Ligand effects in rhodium-catalyzed hydroformylation with bisphosphines: Steric or electronic?

Article abstract of DOI:10.1039/c6cy01990k

Twelve commercially available bisphosphine ligands have been evaluated in rhodium-catalyzed hydroformylation reactions. All ligands exhibited high chemoselectivities for aldehyde formation. The highest enantioselectivity (53% ee) of styrene hydroformylation was achieved with (S)-BTFM-Garphos (L7) substituted with electron withdrawing substituents. High pressure NMR (HP-NMR) spectroscopy and in situ high pressure IR spectroscopy (HP-IR) were used to study the resting states of the catalyst species in the reactions. The ligand effect on the structures of the observable species was examined. Both electronic and steric factors were considered to contribute to the performance of the various ligands. The results showed that decreasing the phosphine basicity increased the enantioselectivity, while in the systems studied here the steric character plays a less important role than the electronic features in achieving good regioselectivities.

Full text of DOI:10.1039/c6cy01990k

View Article Online
View Journal
Catalysis
Science &
Technology
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: Y. Jiao, M. S.
Torne, J. Gracia, J. W. Niemantsverdriet and P. W.N.M. van Leeuwen, Catal. Sci. Technol., 2017, DOI:
10.1039/C6CY01990K.
This is an Accepted Manuscript, which has been through the
Volume 6 Number 1 7 January 2016 Pages 1–308
Royal Society of Chemistry peer review process and has been
accepted for publication.
Catalysis
Science &
Technology
Accepted Manuscripts are published online shortly after
www.rsc.org/catalysis
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
ISSN 2044-4753
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
PAPER
Qingzhu Zhang et al.
Catalytic mechanism of C–F bond cleavage: insights from QM/MM
analysis of fluoroacetate dehalogenase
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
rsc.li/catalysis

Page 1 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
Ligand Effects in Rhodium-Catalyzed Hydroformylation with
Bisphosphines: Steric or Electronic?
Yunzhe Jiao*† ‡, Marta Serrano Torne‡, Jose Gracia†, J. W. (Hans) Niemantsverdriet†#
and Piet W. N. M. van Leeuwen*‡&
‡
Institute of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of Science
and Technology, Avgda. Països Catalans 16, 43007 Tarragona, Spain
†
SynCat@Beijing, Synfuels China Technology Co. Ltd, Beijing, P. R. China
#Syngaschem BV, PO Box 6336, 5600 HH Eindhoven, The Netherlands
&Present address: LPCNO, INSA-Toulouse, 135 Avenue de Rangueil, F-31077 Toulouse, France
β ≈ 120°
β = 123°
n
n
l/b = 1.9
l/b = 99
Do wide bite angles lead to high linear regioselectivity in
hydroformylation, or is an electronic effect operative?
Abstract
Twelve commercially available bisphosphine ligands have been evaluated in
rhodiumꢀcatalyzed hydroformylation reactions. All ligands exhibited high
chemoselectivities for aldehyde formation. The highest enantioselectivity (53% ee) of

Catalysis Science & Technology
Page 2 of 34
View Article Online
DOI: 10.1039/C6CY01990K
styrene hydroformylation was achieved with (S)ꢀBTFMꢀGarphos (L7) substituted with
electron withdrawing substituents. High pressure NMR (HPꢀNMR) and in situ high
pressure IR spectroscopy (HPꢀIR) were used to study the resting states of the catalyst
species in the reactions. The ligand effect on the structures of the observable species
was examined. Both electronic and steric factors were considered to contribute to the
performance of the various ligands. The results showed that decreasing the phosphine
basicity increased the enantioselectivity, while in the systems studied here the steric
character plays a less important role than the electronic feature in achieving good
regioselectivities.
Introduction
Hydroformylation is one of the most extensively studied homogeneous catalytic
processes in industry.1–3 Much effort has been devoted to the development of
catalysts/ligands in order to obtain high catalytic activities and selectivities. Compared
to the diphosphite ligands with good enantioselectivities in limited examples,
bisphosphine ligands such as bisdiazaphospholanes and bisphospholanes have been
reported to achieve excellent ee values and enhanced efficiency in asymmetric
hydroformylation with a wide range of substrates.4,5 For instance, the use of the
diazaphospholane ligands gives ee values of 86ꢀ99% in the asymmetric
hydroformylation of Nꢀvinyl carboxamides, allyl ethers, and allyl carbamates.6
Nevertheless, there is a need for novel bisphosphine ligands, especially chiral ones, to
broaden the substrate scope in hydroformylation.
In order to be able to predict the performance of a candidate ligand, it is necessary to
establish the relationship between ligand structure and the catalytic species formed in
the reaction. In general, bidentate phosphorus ligands form a pentacoordinate

Page 3 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
intermediates HRh(PꢀP)(CO)2 (PꢀP = bisphosphine), which play a decisive role in
determining the regioselectivity and enantioselectiviy of the reaction.7,8 In this study,
in situ spectroscopic experiments were conducted under the ‘real’ catalytic conditions to
monitor the reaction and the catalyst’s resting states. The formation of a single catalytic
species was suggested to be the key to maintain high enantioselectivities while other
species can also be present, as shown in highꢀpressure infrared (HPIR) absorption.9–11
The trigonal bipyramidal hydridorhodium complex equilibrates between two possible
isomeric structures of different phosphineꢀcoordination modes: equatorialꢀapical (ea) or
equatorialꢀequatorial (ee) coordination, as derived from the coupling constants in highꢀ
pressure NMR (HPNMR).12 Although the ee coordination mode is generally related to
a wide bite angle and high linear/branched ratios, the regioselectivities of the reaction
were not predictable solely from the chelation mode of the bisphosphine in HRh(Pꢀ
P)(CO)2.13
Fig. 1 Bite angle effect in l/b ratio using different bisphosphine ligands in the Rhꢀ
catalyzed hydroformylation of 1ꢀalkenes. 12–19
As shown in Fig. 1, the steric effect is demonstrated in the reaction as increasing the
bite angle of the ligand usually leads to a higher regioselectivity in the
hydroformation.12,20 For example, the linear/branched (l/b) ratio increased from 2.6

