Phenoxaphosphine-Based diphosphine ligands. synthesis and application in the hydroformylation reaction

Article abstract of DOI:10.1021/om901041r

The synthesis of a new series of diphosphine ligands based on 2,7-di-im-butyl-9,9-dimethylxanthene (1), P-tolyl ether (2), ferrocene (3), and benzene (4) backbones, containing one or two 2,8-dimethylphenoxaphosphine moieties, is reported. The ligands were employed in the rhodium-catalyzed hydroformylation of 1-octene. For all four ligand backbones, introduction of phenoxaphosphine moieties led to an increase in catalytic activity and a decrease in regioselectivity toward the linear aldehyde product. Xanthene-based ligands la- Ic yielded highly active and regioselective hydroformylation catalysts; ligands containing/p-tolyl ether and ferrocene backbones 2a-2c and 3a-3c provided less active and less regioselective catalysts. Catalysts containing benzene-derived ligands 4a and 4b showed a remarkable preference for the formation of the branched aldehyde product. The coordination behavior of ligands 1-4 under hydroformylation conditions was investigated using high-pressure NMR and IR spectroscopy, revealing the distinct steric and electronic properties of the diphenylphosphine and 2,8-dimethylphenoxaphosphine moieties in ligands 1-4. The phosphacyclic moieties proved to be less basic and less stoically demanding toward other ligands in metal complexes than the acyclic diphenylphosphine moieties. For ligands that contain rigid backbones, the lack of conformational freedom in these phosphacyclic moieties does lead to repulsive interactions between the substituents of the two phosphorus donor atoms, resulting in an increase in the bite angle of the ligand. The low catalytic activity of rhodium catalysts modified by benzene-based ligands 4a-4c was attributed to the quantitative formation of HRh(L)2 under hydroformylation conditions.

Full text of DOI:10.1021/om901041r

1210 Organometallics 2010, 29, 1210–1221
DOI: 10.1021/om901041r
Phenoxaphosphine-Based Diphosphine Ligands. Synthesis and
Application in the Hydroformylation Reaction
Erik Zuidema,† P. Elsbeth Goudriaan,† Bert H. G. Swennenhuis,† Paul C. J. Kamer,*,‡
Piet W. N. M. van Leeuwen,*,†,§ Martin Lutz,^ and Anthony L. Spek^
†Van ’t Hoff Institute for Molecular Sciences, Universiteit van Amsterdam, Nieuwe Achtergracht 166,
1018 WV Amsterdam, The Netherlands, ‡EaStCHEM, Department of Chemistry, St. Andrews University,
St. Andrews, Scotland, U.K., §Institute of Chemical Research of Catalonia (ICIQ), Avenida Paısos Catalans,
¨
16, 43007 Tarragona, Spain, and ^Bijvoet Centre for Biomolecular Research, Utrecht University,
Padualaan 8, 3584 CH Utrecht, The Netherlands
Received December 3, 2009
The synthesis of a new series of diphosphine ligands based on 2,7-di-tert-butyl-9,9-dimethylxanthene (1),
p-tolyl ether (2), ferrocene (3), and benzene (4) backbones, containing one or two 2,8-dimethylphenoxaphos-
phine moieties, is reported. The ligands were employed in the rhodium-catalyzed hydroformylation of
1-octene. For all four ligand backbones, introduction of phenoxaphosphine moieties led to an increase in
catalytic activity and a decrease in regioselectivity toward the linear aldehyde product. Xanthene-based
ligands 1a-1c yielded highly active and regioselective hydroformylation catalysts; ligands containing p-tolyl
ether and ferrocene backbones 2a-2c and 3a-3c provided less active and less regioselective catalysts.
Catalysts containing benzene-derived ligands 4a and 4b showed a remarkable preference for the formation
of the branched aldehyde product. The coordination behavior of ligands 1-4 under hydroformylation
conditions was investigated using high-pressure NMR and IR spectroscopy, revealing the distinct steric and
electronic properties of the diphenylphosphine and 2,8-dimethylphenoxaphosphine moieties in ligands
1-4. The phosphacyclic moieties proved to be less basic and less sterically demanding toward other ligands
in metal complexes than the acyclic diphenylphosphine moieties. For ligands that contain rigid backbones,
the lack of conformational freedom in these phosphacyclic moieties does lead to repulsive interactions
between the substituents of the two phosphorus donor atoms, resulting in an increase in the bite angle of the
ligand. The low catalytic activity of rhodium catalysts modified by benzene-based ligands 4a-4c was
attributed to the quantitative formation of HRh(L)2 under hydroformylation conditions.
1. Introduction
with a bite angle of approximately 120ꢀ were attributed to the
preferred bisequatorial coordination mode of these ligands in
the trigonal-bipyramidal rhodium complexes involved in
the hydroformylation reaction (Figure 1). On the basis of
this observation, several wide bite angle diphosphine ligands
were successfully developed.4 The coordination mode of the
diphosphine ligand is however not the sole factor governing
the regioselectivity of the catalyst.5 Just as for monodentate
Due to its large-scale industrial application, the hydrofor-
mylation reaction is one of the most studied homogenously
catalyzed reactions. The activity and chemo-, regio-, and
enantioselectivity of the catalyst system can be influenced to
a large extent by ligands coordinated to the catalytically
active rhodium, cobalt, or platinum metal centers. In selected
cases, clear relationships between ligand structure and cata-
lyst activity and selectivity have been observed. For instance
for monodentate phosphorus-based ligands, both Moser et
al. and Van Leeuwen et al. have shown that a relationship
exists between ligand basicity and catalyst activity and
selectivity.1 For bidentate ligands the natural bite angle2 of
the ligand proved to be an important factor.3 The increased
activity and (in some cases) selectivity observed for ligands
(4) (a) Van Der Veen, L. A.; Keeven, P. H.; Schoemaker, G. C.; Reek,
J. N. H.; Kamer, P. C. J.; Van Leeuwen, P. W. N. M.; Lutz, M.; Spek,
A. L. Organometallics 2000, 19, 872. (b) Bohnen, H.; Herwig, J.; Joerg, A.;
Hoff, D.; Surm, S.; Van Leeuwen, P. W. N. M.; Bronger, R. P. J.; Stelzer, O.
