
with shows correlation with regioselectivity in rhodium catalyzed
hydroformylation.4 Based on MM3 calculations, recently Briggs
and Whiteker suggested that coordination of rotationally hin-
dered diastereomeric biphenyldiphosphites of the Biphephos
ligand type to the Rh-center may result in different, configura-
tion-dependent structures of the catalysts resting state
[HRh(CO)2(bisphosphite)].5 Differences in both, bite angles
and probably more important the cone angles at each phos-
phorus atom were discussed as a reason for variations in both,
n-regioselectivity and catalytic activity found in the hydroformy-
lation of the standard olefin 1-octene. For bidentate ligands
exhibiting flexible backbones and a tendency for epimerization,
To study the impact of a distinct ligand configuration on the
electron donating properties of compounds 1, diselenides were
prepared.10 Reaction with six equivalents of selenium in boiling
toluene gave quantitative conversion to 1-Se2 within 17 h.
Interestingly, under these conditions, a selenium mediated
a Leibniz-Institut fur Katalyse an der Universitat Rostock e.V.,
A.-Einstein-Str., 29a, 18059 Rostock, Germany.
E-mail: detlef.selent@catalysis.de.
¨
¨
E-mail: armin.boerner@catalysis.de; Fax: +49381128151169
b Evonik Oxeno GmbH, Paul-Baumann-Str., 1, 45 772 Marl, Germany
c Institut fur Chemie der Universitat Rostock, A.-Einstein-Str., 3,
¨
¨
18059 Rostock, Germany
w Electronic supplementary information (ESI) available: Experimental
section and details on PM3 calculations. See DOI: 10.1039/b814120g
Scheme 1 Diastereomeric binaphthol based bisphosphites.
Chem. Commun., 2008, 6203–6205 | 6203
ꢀc
This journal is The Royal Society of Chemistry 2008

View Online
transesterification occurred for 1-(S,R,S) resulting in 39% of 1-
(S,S,R)Se2 in the final product.11 Chemical shifts as well as
31P–77Se couplings obtained from 31P NMR showed only minor
differences (Scheme 1). Thus, for the diselenides a comparable
donor behaviour of phosphorus is proved for the isomers of 1.
In preliminary trials a mixture of diastereomeric ligands 1 as
obtained from racemic 2,20-dihydroxybinaphthyl, consisting
of 55.9% (R,R,R)-, 38.7% (S,S,R)- and 5.4% (S,R,S)-diphos-
phite isomer after workup, was tested in the hydroformyla-
tion of 1-octene. Aldehyde yields ranging from 84 to 89%,
containing a 0.959 n-nonanal molar fraction (ratio 1 : Rh = 5)
were determined after reaction at 100 1C/50 bar (see Table 1).
For a technical mixture of less reactive internal octenes applied
as an substrate at 120 1C/20 bar, the ligand : rhodium ratio
showed a remarkable effect with almost complete inhibition of
hydroformylation reaction at L : Rh = 5 : 1.
dominating in modifying catalytic activity, if the diastereo-
meric mixture of diphosphites is applied. Depending on the
substrates used, yield-based performance can be evaluated as
follows: 1-octene, 1-(R,R,R) B 1-(S,R,S) 4 1-(S,S,R); inter-
nal octenes, 1-(R,R,R) 4 1-(S,R,S) B 1-(S,S,R).
It has to be noted that regioselectivities for n-nonanal obtained
from 1-octene differ only within a range of 1.8% for the three
diastereomers. This may point toward similar geometries of the
catalysts formed. Indeed, PM3 semi-empirical calculations per-
formed for the catalyst’s resting state [HRh(CO)2(1)] gave
relatively small deviations for the P–Rh–P bite angle only,
ranging from 118.6 to 121.81 for the energetically favoured
bisequatorial diphosphite coordination identified for all the three
diastereomeric complexes (ESIw)12 In order to get more informa-
tion on the catalytically active species formed from ligands 1,
hydride complex formation under an atmosphere of syngas was
studied by means of HP-NMR spectroscopy in a modified
sapphire cell equipped with a gas circulation device ensuring full
control of pressure and gas saturation.7 Already at the stage of
dissolving precursors under argon for sample preparation re-
markable differences in reactivity were found.