Catalysis Science & Technology
Page 4 of 34
View Article Online
DOI: 10.1039/C6CY01990K
with DPPE (βn = 87°) to 99 with NAPHOS (βn = 123°
).19–23 An even higher linearity
preference (l/b = 123) was observed for the BISBI derivative with CF3 substitution on
the aryl rings at phosphorus.24 However, a negative effect of electron withdrawing
substituents on regioselectivities was observed in bisphosphines such as the DPPE
ligands, which have small bite angles, with the resulting coordination at ea sites (the
numbers were not corrected for the formation of 2ꢀalkene side products).
In this work, we examined a series of commercially available bisphosphine ligands in
the rhodiumꢀcatalyzed hydroformylation of 1ꢀoctene and styrene. None of these have
been previously applied in alkene hydroformylations. Our experimental results
indicated that both steric and electronic character can affect the catalyst performance,
although unfortunately generalization of these effects cannot be made due to limited
data. Moreover, in situ techniques allowed us to determine the structures of the
dominant species, where the chelation modes of HRh(PꢀP)(CO)2 are largely influenced
by the steric factor of ligands.
Materials and methods
Materials
All reactions were carried out using standard Schlenk techniques under nitrogen or
argon atmosphere. Toluene was collected from the solvent purification system (MBraun
MB SPS800). Tolueneꢀd8 was dried, distilled over sodium/benzophenone for the HPꢀ
NMR experiment. Other solvents were purged with argon before use.
2ꢀMeTHF (anhydrous) was purchased from SigmaꢀAldrich and used without further
Cyclohexane were purchased from Scharlab and used without further drying. DPEphos
was purchased from SigmaꢀAldrich. All the other bisphosphine ligands were purchased
from Strem Chemicals. 2ꢀMeTHF, styrene and 1ꢀoctene were filtered over neutral

Page 5 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
1
31
31
alumina before use to remove hydroperoxides. H and P{1H}, P NMR spectra were
recorded on a Bruker 400 or 500 MHz spectrometer. Gas Chromatography was
performed on an Agilent 6890N, HP5 column. Enantiomeric excesses were measured
on the same equipment with βꢀdex 120 column. Standard infrared spectra were recorded
on a Thermo Nicolet FTꢀIR 5700 Nexus spectrometer equipped with a MCT detector,
KBr beam splitter at 4 cmꢀ1 resolution. Highꢀpressure transmission IR experiments were
performed in a 50 mL autoclave (SS 316) equipped with IRTRAN windows (ZnS), a
mechanical stirrer, a temperature controller, and a pressure transducer; homeꢀbuilt by
the workshop of the FNWI, University of Amsterdam.25
Standard hydroformylation experiments
A solution of Rh(acac)(CO)2 (0.005 mmol) and the phosphorus compound (0.01
mmol) were dissolved in 4 mL toluene and injected into the autoclave under N2. The
syngas was introduced and the heating and stirring was initiated for catalyst incubation.
Then 1 mL solution of substrate (5 mmol) was injected under syngas, followed by
reaching the reaction conditions. Reaction solutions were quenched immediately with
excess triꢀnꢀbutyl phosphite to avoid isomerization after the reaction. Conversions and
regioselectivities were determined by GC analysis of the crude samples after filtration
through silica gel.
In situ HP-IR experiments
The rhodium complex Rh(acac)(CO)2 (0.052 mmol) and the corresponding phosphorus
ligand (2 to 10 equiv.) were dissolved in 20 mL of 2ꢀMeTHF. This solvent was used
due to the lack of absorptions in the range of CO stretching bands and the solvent effect
is small considering the polarity change between toluene and 2ꢀMeTHF.26,27 After the

Catalysis Science & Technology
Page 6 of 34
View Article Online
DOI: 10.1039/C6CY01990K
HPꢀIR autoclave had been pressurized and the mixture heated, the autoclave was placed
in the infrared spectrometer. While the sample was being stirred the infrared spectra
were recorded.
HP-NMR experiments
The rhodium complex Rh(acac)(CO)2 (0.04 mmol) and the corresponding phosphorus
ligand (0.08 mmol) were dissolved in tolueneꢀd8 (0.4 mL), incubated for 1 hour under
10 bar of syngas at 80 °C. A portion of the resulting solution (0.2 mL) was transferred
to a sapphire tube (diameter = 10 mm) and pressurized with 10 bar of syngas. The tube
was placed in a Bruker 500 MHz spectrometer and the spectra were recorded at variable
temperatures.
Computational details
DFT calculations with the B3LYP28 and M06229 functionals have been carried out
with the gaussian09 program package30. In M062 calculations, the combination of a 6ꢀ
31G** basis set for light atoms (C, H, O, N, and F) and SDDALL for heavy atoms (P
and Rh) was used for initial structure optimization. Final optimization was conducted
using LANL2DZ for all atoms in both M062 and B3LYP calculations. Frequency
calculations have been carried out at the optimized structures to confirm that they are
minima on the potential energy surface. Because of the polarity of the structures, the
solvent effects on their relative stabilities were evaluated by calculating the free
energies of the solvation in terms of the polarizable continuum model (PCM). Gibbs
free energies have been calculated at 298.15 K and 1 atm.
Results and discussions