DE 10225282, 2003. (c) Bohnen, H.; Herwig, J. WO 2002068371, 2002. (d)
Bohnen, H.; Herwig, J. WO 2002068369, 2002. (e) Ahlers, W.; Paciello, R.;
Vogt, D.; Hofmann, P. WO 2002083695, 2002. (f) Ahlers, W.; Wiebelhaus,
D.; Paciello, R.; Bartsch, M.; Baumann, R.; Vogt, D.; Hewat, A. WO
2002022261, 2002. (g) Rainer, P.; Ahlers, W.; Mackewitz, T.; Paciello, R.;
Volland, M. WO 2005039762, 2005. (h) Van Leeuwen, P. W. N. M.; Walczuk-
Gusciora, E. B.; Grimmer, N. E.; Kamer, P. C. J. WO 2005049537, 2005. (i)
Ahlers, W.; Roeper, M.; Hofmann, P.; Warth, D. C. M.; Paciello, R. WO
2001058589, 2001.
*Corresponding authors. E-mail: pcjk@st-andrews.ac.uk;
PvanLeeuwen@ICIQ.es.
(1) (a) Van Leeuwen, P. W. N. M.; Roobeek, C. F. J. Organomet.
Chem. 1983, 258, 343. (b) Moser, W. R.; Papile, C. J.; Brannon, D. A.;
Duwell, R. A.; Weininger, S. J. J. Mol. Catal. 1987, 41, 271.
(2) Casey, C. P.; Whiteker, G. T. Isr. J. Chem. 1990, 30, 299.
(3) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.;
J. A. Gavney, J.; Powell, D. R. J. Am. Chem. Soc. 1992, 114, 5535.
(5) (a) Van Der Veen, L. A.; Boele, M. D. K.; Bergman, F. R.; Kamer,
P. C. J.; Van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J.; Schenk,
H.; Bo, C. J. Am. Chem. Soc. 1998, 120, 11616. (b) Nettekoven, U.; Kamer,
P. C. J.; Widhalm, M.; Van Leeuwen, P. W. N. M. Organometallics 2000, 19,
4596. (c) Freixa, Z.; Van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans.
2003, 10, 1890.
r
pubs.acs.org/Organometallics
Published on Web 02/09/2010
2010 American Chemical Society

Article
Organometallics, Vol. 29, No. 5, 2010 1211
complexes that the angles around the phosphorus atom
decrease with increasing π-back-donation from the metal
to the ligand.13 Most cyclic ligand structures applied in the
hydroformylation reaction impose small valence angles
around the phosphorus atom and are therefore expected to
be less effective σ-donors than their acyclic analogues, while
favoring back-donation from the metal to the ligand. This
relationship between C-P-C angles, ligand electronic pro-
perties, and catalyst activity in the hydroformylation was
recently elegantly confirmed by Pringle and co-workers using
a range of isosteric and isoelectronic diphosphine ligands
containing phosphacyclic moieties of different ring size.8d
In addition to this angle effect, it has been postulated
that, for phosphine-based phosphacyclic ligands contain-
ing unsaturated moieties, conjugation between the phos-
phorus orbitals and the π-system of the ligand is responsible
for the unusually high activity in the hydroformylation
reaction observed for these ligands.8a,10b,c,14 The degree
of conjugation in this type of phosphine compound, how-
ever, remains heavily debated.15 Computational studies on
phosphole-type compounds have shown that the inter-
action between the phosphorus orbitals and the π-orbitals
on the cyclic compound is minimal due to the pyramidal
structure around the phosphorus atom.16 It should be noted
that systematic computational studies on the coordination
behavior of these compounds in transition metal complexes
are lacking.
We previously reported the synthesis and application
of wide bite angle ligands that contain dibenzophosphole
and phenoxaphosphine phosphacyclic moieties.4a,8c,14 The
resulting catalysts proved an order of magnitude more
active in the hydroformylation of terminal alkenes than
their acyclic analogues4a and could even be applied in the
hydroformylation of internal alkenes to linear aldehydes.8c
In order to further explore the unusual activity and selec-
tivity of hydroformylation catalysts modified by these
phosphacyclic ligand structures, we have synthesized sev-
eral bidentate ligands containing cyclic 2,8-dimethylphe-
noxaphosphine moieties, based on the xanthene, p-tolyl
ether, ferrocene, and benzene ligand backbones (Chart 1).
The “parent” diphenylphosphine-substituted ligands have
been shown to span a wide range of ligand bite angles,17
and all four have previously been applied in the hydro-
formylation reaction.5b,6a,8a,18 New phosphacyclic ligand
structures were generated by replacing one or both di-
phenylphosphine groups by 2,8-dimethylphenoxapho-
sphine moieties. The catalytic performance of the new
ligands in the hydroformylation of 1-octene was evaluated.