Further experiments verified the contribution of each dia-
stereomer. In 1-octene hydroformylation, 1-(R,R,R) and 1-
(S,R,S) showed almost identical modifying properties to the
catalyst. The (S,S,R)-diphosphite gave poor yields of 11–17%
at L/Rh = 5, but a convenient result with less equivalents of
ligand applied. For L/Rh = 2, 89% of aldehyde yield were
paired with a regioselectivity of 95.6%. The regioselectivities
obtained prove only a minor contribution of unmodified
catalysis, if present at all. From the internal octene mixture,
only traces of aldehyde again are found with the (S,S,R)-
ligand applied at a concentration of 3.75 ꢁ 10ꢂ3 mol lꢂ1
(ligand/Rh = 5). Interestingly, 1-(S,R,S) gave a low yield, too.
The catalyst derived from 1-(R,R,R) showed superior perfor-
mance. This is surprising, because results from 1-octene
hydroformylation did not reveal enhanced capability for olefin
isomerization. From this substrate containing 3.3% 1-octene
regioselectivity reached a value of 71.9% at L/Rh = 2. For all
catalysts, incomplete precursor to catalyst transformation can
be excluded as a factor altering results. Because batches
performed with a prolonged catalyst preformation time of
4 h gave almost identical catalytic performance, the routine
heating methodology (30 min for 100 and 120 1C) seems to be
sufficient. From the results it is obvious that 1-(S,S,R) is
If one equivalent of 1-(S,R,S) was reacted with [acacRh-
(COD-1.5)] under argon in toluene-d8, the 31P NMR spectrum
obtained at room temperature showed that the diphosphite is
not able to replace entirely the diolefin from the metal centre.
The spectrum is characterized by a broadened doublet at
144.6 ppm, (JPRh = 301 Hz) for [acacRh(1-(S,R,S))] covering
60% of overall signal intensity, accompanied by the signal of
free ligand at 146.1 ppm. However, when a stream of syngas
was passed through the solution at 20 bar pressure, CO
accelerated the substitution of diolefin and a pure complex
formed immediately. Spectroscopic data obtained for the
rhodium compound observed did not change, compared to
that measured under argon, therefore the intermediate forma-
tion of [acacRh(CO)(1-(S,R,S))] is unlikely. Already after 10
min of syngas treatment at 25 1C, [HRh(CO)2(1-(S,R,S))]
is detectable. After increasing the temperature to 70 1C,
complete precursor to rhodium hydride transformation was
achieved within 3 h without the reaction system passing
through a detectable intermediate. Only one phosphorus
signal is observed for the complex which is centered at
167.5 ppm (d, JPRh = 237.5 Hz). The hydride signal is found
at ꢂ10.57 ppm (d, JHRh = 2.9 Hz). The hydride complex
solution remained unchanged for several hours after cooling to
room temperature and depressurization to one bar of
CO/H2. This allowed IR spectroscopic measurement at the
same gas-saturated sample, resulting in stretching frequencies
n(CO) = 2020, 2070 cmꢂ1. Further details, as is the absence of
detectable H–P coupling, verified a bisequatorial coordination
of the bisphosphite.13
Table 1 Hydroformylation of 1-octenea and isomeric n-octenesb
1-Octene
n-Octenes
L/Rh = 5
Yield/
L/Rh = 2
Yield/
L/Rh = 5
Yield/
L/Rh = 2
Yield/
n-nonanal
(%)
n-nonanal
(%)
n-nonanal
(%)
n-nonanal
(%)
mixture
1-(R,R,R)
84/95.9
87/94.4
89/94.8
86/93.5
88/92.8c
92/94.8
2/n.d.
20/76.1
41/67.0
68/71.9
71/71.4c
34/63.4
1-(S,R,S)
1-(S,S,R)
87/94.8
87/94.9c
11/96.2
17/95.8c
4/85.4
4/86.1c
0.5/n.d.
0.3/n.d.c
Analogous measurements were performed with 1-(R,R,R).
The corresponding enantiomer, 1-(S,S,S), combined with
[acacRh(CO)2] and cationic rhodium precursors was already
subject to an HP-NMR investigation for asymmetric hydro-
formylation.14 Hydride complex formation was described to
take 16 h at 40 1C. With [acacRh(COD-1.5)], 1-(R,R,R)
under argon reacts promptly to [acacRh(1-(R,R,R))] (d =
140.0 ppm, JPRh = 309.7 Hz). The latter complex immediately
disappears upon syngas treatment of the solution. First, only
89/95.6
40/64.3
a
T = 100 1C, p = 50 bar, [Rh] = 0.3 ꢁ 10ꢂ3 mol lꢂ1, substrate/
b
catalyst = 5410, toluene, t = 4 h. 120 1C, 20 bar, [Rh] = 0.75 ꢁ
10ꢂ3 mol lꢂ1, substrate/catalyst = 2160), toluene, 4 h. A technical
substrate mixture was applied: 3.3% 1-octene; 48.4% Z/E-2-octene;
29.2% Z/E-3-octene; 16.4% Z/E-4-octene; 2.1% skeletal C8-olefinic
c
isomers; 0.6% n-octane. Additional catalyst preformation time ap-
plied at reaction temperature: 4 h.