Page 7 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
In Situ HP-IR Analysis of Rhodium-DPEphos/DTP-DPEphos
As shown in Scheme 1, DPEphos (L1) has proved to be a good ligand for
hydroformylation with a high selectivity and a moderate activity. Its o-tolyl derivative
DTPꢀDPEphos (L2) has a few applications in Heck coupling and aryl halide amination
but has not been tested in hydroformylation.31,32 While steric factors are held
responsible for the determination of the l/b ratio in hydroformylation33, these steric
effects are brought about by changing the backbone and there are hardly examples of
steric variations of the substituents on phosphorus using the same backbone. To
compare the substituent effect, in situ HPꢀIR and HPꢀNMR techniques were used to
determine the dominant species present during the reaction of the typical rhodium
precursor Rh(acac)(CO)2 (1) and the ligand. HPꢀFTIR was performed under 20 bar of
syngas at RT to 80 °C. In the reaction of L1, precursor 1 was depleted too fast to be
observable from comparison with the reported IR absorptions (2082, 2011 cmꢀ1).
Instead, two peaks at 1992 and 1943 cmꢀ1 appeared immediately after pressurization at
r.t. (Fig. 2(a)). The frequencies can be assigned as HRh(CO)2(L1) by comparison with
the reported absorptions for the type of hydride species HRh(CO)2(PꢀP) with ea
phosphorous coordination (Table 1).12,34
Scheme 1 Nonchiral bisphosphine ligands L1 and L2 for hydroformylation.

Catalysis Science & Technology
Page 8 of 34
View Article Online
DOI: 10.1039/C6CY01990K
(1)
(2)
During heating to 80 °C, HRh(CO)2(L1) remained as the only formed species within
terminal carbonyl region and did not change after substrate addition, suggesting
HRh(CO)2(L1) as the resting state in the catalytic cycle. The same species were
observed in using cyclohexane instead of 2ꢀMeTHF as the solvent, indicating the
independence on the solvent polarity. As shown in Fig. 2(b), a new band appeared at
1885 and 2043 cmꢀ1 with the decay of 1 under pressurization of syngas in the absence of
any phosphine ligand. This species is attributed to Rh4(CO)12 (2) (reported ν
/cm–1:
1885.4, 2044.2, 2071.2)35 with the assumption that the other major bands are covered by
the peak at 2077 cm–1 from the subsequent rapid formation of Rh6(CO)16 (3) (observed
ν
/cm–1: 1816, 2077; reported /cm–1: 1818.4, 2075.4)35 (Fig. 2(b), eq 1). The formation
ν
of 2 is irreversible as the following injection of ligand did not give the expected
formation of HRh(CO)2(L1).

Page 9 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K

Catalysis Science & Technology
Page 10 of 34
View Article Online
DOI: 10.1039/C6CY01990K
Fig. 2 (a) HPꢀFTIR spectra for the reaction of 1 with L1 in 2ꢀMeTHF at 80
°C, 20 bar
of CO/H2 (1:1) (carbonyl region). (b) HPꢀFTIR spectra for the reaction of 1 without any
phosphine ligand in 2ꢀMeTHF at 80 C, 20 bar of CO/H2 (1:1) (carbonyl region).
Spectra were recorded from top to bottom.
°
In contrast, no occurrence of HRh(CO)2(L2) was observed during an HPꢀIR
experiment of L2 under the same condition (Fig. 3, eq 2). The metal precursor 1
disappeared within one hour during heating and correspondingly 3 formed as the first
intermediate. The minor peak at 2042 cm–1 might be from the formation of 2. Within
four hours at 80 °
C, decomposition of 3 occurred, leaving one major peak at 2038 cm–1.

Page 11 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
Fig. 3 HPꢀFTIR spectra for the reaction of 1 with L2 in 2ꢀMeTHF at 80 °C, 20 bar of
CO/H2 (1:1) (carbonyl region). Spectra were recorded from top to bottom.
Table 1 HPꢀFTIR data of the species formed in the hydroformylation reaction.
ee (cm−1)
IR ν
ea (cm–1)
Complexes
IR
ꢀꢀ
ν
HRh(CO)2(L1)
HRh(CO)2(L4)
HRh(CO)2(L6)
HRh(CO)2(L7)
1992, 1943
ꢀꢀ
2044, 1967
ꢀꢀ
ꢀꢀ
1990, 1948
2010, 1970
2000, 1961
1988, 1944
2039, 2005, 1949
HRh(CO)2(CHIRAPHOS) a ꢀꢀ
HRh(CO)2(BDPP) a
ꢀꢀ
HRh(CO)2(SPANPhos) a
2027, 1978