In addition, the coordination behavior of the bidentate
ligands under hydroformylation conditions was studied
using both high-pressure NMR and IR spectroscopic
techniques.
Figure 1. Equatorial-axial (left) and bisequatorial (right) coor-
dination modes of bidentate ligands in the resting state of the
hydroformylation catalyst.
ligands, the electronic properties of the donor moieties6 and
the steric bulk of the ligand5c,7 have a profound influence on
the activity and selectivity of the catalyst system.
Several authors have also reported that enclosing the
ligand’s phosphorus donor atom(s) inside a cyclic structure
leads to more stable, active, and/or regio- and enantioselec-
tive catalysts.8 The increased regio- and enantioselectivity of
catalysts containing phosphacyclic ligands compared to
catalysts containing noncyclic ligands are often attributed
to a more effective steric interaction between ligand and
substrate, caused by the rigid structure of the phosphacyclic
ligand. Indeed, in the rhodium-catalyzed asymmetric hydro-
formylation of prochiral olefins, where effective transfer of
chirality from the ligand to the substrate is crucial, the
highest enantioselectivities are obtained for catalysts con-
taining phosphacyclic ligand structures.9 The increase in
activity observed for many phosphacyclic ligands has been
ascribed to a lower basicity of the phosphorus moiety in the
cyclic ligand structure.8a,b,d,10 Early studies have suggested
that the lower basicity of cyclic phosphite or phosphorami-
dite structures compared to their noncyclic analogues is
caused by a less effective overlap between the phosphorus
and oxygen or nitrogen lone-pair orbitals.11 Alder and
co-workers have shown that the valence angles around
the phosphorus atom of noncyclic phosphine compounds
increase upon protonation.12 Consequently, cyclic structures
that impose large valence angles yielded highly basic phos-
phorus compounds. Orpen et al. showed by detailed analysis
of known crystal structures of a large number of metal
(6) (a) Unruh, J. D.; Christenson, J. R. J. Mol. Catal. 1982, 14, 19. (b)
Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.; Petrovich,
L. M.; Matter, B. A.; Powel, D. R. J. Am. Chem. Soc. 1997, 119, 11817. (c)
Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.; Matter, B. A.;
Powel, D. R. J. Am. Chem. Soc. 1999, 121, 63.
(7) (a) Carbo, J. J.; Maseras, F.; Bo, C.; Van Leeuwen, P. W. N. M.
J. Am. Chem. Soc. 2001, 123, 7630. (b) Landis, C. R.; Uddin, J. J. Chem.
Soc., Dalton Trans. 2002, 729.
(8) (a) Van Der Veen, L. A.; Kamer, P. C. J.; Van Leeuwen, P. W. N. M.
Organometallics 1999, 18, 4765. (b) Van Der Veen, L. A.; Kamer, P. C. J.; Van
Leeuwen, P. W. N. M. Angew. Chem., Int. Ed. 1999, 38, 336. (c) Bronger,
R. P. J.;Bermon, J. P.;Herwig, J.;Kamer, P. C. J.;VanLeeuwen, P. W. N. M.Adv.
Synth. Catal. 2004, 346, 789. (d) Haddow, M. F.; Middleton, A. J.; Orpen, A. G.;
Pringle, P. G.; Papp, R. J. Chem. Soc., Dalton Trans. 2009, 202.
(9) (a) Nozaki, K.; Sakai, N.; Nanno, T.; Higashijima, T.; Mano, S.;
Horiuchi, T.; Takaya, H. J. Am. Chem. Soc. 1997, 119, 4413. (b) Breeden,
S.; Cole-Hamilton, D. J.; Foster, D. F.; Schwarz, G. J.; Wills, M. Angew.
Chem., Int. Ed. 2000, 39, 4106. (c) Dieguez, M.; Pamies, O.; Claver, C.
Tetrahedron: Assymmetry 2004, 15, 2113. (d) Clark, T. P.; Landis, C. R.;
Freed, S. L.; Klosin, J.; Abboud, K. A. J. Am. Chem. Soc. 2005, 127, 5040.
(e) Axtell, A. T.; Cobley, C. J.; Klosin, J.; Whiteker, G. T.; Zanotti-Gerosa, A.;
Abboud, K. A. Angew. Chem., Int. Ed. 2005, 44, 5834. (f) Axtell, A. T.;
Klosin, J.; Abboud, K. A. Organometallics 2006, 25, 5003.
(13) (a) Orpen, A. G.; Connelly, N. G. Chem. Commun. 1985, 1310.
(b) Orpen, A. G.; Connelly, N. G. Organometallics 1990, 9, 1206.
(14) Bronger, R. P. J.; Kamer, P. C. J.; Van Leeuwen, P. W. N. M.
Organometallics 2003, 22, 5358.
(10) (a) Tanaka, M.; Ikeda, Y.; Ogata, I. Chem. Lett. 1975, 1115. (b)
Hayashi, T.; Tanaka, M.; Ogata, I. J. Mol. Catal. 1979, 6, 1. (c) Hayashi, T.;
Tanaka, M.; Ikeda, Y.; Ogata, I. Bull. Chem. Soc. Jpn. 1979, 52, 2605. (d)
Hobbes, C. F.; Knowles, W. S. J. Org. Chem. 1981, 46, 4422.
(11) (a) Verkade, J. G. Coord. Chem. Rev. 1972, 9, 1. (b) Verkade,
J. G. Bioinorg. Chem. 1974, 3, 165. (c) Albright, J. O.; Tanzella, F. L.;
Verkade, J. G. J. Coord. Chem. 1976, 6, 225.