ꢀc
This journal is The Royal Society of Chemistry 2008
6204 | Chem. Commun., 2008, 6203–6205

View Online
clean formation of a single hydride was not achieved. About
10% of signal intensity is calculated for a badly resolved,
virtual triplet of quintets shaped phosphorus signal at 176.6 ppm
corresponding to a broad hydride resonance at ꢂ11.1 ppm.
Thus, the formation of the two possible isomeric hydride
complexes with an ax,eq ligand coordination may be seen here
also. An investigation to the real nature of these hydrides and
to their contributions to catalysis is underway.
Fig. 1 31P NMR spectra of the reaction of [acacRh(COD-1.5)] with
1-(S,S,R) (0.1 mmol each) in toluene-d8 (2.2 ml). (a) 30 min after
mixing at 25 1C, argon. (b) 15 min after supplying 20 bar CO/H2 at
25 1C. (c) After cumulative heating at 52 1C, 30 min, and 74 1C, 1.5 h.
For (b) and (c), syngas flow was adjusted at 1 ml minꢂ1. Saturation of
solution with syngas was proved by following the signal intensity of
dissolved H2, d = 4.44 ppm.
In conclusion, the formation pathway and the composition
of the hydroformylation catalysts is dependent on the config-
uration of the diastereomeric ligand applied. The results from
catalysis with different substrates, especially internal octenes,
reveal that the bisphosphite isomers form intrinsically different
catalysts. These differences will be overseen if interpretation of
performance is based only on a linkage between the compar-
able regioselectivities obtained for 1-octene hydroformylation,
and spectroscopic data pointing toward similar geometries
around the rhodium centers in the hydride complexes.
We appreciate financial support from Evonik-Oxeno
GmbH. We thank Mrs. M. Geisendorf and Mrs. K. Romeike
for skilled technical assistance.
the low field shifted signal of [acacRh(CO)(1-(R,R,R))] at
143.5 ppm (JPRh = 273.7 Hz) is observed. Within 45 min an
additional set of multiplets in the range of 158 and 172 ppm
acquiring 10% of signal intensity, together with a broad
hydride signal at ꢂ10.65 ppm in the proton spectrum is
formed. We were not able to identify these latter compounds.
They are of intermediate character holding a maximum of
70% of the phosphorus signal intensity after 2.5 h at 25 1C. At
this stage, still 18% of the acetylacetonate educt is present.
Further cumulative heating at 40 1C/1.0 h, 55 1C/1.5 h and
80 1C/3 h brings the reaction to equilibrium with the forma-
tion of [HRh(CO)2(1-(R,R,R))], resonating at 167.0 ppm (JPRh
= 232.8 Hz). As expected, NMR spectroscopic data obtained
are almost identical to [HRh(CO)2(1-(S,S,S))], however,
for the latter complex a chemical shift for phosphorus of
162.0 ppm (JPRh = 233 Hz) was reported in the literature.
IR spectroscopy verified vibrations n(CO) at 2067, 2013, 2023
Notes and references
1 M. Beller, B. Cornils, C. D. Frohning and C. W. Kohlpaintner,
J. Mol. Catal. A, 1995, 104, 17–85; M. Beller and K. Kumar,
Transition Metals for Organic Synthesis, eds. M. Beller and
C. Bolm, Wiley-VCH, Weinheim, 2004, vols. 1 and 2, 29-56;
Rhodium Catalyzed Hydroformylation, eds. P. W. N. M. van
Leeuwen and C. Claver, Kluwer, Dordrecht, 2000.
2 F. Agbossou, J.-F. Carpentier and A. Mortreux, Chem. Rev., 1995,
95, 2485–2506; M. Dieguez, O. Pamies and C. Claver, Chem. Rev.,
´
2004, 104, 3189–3215.
3 This concept was introduced by Mikami et al. in asymmetric
catalysis under the term ‘‘asymmetric activation’’ see: e.g.
K. Mikami, K. Aikawa, Y. Yusa, J. J. Jodry and M. Yamanaka,
Synlett, 2002, 1561–1578.