Catalysis Science & Technology
Page 12 of 34
View Article Online
DOI: 10.1039/C6CY01990K
HRh(CO)2(pySPAN) a
2063, 2009
2042, 1981
2037, 1979
ꢀꢀ
a
HRh(CO)2(PPh3)2
1992, 1947
1990, 1947
a
HRh(CO)2(PEtPh2)2
a Reported absorptions for the type of hydride species HRh(CO)2(bisphosphine).12,33
High Pressure NMR Analysis of Rhodium-DPEphos Pretreatment
To confirm the structures of the resting states formed from precursor and ligand, HPꢀ
NMR spectra were measured after the reaction between 1 and L1. Concentrated
1
rhodium solution was prepared in order to achieve good signal to noise ratio. The H
NMR signals (Fig. 4) support the formation of HRh(CO)2(L1), with a typical triplet of
2
doublets for the hydride at
δ
–8.78 (1JRh−H = 10.8 Hz, JP−H = 44.7 Hz) due to the
coupling of the hydride with two (timeꢀaveraged) equivalent phosphorus atoms and the
rhodium atom. The coupling constants smaller than those reported for HRh(CO)2(BDPP)
(δ
–8.8, 1JRh−H = 11 Hz, 2JP−H = 57 Hz)14 correspond to the decreased phosphine basicity
of L1. The minor signal at δ −9.23 (quartet, JRh−H = 12,6 Hz) is assigned to
1
HRh(L1)2(CO) based on the reported BDPP analog HRh(BDPP)2(CO) (
δ
–10.2, JRh−H
= 16 Hz)36, in which one of the two bisphosphines coordinates to rhodium in a
monodentate fashion (Table 2).

Page 13 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
1
Fig. 4 H NMR spectra for the reaction of 1 with L1 in tolueneꢀd8 at 80 °C, 10 bar of
CO/H2 (1:1) after 1 h.
The corresponding 31P{1H} NMR spectrum showed a mixture of two doublets as well
as the singlets belonging to the free phosphines and phosphine oxide. The doublet at
δ
28.03 ppm was attributed to HRh(CO)2(L1), which has an averaged Rh–P coupling
constant of 117.4 Hz (Table 2; see SI, Fig. Sꢀ1). When the temperature was decreased to
–80
and 25.5 ppm. All evidence suggests an ea geometry for HRh(CO)2(L1), in which the
equatorial phosphorus atom has a small Rh Pcis coupling constant and the apical
°C, the doublet broadened and resolved into two separate broad doublets at δ 32.4
−
phosphorus has a typical larger Rh
−Ptrans coupling constant. Due to the electronic
difference, the Rh
−
Ptrans coupling constant in HRh(CO)2(L1) (1JRh−P = 117.4 Hz) is
slightly smaller than that in eaꢀHRh(BDPP)2(CO) (1JRh−P = 120.0 Hz). 31P NMR showed
2
a double doublet at RT, in which the size of JHꢀP of 43.2 Hz again indicated the fast

Catalysis Science & Technology
Page 14 of 34
View Article Online
DOI: 10.1039/C6CY01990K
exchange of the two phosphorus atoms in equatorial and apical positions at room
temperature (see SI, Fig. Sꢀ2). However, in this instance the distinct axial P–H and
equatorial P–H coupling was not observable even at –80
°C.
The other broad doublet at 11.56 (JRh–P = 153.8 Hz) was tentatively assigned to a
δ
dimer species [(L1)Rh(CO)2]2 (see SI, Fig. Sꢀ2). However, instead of the expected two
pairs of doublets for restricted rotation at low T as reported previously for
[Rh(BDPP)(CO)2]2 36, four sets of doublets appeared at
δ
20.9, 12.7, 10.2, 3.5 with
C for the unknown
respective coupling constants of 187, 113, 113, 187 Hz at –80
°
structure. Skewed structures of L1 could lead to a lower symmetry, but we have no
proof for that.
1
Table 2 H and 31P{1H} NMR data of the species formed under hydroformylation
conditions (80
Complex
°
C and 10 bar of 1:1 CO/H2).
1H NMR
31P{1H} NMR
1JRh–P(Hz)
(ppm)
2
δ(ppm)
1JRh−H(Hz) JP−H(Hz)
δ
eaꢀHRh(CO)2(L1)
HRh(L1)2(CO)
–8.78
–9.23
–9.21
10.8
12.6
9.6
11
44.7
ꢀꢀꢀ
28.03
ꢀꢀꢀ
117.4
ꢀꢀꢀ
eaꢀHRh(CO)2(L7)
62.5
57
44.71
29.8
ꢀꢀꢀ
120.0
112
ꢀꢀꢀ
eaꢀHRh(CO)2(BDPP) –8.8
HRh(BDPP)2(CO)
HRh(PEtPh2)3(CO)
–10.2
–10.1
16
ꢀꢀꢀ
br
ꢀꢀꢀ
34.5
150

Page 15 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
Hydroformylation of 1-Octene with L1 and L2: Ligand Effect on Catalyst
Performance
L1 and L2 were examined in the hydroformylation of 1ꢀoctene under various
conditions (eq 3). The reaction was carried out at 80 °C and 20 bar of 1:1 CO/H2 using a
1.0 mM solution of rhodium precursor 1 and the corresponding bisphosphine ligand.
The ligand/metal (L/M) ratio was varied at 3, 5, 10 and 15 to study the effect of ligand
concentrations on the catalytic activity of hydroformylation. The product mixtures were
monitored by gas chromatography and the results are summarized in Table 3. For L1,
our results are identical to the reported value with a smaller L/M ratio.12 The increase
of ligand/metal ratio slows down the reaction a little and has minimal influence on the
linear selectivity. Compared to L1, the tolyl derivative L2 showed higher activities but
lower selectivities for aldehyde due to alkene isomerization. Also, the linearity (l/b
ratio) is less favored with L2 than L1, which can be attributed to the fact that L2 cannot
coordinate to rhodium as found in the in situ spectroscopic studies. The failure of
phosphine coordination to rhodium is due to the increased steric hindrance of L2 and
the competition of CO. The resulting high activities are not surprising because the
dominant species HRh(CO)4 (2) is an active catalyst in hydroformylation but usually
has low selectivity control. The dissociative behavior of bulky monodentate ligands has
been previously reported under hydroformylation conditions, forming unselective
Rh6(CO)16 (3).37 Addition of excess L2 can inhibit the olefin isomerization and slightly
increase the l/b ratio.
(3)