(15) Nyulaszi, L. Chem. Rev. 2001, 101, 1229.
(16) Schleyer, P. v. R.; Freeman, P. K.; Jiao, H.; Goldfuss, B. Angew.
Chem., Int. Ed. Engl. 1995, 34, 337.
(17) Dierkes, P.; Van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton
Trans. 1999, 1519.
(18) (a) Kranenburg, M.; Van Der Burgt, Y. E. M.; Kamer, P. C. J.;
Van Leeuwen, P. W. N. M. Organometallics 1995, 14, 3081. (b) Bianchini,
C.; Oberhauser, W.; Orlandini, A.; Gianelli, C.; Frediani, P. Organometallics
2005, 24, 3692.
(12) Alder, R. W.; Butts, C. P.; Orpen, A. G.; Read, D.; Oliva, J. M.
J. Chem. Soc., Perkin Trans. 2 2001, 282.

1212 Organometallics, Vol. 29, No. 5, 2010
Zuidema et al.
Chart 1. Ligands Used in This Study
Scheme 1. Synthesis of Benzene-Based Ligandsa
a (i) n-Butyllithium, diethyl ether, -80 ꢀC; (ii) 2,8-dimethylphenoxaphosphine chloride, THF, -80 ꢀC; (iii) n-butyllithium, ether/pentane/THF,
-120 ꢀC; (iv) bis(diethylamino)phosphine chloride, ether/pentane/THF, -120 ꢀC; (v) PCl3, 50 ꢀC; (vi) 2,20-dilithio-p-tolyl ether, diethyl ether, -80 ꢀC.
2. Results
formation of a phenoxaphosphine structure during the second
bromide-lithium exchange reaction. Introducing the phenoxa-
phosphine moiety first prevented the occurrence of this side
reaction.
2.1. Ligand Synthesis. The synthesis of ligands 1a-1c was
previously reported by Van der Veen et al.4a Symmetrical
p-tolyl ether- and ferrocene-based ligands 2c and 3c were
synthesized via direct dilithiation of the backbone or by
bromide-lithium exchange of the dibrominated backbone,
followed by a reaction with two equivalents of 2,8-dimethyl-
phenoxaphosphinous chloride. For the ligands containing two
different phosphorus moieties based on these backbones, 2b
and 3b, the two phosphine donor groups were introduced
sequentially, starting from dibromo-substituted ligand back-
bones. The synthesis of 1-bromo-10-diphenylphosphinoferro-
cene was previously reported by Butler and co-workers.19 Also
for 2,20-dibromo-p-tolyl ether, bromide-lithium exchange
of a single bromide moiety occurred selectively. In contrast
to the synthesis of ferrocene ligand 3b, initial introduction
of the phenoxaphosphine moiety, generating 2-bromo-
20-(2,8-dimethylphenoxaphosphino)-p-tolyl ether (8), proved
crucial for obtaining ligand 2b selectively. Initial introduction
of the diphenylphosphine group in the synthesis of 2b led to
cleavage of one of the phosphorus-phenyl bonds and the
Dissymmetric benzene-based ligand 4b was synthesized
directly from previously reported o-(bromophenyl)diphenyl-
phosphine.20 Symmetric ligand 4c was synthesized in a
sequence of steps (Scheme 1), starting from 1,2-dibromo-
benzene. Direct reaction of 1-lithio-2-bromobenzene and
2,8-dimethylphenoxaphosphinous chloride at low tempera-
ture resulted in the formation of significant amounts
of unidentified side products. Reetz and co-workers have
reported the synthesis of 1,2-bis(dichlorophosphino)ben-
zene, starting from 1,2-dibromobenzene, butyllithium, and
bis(diethylamino)phosphinous chloride.21 Using a similar
approach, we synthesized 2-bromophenyl-bis(diethylamino)-
phosphine 5. Reaction of 5 with butyllithium and 2,8-dimethyl-
phenoxaphosphine chloride yielded diphosphine 6. The bis-
(diethylamino)phosphine moiety was subsequently converted
to a phosphonous dichloride moiety using PCl3. In the final
(20) Machnitzki, P.; Nickel, T.; Stelzer, O.; Landgrafe, C. Eur. J.
Inorg. Chem. 1998, 1029.
(19) Butler, I. R.; Davies, R. L. Synthesis 1996, 11, 1350.
(21) Reetz, M. T.; Moulin, D.; Gosberg, A. Org. Lett. 2001, 3, 4083.

Article
Organometallics, Vol. 29, No. 5, 2010 1213
Figure 2. Two separate views of the crystal structure of ligand 4c. Displacement ellipsoids are drawn at the 50% probability level.
Symmetry operation a: 1-x, y, 0.5-z.
Table 1. Selected Distances and Angles in the Crystal Structure of Ligand 4c
˚
distances (A)
angles (deg)
interplanar angles (deg)
Ph2-Ph3 2.40(6)
P1-C11
P1-C21
P1-C31
P1-P1a
P1-O1
1.8604(11)
1.8140(12)
1.8206(12)
3.2553(6)
C11-P1-C21
C11-P1-C31
C21-P1-C31
C26-O1-C36
100.83(5)
100.55(5)
98.38(5)
122.60(10)
3.2328(13)
step, this species was reacted with 2,20-dilithio-p-tolyl ether,
forming ligand 4c. For this ligand, crystals suitable for X-ray
diffraction were obtained. The elucidated structure is shown in
Figure 2.