4 L. A. Van der Veen, P. H. Keeven, G. C. Schoemaker, J. N.
H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen, M. Lutz and
A. L. Spek, Organometallics, 2000, 19, 872–883.
(split asymmetric band) and n(RhH) = 1991 cmꢂ1
.
Comparable to 1-(R,R,R) and in contrast to 1-(S,R,S) the
S,S,R-diastereomer at 25 1C under argon spontaneously re-
acted to give [acacRh(1-(S,S,R))]. The 31P NMR spectrum (see
Fig. 1(a)) is characterized by two double doublets at 133.4
(1JPRh = 307 Hz) and 152.0 (1JPRh = 311 Hz) ppm, each
2
exhibiting a JPP coupling constant of 111 Hz between the
5 J. R. Briggs and G. T. Whiteker, Chem. Commun., 2001, 2174–2175.
6 J. M. Smith, B. C. Taverner and N. J. Coville, J. Organomet.
Chem., 1997, 530, 131–140.
configurationally different phosphorus atoms. The minor
component seen at 140.0 ppm (1JPRh = 309 Hz) is assigned
to [acacRh(1-(S,S,S))] derived from the diastereomeric impur-
ity present in the ligand used. Addition of syngas (20 bar) at
25 1C resulted in the formation of [acacRh(CO)(1-(S,S,R))]
showing a multiplet centered at 144.1 ppm, accounting for
67% of signal intensity after 15 min (Fig. 1(b)). The formation
of intermediate hydride complexes is deduced from a signal set
in the range of 163 to 177 ppm, accompanied by a sharp signal
at ꢂ10.56 ppm (1JHRh = 4.1 Hz) and additional very broad
signals at ꢂ10.1 and ꢂ10.4 ppm in the proton NMR spectrum.
Further heating and holding the solution at 74 1C for 1.5 h led
to equilibrium of reaction, where [HRh(CO)2(1-(S,S,R))] is the
major component. This rhodium complex exhibits a hydride
signal at ꢂ10.37 ppm (1JHRh = 3.8 Hz, 2JHP not resolved). In
31P NMR it is characterized by two phosphorus signals at
169.4 ppm (dd, 1JPRh =240 Hz, 2JPP =254 Hz), and 164.6 ppm
7 D. Selent, K.-D. Wiese, A. Borner and W. Baumann, Chem.
Comm., to be submitted.
¨
8 M. J. Baker and P. G. Pringle, J. Chem. Soc., Chem. Commun., 1991,
1292–1293; M. Yan, L. W. Yang, K. Y. Wong and A. S. C. Chan,
J. Chem. Soc., Chem. Commun., 1999, 11–12; J. Scherer, G. Huttner,
M. Buchner and J. Bakos, J. Organomet. Chem., 1996, 520, 45–58.
9 A through space coupling is probable, because the phosphorus
atoms in sterically crowded 1-(S,S,R) are separated by seven
bonds. See also: R. Holmes, in Phosphorus—31 NMR Spectral
Properties in Compound Characterization and Structural Analysis,
eds. L. D. Quin and J. G. Verkade, VCH Publishers, New York,
1994, pp. 27–39.
10 P. N. Bungu and S. Otto, J. Organomet. Chem., 2007, 692,
3370–3379, and refs. therein.
11 For further details see ESIw.
12 PM3 calculations for geometry optimization were performed within
the WAVEFUNCTION Spartan06 software package. See ESIw.
13 A. van Rooy, P. C. J. Kamer, P. W. N. M. van Leeuwen, K. Goubitz,
J. Fraanje, N. Veldman and A. L. Spek, Organometallics, 1996, 15,
835–847; G. J. H. Buisman, L. A. van der Veen, A. Klootwijk, W. G.
J. de Lange, P. C. J. Kamer, P. W. N. M. van Leeuwen and D. Vogt,
1
2
(dd, JPRh =234 Hz, JPP
= 255 Hz). These data,
Organometallics, 1997, 16, 2929–2939; A. Castellanos-Pa
S. Castillon, C. Claver, P. W. N. M. van Leeuwen and W. G. J. de
Lange, Organometallics, 1998, 17, 2543–2552.
´
ez,
together with the observed IR stretching vibrations n(CO)
located at 2018, 2078 cmꢂ1 point toward a favoured bisequa-
torial coordination of the bisphosphite at the metal centre,
similar to the other diastereomers. It must be noted that a
´
14 P. Uriz, E. Fernandez, N. Ruiz and C. Claver, Inorg. Chem.
Commun., 2000, 3, 515–519.
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 6203–6205 | 6205