Catalysis Science & Technology
Page 16 of 34
View Article Online
DOI: 10.1039/C6CY01990K
Table 3 Hydroformylation of 1ꢀoctene with the Rh(acac)(CO)2/Ligand system. a
Ligand L/M Conv. (%) Aldehyde Regioselectivity. l/b ratio
Sel. (%)
100
100
100
100
42
(linear %)
3
67
61
57
59
99
99
99
99
ꢀꢀꢀ
87
87
87
87
25
27
36
38
87
6.7
6.8
6.8
6.9
1.4
1.4
1.6
1.7
6.7
L1
L1
L1
L1
L2
L2
L2
L2
5
10
15
3
5
46
10
15
2.2
59
61
L1b
100
a
Reaction conditions: P(CO)/P(H2) = 1, P(CO/H2) = 20 bar, substrate/Rh = 667,
Rh(acac)(CO)2 = 1 mM. Incubation at 60 °C, 45 mins, 10 bar; Reaction at 80 °C, 12 h,
b
20 bar.
Previously reported reaction at P(CO)/P(H2) = 1, P(CO/H2) = 10 bar,
C, 16 h.
substrate/Rh =674, Rh(acac)(CO)2 = 1.78 mM , 80
°
The fast formation of the active species at r.t. in the HPꢀIR experiment implies that
pretreatment of catalyst with L1 is not necessary. Therefore, catalytic activities were
compared under different incubation conditions (see SI, Table S–2). As expected from
the spectroscopic studies, similar conversions, aldehyde selectivities, and l/b ratios were
observed regardless of the incubation conditions.
Hydroformylation of 1-Octene with Chiral Ligands

Page 17 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
Seven commercially available chiral bisphosphines L3–9 were tested in rhodiumꢀ
catalyzed hydroformylation of 1ꢀoctene (Scheme 2, Table 4). Dominating formation of
internal octenes was seen when L5 and L8 were used. For L8 this can be understood
from the steric bulk on the aryl groups; from DFT calculations it is known that two aryl
groups on two phosphine atoms during hydroformylation show πꢀstacking and the bulky
groups prevent this. All other ligands showed high activities and selectivities to
aldehydes. The formation of the linear aldehyde (1ꢀnonanal) is slightly favored over the
branched product (2ꢀnonanal). The low 1/b ratios for L3 and L4 are unexpected
considering their wide bite angles of over 120°, although at too wide angles the
relationship breaks down. In addition L3 is highly flexible and the amides may
participate in the coordination, as seen in Pd chemistry. As a high linearity is usually
correlated to bidentate phosphines with large βn (NAPHOS: βn = 123
inclusion of the hetero Nꢀatom in L4 must change the selectivity by electronic means as
a considerable P–N double bond character via ꢀinteraction, indicated by the trigonal
°
, l/b = 99)19, the
π
planar geometry around the nitrogen atoms, has been observed in aminophosphines.38
For L6 a behaviour similar to DPEphos might be expected in view of the bite angles,
but perhaps this is counteracted by the rigidity of L6. On BISBI and Xantphos ligands
electron withdrawing groups such as CF3 (as present in L7 and L9) favour higher l/b
ratios. For bite angles below 90°, however, Casey reported a deleterious effect of
electronꢀwithdrawing groups and thus the effect on L7 and L9 could be either way. In
dissymmetric diphos ligands the electron withdrawing phosphine occupies the
equatorial position and this leads to an increase of l/b.39 The most regioselective ligand
is ferroceneꢀtype bisphosphine L9 but the l/b ratio of 3.4 is smaller than those of the
nonchiral derivatives in literature (l/b ratios from 5.4 to 11.4 for DPPF ligands).40,41 L9

Catalysis Science & Technology
Page 18 of 34
View Article Online
DOI: 10.1039/C6CY01990K
is highly sterically hindered and the hampered πꢀstacking may also contribute to the
modest, positive effect caused by the equatorial electronꢀwithdrawing phosphine.
Scheme 2 Chiral ligand scope for rhodiumꢀcatalyzed hydroformylation.
Table 4 Hydroformylation of 1ꢀoctene with chiral bisphosphines.a
b
Ligand
P−
Rh
−P (°)
Conv. (%) Aldehyde Sel. (%) l/b ratio
125
121
91
96
100
100
6
1.5
1.9
1.4
2.2
2.8
0.8
3.4
L3
L4
L5
L6
L7
L8
L9
99
100
99
105
91
100
97
100
100
100
95
5
ꢀꢀꢀ
100
a Reaction conditions: P(CO)/P(H2) = 1, P(CO/H2) = 20 bar, substrate/L/Rh = 1000/2/1,
b
Rh(acac)(CO)2 = 1 mM. 80
°C, 12 hrs.
Calculated P−Rh−P bond angles in