Table 2. Calculated Natural Bite Angles, Flexibility Ranges, and
Solid Angles for Ligands 1-4
ligand
βn (deg)a
flexibility range (deg)a,b
solid angle (%)c
1a
1b
1c
2a
2b
2c
3a
3b
3c
4a
4b
4c
113.5
119.2
120.6
108.0
106.6
106.8
113.4
109.7
109.8
86.0
100-131
104-137
104-136
93-124
94-123
94-124
94-131
94-125
95-123
78-93
50.6
49.1
45.8
46.3
45.1
43.3
46.4
46.6
47.4
37.5
38.3
38.1
4c crystallizes in the centrosymmetric space group C2/c,
and the molecule has an exact, crystallographic 2-fold
symmetry. The phosphorus-phosphorus distance observed
in the crystal structure of 4c corresponds to a natural bite
˚
angle of 89.4ꢀ (at a M-P distance of 2.315 A), considerably
larger than the value of 83ꢀ reported for 1,2-bis(diphenyl-
phosphino)benzene (4a).17 Van Leeuwen et al. have pre-
viously observed a similar increase in bite angle going from
ligand 1a to ligand 1c and attributed this to steric repulsion
between the two phosphacyclic moieties in ligand 1c.14 The
geometry of 4c in the crystal structure minimizes both the
steric repulsion between the two rigid phosphacyclic struc-
tures and the electronic repulsion between the lone pairs of
the two phosphorus atoms. The two phosphacyclic moieties
are almost planar, and the geometry around the phosphorus
atoms is trigonal pyramidal. The C-P-C angle in the
phosphacyclic ring is somewhat smaller than the other two
C-P-C angles in the structure.
85.6
84.7
78-93
78-92
a Calculated using the method of Casey and Whiteker.2 b Accessible
bite angle range within 3 kcal mol-1 excess strain energy of βn.
c Calculated using the method of White and co-workers,22 as implemen-
ted in the Steric program.23
angle methodology introduced by White et al.22 The calcu-
lated bite angles and solid angles are reported in Table 2.
As expected, the different ligand backbones span a wide
range of natural bite angles. The calculated bite angles
and flexibility ranges are in line with previous observa-
tions that xanthene-based ligands 1 preferentially adopt
The natural bite angles (βn) of ligands 1-4 were deter-
mined using molecular mechanics-based calculations on
rhodium(diphosphine) complexes following the methodo-
logy developed by Casey and Whiteker.2 The optimized
rhodium(diphosphine) structures were subsequently used
to evaluate the steric properties of the ligands using the solid
(22) (a) White, D.; Tavener, B. C.; Leach, P. G. L.; Coville, N. J.
J. Organomet. Chem. 1994, 478, 205. (b) White, D.; Tavener, B. C.; Coville,
N. J.; Wade, P. W. J. Organomet. Chem. 1995, 495, 41.
(23) Taverner, B. C. Steric 1.2.8, www.gh.wits.ac.za/craig/steric.

1214 Organometallics, Vol. 29, No. 5, 2010
Zuidema et al.
Table 3. Results of the Hydroformylation of 1-Octene Using
Ligands 1-4a
ligand
TOFb
l/b
% isomersc
% n-aldehyded
403
163
346
812
38
77
34
79
80
2.8
44
31
22
7.2
3.2
3.0
5.3
1.9
1.5
0.9
0.6
77
17
93
93
86
86
74
74
80
65
56
46
35
1a
1b
1c
2a
2b
2c
3a
3b
3c
4a
4b
4c
5.0
5.5
9.8
2.3
2.5
2.6
4.8
1.6
5.0
3.1
7.0
Figure 3. Complexes formed under hydroformylation condi-
tions.
224
32
89
previously investigated extensively by both Lazzaroni et al.24
and Garland et al.25 Also under our experimental conditions,
high hydroformylation rates are observed for this catalyst
system, accompanied by high rates of alkene isomerization.
At low conversion of 1-octene, the catalyst system shows a
surprisingly high l/b ratio of 2.8 for the formation of
aldehyde products. At higher octene conversions, we observe
a decrease in the regioselectivity of the catalyst system due to
the hydroformylation of alkene isomerization products.
Xanthene-based ligands 1a8a and 1c8c and ferrocene-based
ligand 3a5b,6a,18b were previously applied as ligands in the
rhodium-catalyzed hydroformylation of 1-alkenes. The
xanthene-based ligands yielded highly active and regioselec-
tive catalyst systems, while ferrocene-based ligand 3a was
shown to be significantly less active and regioselective in the
hydroformylation of 1-alkenes. Also in this study, the most
active catalyst systems are obtained for xanthene-based
ligands 1. Ferrocene-based ligands 3 yielded moderately
active catalyst systems, and p-tolyl ether-based ligands 2
and benzene-based ligands 4 formed the least active hydro-
formylation catalysts. As a general trend, a decrease in bite
angle (and solid angle) of the diphosphine ligand leads to a
lower regioselectivity in the hydroformylation of 1-octene.
Catalysts modified by benzene-based ligands 4a and 4b show
a remarkable preference for the formation of the branched
aldehyde product in the hydroformylation of 1-octene.
For all four ligand backbones, introduction of phospha-
cyclic moieties leads to an increase in catalytic activity,
accompanied by a decrease in regioselectivity toward the
linear aldehyde product and an increased formation of
octene isomers. P-tolyl ether-based ligand 2c, however,
exhibits a significantly lower catalytic activity than both
ligand 2a and 2b. Van der Veen and co-workers previously
reported similar observations for the related bis(phenoxa-
phosphino)-p-tert-butyl-diphenyl-ether ligand4a and attri-
buted this to the rigid equatorial-axial coordination of the
ligand in the resting state of the catalyst A-ea (Figure 3).
Surprisingly, benzene-based ligand 4c proved to be comple-
tely inactive in the hydroformylation of 1-octene under the
experimental conditions.
a T = 80 ꢀC, pCO = pH2 = 10 bar, [Rh] = 1.0 mM, Rh:L:sub-
strate = 1:10:637. b Turnover frequency in mol aldehyde (mol Rh)-1
h-1 c Percentage octene isomers in all products formed in the reaction.
.
d Percentage nonanal in all products formed in the reaction.
a bisequatorial coordination mode in trigonal-bipyramidal
complexes.4a,8a,14 For benzene-based ligands 4, this coordi-
nation mode is predicted to be inaccessible, and consequently
these ligands will prefer an equatorial-axial coordination
mode. Flexible ligands 2 and 3 are predicted to be capable
of accommodating both coordination modes in trigonal-
bipyramidal complexes. The calculated solid angles follow
the observed bite angle trend; wide bite angle ligands 1 are
predicted to induce the most steric hindrance around the
metal center in transition metal complexes. For ligands
having the same backbone, the differences in bite angle are
much smaller. It was already reported that on going from
ligand 1a to 1c, the natural bite angle of the ligand increases
due to steric repulsion between the two rigid phosphacyclic
moieties in ligand 1c.14 Interestingly, the calculated solid
angles in xanthene-based ligands 1 and p-tolyl ether-based
ligands 2 decrease when the diphenylphosphine moieties of
the ligands are replaced by phenoxaphosphine moieties,
suggesting that these phosphacyclic moieties impose less
steric hindrance around the metal center than acyclic diphe-
nylphosphine moieties. For ligands 3 and 4 both the calcu-
lated bite angles and solid angles remain nearly constant.
Steric differences between the phenoxaphosphine and diphe-
nylphosphine moieties in these ligands are relatively small.
2.2. Hydroformylation of 1-Octene. The catalytic perfor-
mance of rhodium diphosphine catalysts containing ligands
1-4 in the hydroformylation of 1-octene was investigated.
The hydroformylation reactions were performed at 80 ꢀC
under an atmosphere of 20 bar of CO/H2 (1:1) using a
1.0 mM solution of rhodium catalyst prepared in situ from
Rh(CO)2(acetylacetonate) and 10 equivalents of diphos-
phine ligand. The production of nonanal, 2-methyloctanal,
and octene isomers was determined using gas chromatogra-
phy. Turnover frequencies and regioselectivities were deter-
mined as average values at 15-20% alkene conversion. The
results are shown in Table 3.
2.3. Complex Characterization. For all ligands, the com-
plexes formed under hydroformylation conditions were stu-
died using high-pressure variable-temperature NMR and
in situ IR spectroscopy. In order to facilitate the comparison
between the hydroformylation results and the spectroscopic
data, the in situ IR spectroscopic experiments were per-
formed at actual catalytic rhodium (1.0 mM) and ligand
(10.0 mM) concentrations. Due to the relatively low sensi-
tivity of the NMR technique, the NMR spectroscopic studies
required significantly higher rhodium (12 mM) and ligand
(26 mM) concentrations. Upon reacting Rh(acetylacetonate)-
(CO)2 with an excess of ligand under an atmosphere of CO/H2
The catalytic performance of the unmodified rhodium
catalyst in the hydroformylation of terminal alkenes was
(24) Lazzaroni, R.; Pertici, P.; Bertozzi, S.; Fabrizi, G. F. J. Mol.
Catal. 1990, 58, 75.
(25) (a) Feng, J. H.; Garland, M. Organometallics 1999, 18, 417. (b)
Liu, G. W.; Volken, R.; Garland, M. Organometallics 1999, 18, 3429.
(26) (a) Horvath, I. Organometallics 1986, 5, 2333. (b) Van Rooy, A.;
Kamer, P. C. J.; Van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J.; Veldman,
N.; Spek, A. L. Organometallics 1996, 15, 835.

Article
Organometallics, Vol. 29, No. 5, 2010 1215
Table 4. High-Pressure NMR Spectroscopic Dataa
L
structure
δ31P (ppm)
JRh,P (Hz)
JP,P (Hz)
δ1H (ppm)
|JP,H| (Hz)
JRh,H (Hz)
1a
1b
A
A
19.9(d)
-24.8(dd)
15.9(dd)
-22.9(d)
26.6(d)
-19.6(dd)
29.2(dd)
-20.0(d)
-37.4(m)
31.7(d)
-11.5(bd)
33.9(bd)
-8.1(d)
-6.6(m)
-5.7(m)
62.7(d)
151
150
141
151
117
141
66
128
159; 9
124
133
109
130
159
165
118
143
136; 3
143
146
152
-8.4(dt)
-8.9(bd)
17
7
11
73
73
1c
2a
2b
A
A
A
-9.1(bt)
-8.6(dt)
-9.2(ddd)
12
47
11
11
23
23
116; -19 b
2c
A
C
A
A
-9.0(dt)
33
8
67
3a
3b
-8.9(dt)
-9.3(dd)
42
88
9
8
3c
4a
A
D
-9.1(dt)
-9.0(bd)
82
-
9
11
127
127
A
B
C
B
-8.4(dt)
-9.6(bm)
61
10
59.2(d)
43.6(m)
23.3(ddd)
59.6(ddd)
33.5(d)
54
53; 46
53; 46
4b
4c
-9.5(bdt)
-9.2(bp)
40
10
B
8
a Conditions: T = 0 ꢀC, pCO = pH2 = 10 bar, [Rh] = 12.0 mM. [Rh]:[L] = 1:2.2. b The sign of these coupling constants was not determined, but
previous studies on similar complexes suggest they have opposite signs.3,5b,6c,26
Table 5. High-Pressure Infrared Spectroscopic Dataa
spectra and the low intensities of the carbonyl absorptions
corresponding to A-ea in the IR spectra.