Page 19 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
HRh(CO)2(P−P) (ee); optimization of L9 could not be achieved probably due to the
steric bulkiness. Note that the bite angles reported here are those calculated for the
trigonalꢀbipyramidal rhodium complexes and that they are not "natural bite angles"
calculated with MM2 according to Casey and Whitteker, using a dummy atom without
angular preferences42.
Asymmetric Hydroformylation of Styrene with Chiral Ligands
Screening of L3–L13 as new bisphosphine ligands in hydroformylation of styrene
revealed very high regioselectivites in all cases except for L4 (Scheme 3, eq 4, Table 5).
In all other instances the formation of the branched aldehyde was preferred and even
exclusive formation of the branched product was seen for L5 and L8; interestingly these
are the ligands that gave mainly isomerization for 1ꢀoctene as the substrate, indicating
that the terminal alkylꢀRh is not very stable. Comparison of L7 and L8 indicates that
electronꢀdonating groups on the phosphine aryl rings reduces the catalytic activity and
enantioselectivity. A similar electronic effect is also demonstrated in the case of L12
and L11, where fluorine substitution on the backbone has increased the conversion from
20% to 39%. The trend of inhibited activity by electron rich ligands cannot be
generalized to a wider scope as the electronꢀricher methyl derivative L10 has slightly
higher conversion than its parent complex L11 in Scheme 3. The formation of linear
aldehyde is favored with less basic phosphine ligands as indicated by a trend of l/b of 0
(L8) < 0.09 (L10) < 0.56 (L11). We did observe a correlation of ligand basicity on the
one hand and enantioselectivity and linear preference on the other hand within the
SEGPHOS ligands L8, L10 and L11. The increase of enantiomeric excess runs in
parallel with a decrease of ligand basicity: 19% (L8) < 22% (L10) < 29% (L11),
although the ee range is too small to arrive at a firm conclusion.

Catalysis Science & Technology
Page 20 of 34
View Article Online
DOI: 10.1039/C6CY01990K
The highest enantioselectivity of 53% ee was obtained for (S)ꢀBTFMꢀGarphos (L7)
with the electronꢀwithdrawing CF3 substituents. L7 seems to have the greatest potential
in asymmetric hydroformylation evidenced as well by its high activity and a good
branched preference.
CH3
O
O
O
O
F
F
PPh2
PPh2
PPh2
PPh2
PPh2
PPh2
O
O
CH3
CH3
O
O
P
P
O
O
2
2
O
O
F
F
O
O
O
O
CH3
(S)-(-)-DM-SEGPHOS
(S)-(-)-SEGPHOS
(S)-DIFLUORPHOS
(S)-SYNPHOS
L10
(
)
L11
L12
L13
( )
(
)
(
)
Scheme 3 New Ligands for asymmetric hydroformylation of styrene.
(4)
Table 5 Hydroformylation of styrene with chiral bisphosphines.a
Ligand Conv. (%)
Aldehyde
Sel. (%)
l/b ratio
ee (S)
(%)
98
100
100
100
100
100
100
14/86
50/50
ꢀꢀ/100
7/93
–3
1
L3
L4
L5
L6
L7
L8
98
11
–21
–4
53
19
100
100
12
7/93
ꢀꢀ/100

Page 21 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
100
65
23
20
39
18
96
13/83
10/90
8/92
–11
–2
22
29
33
19
L9
L9b
L10
L11
L12
L13
100
100
100
100
100
36/64
9/91
8/92
a Reaction conditions: P(CO)/P(H2) = 1, P(CO/H2) = 20 bar, substrate/L/Rh = 1000/2/1,
b
Rh(acac)(CO)2 = 1 mM. , incubation 80
excess ligand inhibits conversion and ee.
°C, 1 hr, reaction 80
°
C, 12 hrs. L/M = 10:
Condition Optimization of Hydroformylation with L7
With the promising results for L7 in asymmetric hydroformylation under the
“standard” condition, the reaction parameters such as T, P, solvent, incubation and L/M
were varied to optimize the ee as shown in Table 6. Exclusive formation of aldehyde
product was observed in all cases. The reaction time of 12 h is necessary to drive the
reaction to completion (entry 1 vs 2). The process of catalyst pretreatment does not
affect the enantioselectivity much (entry 1 vs 5). The optimal pressure is 20 bar as a
poor enantioselectivity of 4% ee was obtained at 8 bar and the selectivity was not
further increased using a higher pressure of 30 bar (entry 1 vs 3, 5 vs 6). Addition of
excess ligand (L/M = 2) is critical to maintain the enantiomeric excess around 50%
(entry 2 vs 9). Increasing the ligand/metal ratio to 5 equivalents slightly hampered the
catalyst performance (entry 1 vs 10). The influence of ligand excess and pressure cannot
be generally applied to all bisphosphine ligands as the opposite trend was seen for L8
(see SI, Table Sꢀ1).
Table 6 Results of hydroformylation of styrene with L7 under different conditions. a

Catalysis Science & Technology
Page 22 of 34
View Article Online
DOI: 10.1039/C6CY01990K
L/
M
Conv.
(%)
l/b
ratio
ee
(%)
#
Incubation b Reaction
Solvent
1
2
3
4
5
6
7
8
9
1
80
2
Toluene
Tol
100
95
97
50
98
98
81
70
48
99
7/93
53
45
46
48
48
4
80 °C
80 °C
80 °C
60 °C
None
None
2
10/90
10/90
8/92
80 °C/20bar/8h
80 °C/30b/12h
60 °C/20b/12h
80 °C/20b/12h
80 °C/8b/12h
80 °C/20b/12h
80 °C/20b/12h
80 °C/20b/8h
80 °C/20b/12h
2
Tol
2
Tol
2
Tol
10/90
41/59
52/48
17/83
30/70
10/90
2
Tol
2
2ꢀMeTHF
cꢀhexane
Tol
2
80 °C
80 °C
80 °C
80 °C
2
39
28
49
1.2
5
Tol
a Exclusive formation of aldehyde was observed in all entries. b Incubation condition: 10 bar of
CO/H2 (1:1) for 1 h
.
To understand the better catalytic performance of L7, in situ HPIR experiments for L7
were performed under the catalytic conditions (80 C, 20 bar). At the beginning of
°
heating, the catalyst precursor 1 was observed with two absorptions at 2082 and 2011
cm–1 in the terminal carbonyl region (Fig. 5). When the temperature increased to 65
°C,
1 was completely converted to a new species, assigned to HRh(CO)2(L7) with a pair of
absorptions at 2010 and 1970 cm–1. The carbonyl bands of HRh(CO)2(L7) shift to
higher wavenumbers (respectively 18 and 27 cm–1) compared to HRh(CO)2(L1),
consistent with the electron deficiency in L7. Additional evidence for the ea structure of
HRh(CO)2(L7) was seen from the HPꢀNMR spectra (Fig. 6). Similar to eaꢀ
1
2
1
HRh(CO)2(L1) (Table 2,
δ
–8.78, JRh−H = 10.8 Hz, JP−H = 44.7 Hz), the H NMR of
HRh(CO)2(L7) showed a typical triplet of doublets at
δ
–9.21 with 1JRh−H = 9.6 Hz and
2JP−H = 62.5 Hz for the hydride resonances. Consistently, a doublet signal at
δ
44.71
31
1
(1JRh−P = 120.0 Hz) was observed in P{1H} NMR (eaꢀHRh(CO)2(L1):
δ
28.03, JRh−P
= 117.4 Hz) and the H–P coupling constant of 63.5 Hz was obtained from 31P NMR (see
1
1
SI, Fig. Sꢀ3 and Sꢀ4). The smaller JRh−H and larger JRh−P are indicative of the nonꢀ

Page 23 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
basicity of L7 relative to L1 and BDPP.
Fig. 5 HPꢀFTIR spectra for the reaction of 1 with L7 in 2ꢀMeTHF at 80 °C, 20 bar of
CO/H2 (1:1) (carbonyl region). Spectra were recorded from top to bottom.

Catalysis Science & Technology
Page 24 of 34
View Article Online
DOI: 10.1039/C6CY01990K
1
Fig. 6 H NMR spectra (hydride region) for the reaction of 1 with L7 in Tolueneꢀd8 at
80 C, 10 bar of CO/H2 (1:1) after 1 h.
°
L6 has a calculated bite angle of 105°, slightly wider than that of L7. The IR spectra
recorded under catalytic conditions showed the formation of a mixture of intermediates.
The dominating resting state is HRh(CO)2(L6) with the two absorptions at 1990 and
1948 cm–1 (Fig. 7(a)). There are other weak bands appearing at 2046, 2014, 2004, 1797
cm–1 which can probably be attributed to rhodium carbonyl species or/and the
equatorialꢀequatorial isomer of HRh(CO)2(L6). Under the same experimental
conditions, the IR spectra of L4 revealed two absorptions of terminal carbonyl
stretching at 2044, 1967 cm–1 from the beginning of heating (Fig. 7(b)). This suggests a
rapid formation of the resting state HRh(CO)2(L4). The higher frequencies of the
carbonyl stretching vibrations of HRh(CO)2(L4) are in accordance with the equatorialꢀ
equatorial coordination of the bisphosphine, which can be expected from its calculated
natural bite angle of around 120°. The exceptional ee geometry of HRh(CO)2(L4) can

Page 25 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
explain the formation of linear aldehyde for styrene (l/b = 1/1) as shown in Table 5,
similar to Xantphos (l/b = 4.7/5.3).12

Catalysis Science & Technology
Page 26 of 34
View Article Online
DOI: 10.1039/C6CY01990K
Fig. 7 (a) HPꢀFTIR spectra for the reaction of 1 with L6 in 2ꢀMeTHF at 80
°C, 20 bar
of CO/H2 (1:1) (carbonyl region). (b) HPꢀFTIR spectra for the reaction of 1 with L4 in
2ꢀMeTHF at 80 C, 20 bar of CO/H2 (1:1) (carbonyl region). Spectra were recorded
°
from top to bottom.
Calculated free energy differences between isomers.
We have calculated the free energies of the two coordination modes (ea and ee) of
HRh(CO)2(P−P) in both gas phase and solution, using the Polarizable Continuum
Model (PCM), see Table 7. Calculations with the B3LYP hybrid functional including
PCM seem generally in better agreement with the IR observations than those with the
M062 functional. Nonetheless, the energy difference between isomers seems
overestimated for L7 by B3LYP/PCM. Dominating ea geometry is seen in cases of L7,
BDPP and CHIRAPHOS, where the free energy differences between ee and ea isomers
are over 2 kcal·mol–1, equivalent to a ratio of ea/ee of ~ 30. Isomers with smaller free
energy differences display little preferences in geometry distributions for ligands such
as SPANphos, DPEphos, pySPAN and monodentate phosphines. As one can see from
the table we have obtained excellent agreement between calculated energies and
experimentally observed isomer distributions.
Table 7 Free energies (kcal·mol–1) for stationary points in the two coordination modes
(ea and ee) of HRh(CO)2(bisphosphine) in comparison with experimentally observed
coordination modes. Toluene was used as the solvent in PCM.
Isomer
ꢀ
Gea −ꢀGee (kcal·mol−1) of
Ligand
(observed)
HRh(CO)2(bisphosphine)
B3LYP/LANL2DZ
M062/LANL2DZ
Gas
Phase
Gas
PCM
PCM
Phase