A-ee
A-ea
p-Tolyl ether-based ligands 2a and 2b also showed exclu-
sive formation of structure A under hydroformylation con-
ditions. The NMR data collected for ligand 2a showed a
ligand
ν1
ν3
ν2
ν4
1a
1b
1c
2a
2b
2c
3a
3b
3c
4a
4b
4c
2039(s)
2040(vs)
2043
1974(s)
1977(vs)
1983
1996(w)
1997(vw)
1949(w)
1953(vw)
2
considerably larger JP,H coupling constant than the data
collected for ligand 1a. This suggests that the equilibrium
between A-ee and A-ea is shifted toward the equatorial-axial
isomer A-ea. This was confirmed in the high-pressure IR
experiments involving ligand 2a, where the main absorptions
were attributed to the equatorial-axial isomer of A. Also for
ligand 2b, the main signals observed in the carbonyl region of
the IR spectrum were attributed to the equatorial-axial
2049(m)
1979(m)
1981(vw)
1998(s)
1993(vs)
1950(s)
1951(vs)
2022(m)
2043(w)
2040(m)
1971(m)
1985(w)
1982(m)
1997(s)
1999(s)
1999(s)
1956(s)
1959(s)
1954(s)
1
isomer A-ea. Both the H and 31P{1H} NMR data were
consistent with an equatorial-axial coordination mode of the
ligand in structure A. Proton-coupled 31P NMR spectra
revealed that the hydride coupled predominantly with the
phosphorus signal at 29 ppm. On the basis of the phosphorus
chemical shifts observed for ligands 2a and 2c, this phos-
phorus signal is attributed to the diphenylphosphine moiety
of ligand 2b. Clearly, the diphenylphosphine moiety of
ligand 2b is coordinated in the apical position of A-ea, trans
toward the hydride moiety, and the phosphacyclic moiety is
coordinated in the equatorial plane of structure A-ea.
The high-pressure NMR experiments involving ligand 2c
showed not only the formation of structure A but also the
formation of significant amounts of rhodium dimer C, as
well as smaller quantities of unidentified complexes. The
high-pressure IR experiments, however, did not show the
expected signals for structures A and C. Instead, several
absorptions in the terminal carbonyl region were observed,
which could not be assigned on the basis of the available
spectroscopic data.
As was previously reported for other ferrocene-based
diphosphine ligands,5b reaction of Rh(acac)(CO)2 with an
excess of 3c under an atmosphere of H2/CO at the high
concentrations necessary for NMR spectroscopy yielded a
mixture of dimeric structure D and monomeric structure A.
Due to the similar chemical shifts of all three phosphine
moieties in structure D, highly complex second-order
31P{1H} NMR spectra are obtained for this complex at 0 ꢀC.
a Conditions: T = 40 ꢀC, pCO = pH2 = 10 bar, [Rh] = 1.0 mM.
[Rh]:[L] = 1:10, in cm-1
.
(1:1), the formation of several rhodium complexes was observed.
The observed structures are shown in Figure 3. The correspond-
ing NMR data are summarized in Table 4, and the IR data are
summarized in Table 5.
Despite the excess of ligand used, for all three xanthene-
based ligands 1a-1c, only the formation of the well-known
HRh(L)(CO)2 structure A was observed. The coordination
behavior of ligands 1a and 1c in structure A was previously
reported.5a,8a,b For ligand 1a, a mixture of structure A-ee
and A-ea was observed in the high-pressure infrared experi-
ments. Even at -90 ꢀC, these two structures are in rapid
equilibrium on the NMR time scale, yielding a single doublet
in the 31P NMR spectra and an averaged triplet of doublets
1
for the hydride in the H NMR spectra. Ligand 1c coordi-
nates exclusively in a bisequatorial fashion in structure A.
No signals corresponding A-ea were observed in the high-
pressure IR spectra for this ligand system. Furthermore, the
measured phosphine-hydride coupling constant was small,
indicating a mutual cis arrangement of both phosphine
moieties and the hydride ligand. Dissymmetric ligand 1b
also exhibited predominantly bisequatorial coordination in
2
structure A, as was evidenced by the small JP,H coupling
1
constant observed in the hydride signal in the H NMR

1216 Organometallics, Vol. 29, No. 5, 2010
Zuidema et al.
Upon cooling the NMR sample, the signals broaden, indicative
of fluxional behavior in this complex. Unfortunately, the limit
of slow exchange could not be attained. Because of the absence
of a coupling between the hydride signal and phosphine signals,
and for steric reasons, we favor a structure of D in which the
three phosphine moieties are predominantly coordinated in the
equatorial plane of the complex. The observed fluxional beha-
vior might indicate, however, that a rapid equilibrium exists
between this major isomer and a minor one in which one of the
phosphine moieties is coordinated in the axial position of the
complex, trans to the hydride. Surprisingly, no formation of
dimer Dwas observed for ligand 3a and dissymmetric ligand 3b.