Page 27 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
–2.51
–2.45
–1.6
–3.52
–2.76
–2.74
–2.18
–1.3
–1.22
–2.21
–1.35
–1.55
–2.3
–1.52
–2.58
–2.14
–1.58
–2.44
–0.81
–2.37
–0.51
0.19
L7
ea
BDPP
L6
ea
ea
CHIRAPHOS
SPANphos
L1
–2.18
–0.64
–0.57
0.94
ea
ee/ea
ea
–1.09
0.16
–0.24
–2.8
2 PPh3
2 PEtPh2
L4
ee/ea a
ee/ea
ee
1.22
0.24
0.24
0.73
0.65
–1.19
–1.35
3.29
pySPAN
L2
1.63
1.68
–1.37
2.84
ee
2.14
2.31
none
a The major isomer has ee coordination mode.
Solvent effects were also studied under the “standard” conditions. As previously
mentioned, in situ HPꢀIR studies of the catalyst pretreatment using L7 in 2ꢀMeTHF
displayed the successful formation of eaꢀHRh(CO)2(L7). However, the actual
hydroformylation reaction of styrene in 2ꢀMeTHF is much less regioselective and
enantioselective than that in toluene (Table 6, entry 1 vs 7). Cyclohexane, less polar
than toluene, also gave a slower reaction. Attempts to clarify the origin of solvent effect
in L7 were undertaken by DFT calculations. Both the ea/ee ratio and the interaction
energy difference go up, opposite to the decline of the interaction energy between
rhodium and the ligand, as the solvent polarity increases (Table 8). It remains unclear
how the solvent affects the rate of reaction in the current work. Perhaps the solvent can
play a directing role in the geometry of alkene coordination by competitive interaction
with the metal center; this is beyond the scope of the present paper.

Catalysis Science & Technology
Page 28 of 34
View Article Online
DOI: 10.1039/C6CY01990K
Table 8 Free energies (kcal·mol–1) for stationary points in the two coordination modes
(ea and ee) of HRh(CO)2(L7) using PCM for three solvents: nꢀhexane, toluene and
tetrahydrofuran.
Solvent
ꢀGea−ꢀGee(kcal·mol−1)
Einteraction (kcal·mol−1) a
B3LYP/LANL2DZ M062/LANL2DZ B3LYP/LANL2DZ M062/LANL2DZ
ea
ee
ea
ee
Gas phase
nꢀHexane
Toluene
–2.51
–3.30
–3.52
–4.15
–1.22
–1.45
–1.52
–1.76
–30.68
–29.57
–29.20
–26.98
–28.18
–26.27
–25.68
–22.84
–38.79 –37.57
–36.93 –35.48
–36.52 –35.00
–34.58 –32.82
THF
a
Einteraction is the interaction energy between L7 and the metal center. It is calculated
with the equation Einteraction = E(HRh(CO)2(L7)) – E([HRh(CO)2]) – E(L7).
Conclusions
A series of new bisphosphine ligands has been examined in rhodiumꢀcatalyzed
hydroformylation. The presence of (bisphosphine)Rh(CO)2H was detected by HPꢀIR
and HPꢀNMR spectra in selected cases. The low selectivity in 1ꢀoctene
hydroformylation for DTPꢀDPEphos (L2) can be explained by the difficulty of the
bulky ligand to coordinate to the rhodium center in its competition with CO, while its
parent bisphospine DPEphos (L1) forms stable, active catalysts under the conditions.
Both spectroscopic and theoretical studies indicate that the ee/ea equilibrium is
controlled by the steric factor of the bisphosphines. Exclusive formation of ea isomers
was seen for the ligands with small bite angles of around 100
°. For CTHꢀ(R)ꢀBINAM
(L4), which has a wider bite angle of around 120 , both phosphorus atoms chelate in the
°

Page 29 of 34
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C6CY01990K
equatorial sites. Moreover, a simple correlation between ea/ee ratio and the l/b ratio
does not exist within the scope of this work. L4, obtained by replacing the methylene
groups in NAPHOS by NH, displays structural similarity with the well performing
NAPHOSꢀtype ligand (l/b = 99) and adopts an equatorialꢀequatorial resting state in
HRh(CO)2(L4). However, the l/b ratio is low in the hydroformylation of 1ꢀoctene with
L4. The poor regioselectivity of L4 disproved once again that a wide bite angle cannot
be used as the exclusive parameter in predicting the linear selectivity in
hydroformylation reactions. Instead, the electronic change brought about by heteroatom
incorporation dramatically lowers the selectivity. For 1ꢀoctene hydroformylation, the
highest linearity was achieved with the less basic bisphosphine L7 and the ferrocene
bisphosphine L9. Hence, within the ligands for rhodiumꢀcatalyzed hydroformylation
considered in this paper, it appears that electronic control is more important than bite
angle. Among all the bisphosphine complexes, L7 was the most promising ligand for
asymmetric hydroformylation of styrene with a high conversion and branched
selectivity (b/l = 13) (53% ee), although better alternatives are available from literature.
We have shown that DFT can very well predict the structure of the resting state
HRh(CO)2(P−P), be it equatorialꢀequatorial or equatorialꢀapical, which offers a useful
alternative in case in situ NMR or IR would not be available.
Acknowledgements
Y.J. thanks the Institute of Chemical Research of Catalonia (ICIQ), 43007 Tarragona,
Spain for hosting her, and Syngaschem BV for enabling her stay in The Netherlands and
Spain. Syngaschem BV acknowledges financial support from Synfuels China
Technology Co. Ltd.