Similar to ligand 2b, ligand 3b showed the selective formation of
A-ea, with the diphenylphosphine moiety predominantly occu-
pying the axial position trans toward the hydride. For all three
ferrocene-based ligands studied here, no signals originating
from dimeric structure Dcould be observed in the high-pressure
IR experiments. At the low rhodium concentration used in
these IR experiments, the formation of monomeric structure A
is strongly favored over the formation of dimer D.
Under the high concentration conditions of the high-
pressure NMR experiments, benzene-based ligand 4a formed
three major species. Next to species A, dimer C and mono-
meric bischelate structure B were formed in significant
quantities. The synthesis of bischelate complexes B has
previously been reported for several ligands,27 including
ligand 4a.28 James and co-workers showed that structure B
containing either 1,2-bis(diphenylphosphino)ethane or 1,3-
bis(diphenylphosphino)propane ligands is quantitatively
converted to dimer C under hydroformylation conditions,
presumably via compound A.29 Also for ligand 4a, we
observed the slow conversion of B to dimer C in the high-
pressure NMR experiments.
at -90 ꢀC. For symmetrical bischelate complexes B contain-
ing either ligand 4a or 4c one sharp doublet was observed in
the 31P{1H} NMR spectra, even at low temperatures. For
these ligands, all four phosphine moieties are equivalent due
to fast fluxional processes within the complex.
In contrast to the dimeric structures observed for ligands 2
and 3, which are formed only at high catalyst concentrations,
the absence of strong carbonyl absorptions in the in situ IR
spectra suggests that for ligands 4a-4c bischelate species B is
also the main species formed under catalytic conditions.
Unfortunately, due to the low catalyst concentration and
the relatively low resolution in the high-pressure IR experi-
ments, the weak rhodium-hydride vibration band could not
be observed for these compounds.
3. Discussion
The observed coordination behavior of ligands 1-4 in
rhodium structures A-D under an atmosphere of CO/H2 is
consistent with previous high-pressure spectroscopic studies.
Depending on the natural bite angle of the ligand, a bis-
equatorial and/or equatorial-axial coordination mode of the
diphosphine ligand in structure A is observed. Xanthene-
based ligands 1 show predominantly a bisequatorial coordi-
nation mode in species A. As was previously reported,
introduction of 2,8-dimethylphenoxaphosphine moieties in
ligands 1 led to an increase in the ratio A-ee:A-ea.14 This was
attributed to the increased bite angle of 1c compared to 1a
caused by steric repulsion between the two rigid phospha-
cyclic moieties in 1c. We also observe an increase in bite angle
in the crystal structure of benzene-based ligand 4c compared
to reported values for ligand 4a,17 but not in the calculated
bite angles of ligands 2-4 (Table 2).
The increase in the ratio of A-ee:A-ea upon introduction of
phosphacyclic moieties in the ligand structure is not ob-
served for p-tolyl ether and ferrocene-based ligands 2 and 3,
which is in agreement with the calculated natural bite angles
for these ligands. On the basis of the high flexibility of the
ferrocene and p-tolyl ether backbones compared to the
xanthene and benzene backbones, one would expect the bite
angles of ligands 2 and 3 to be more sensitive to steric
interactions between the two phosphine moieties than the
bite angles of ligands 1 and 4. Probably, the high flexibility of
the ferrocene and p-tolyl ether backbones permits reorienta-
tion of the rigid phenoxaphosphine moieties in structure A-
ea and so minimizes steric interactions between the two
phosphine moieties of the ligand.
The NMR experiments show that the predominant coordi-
nation mode of ligands 2b and 3b in species A is equatorial-
axial, their diphenylphosphine moieties coordinating in the
axial position. Previous studies on equatorial-axial coordina-
ting dissymmetric diphosphine ligands have shown that there is
a strong preference for the coordination of the least basic
moiety of the diphosphine ligand in the equatorial plane
of complex A-ea,6c in accordance with simple orbital pre-
ferences.30 This confirms that 2,8-dimethylphenoxaphosphine
moieties are indeed less basic than diphenylphosphine moieties.
The lower basicity of the phenoxaphosphine moieties com-
pared to the diphenylphosphine moieties in ligands 1-3 is also
reflected by the increase in carbonyl stretching frequencies
of species A-ee when the diphenylphosphine moieties of
the ligands are replaced by phosphacyclic moieties (Table 5).
Ligands 4b and 4c formed bischelate structure B selectively
under the experimental conditions. For neither of them was
formation of dimer C observed, not even after prolonged
reaction times. For dissymmetric ligand 4b, the hydride
signal attributed to B was a doublet of triplets, consistent
with fast fluxional processes within the complex.27a The
coupling pattern suggests that the hydride couples with a
single donor atom of each diphosphine ligand coordinated to
the metal. Proton-coupled 31P NMR spectra were collected,
and these revealed that the hydride couples solely with
the phosphorus signal at 23.3 ppm. On the basis of the
chemical shifts of the phosphine moieties in complexes B(4a)
and B(4c), this signal was attributed to the phosphacyclic
moiety of ligand 4b. Upon cooling of the reaction mixture to
-60 ꢀC, both phosphine signals coalesced simultaneously.
At -80 ꢀC, four broad signals are observed, indicating that
in the absence of fast fluxional processes all four phosphorus
moieties in complex B(4b) become nonequivalent. The limit
of slow exchange, however, could not be attained, not even
(27) (a) Meakin, P.; Muetterties, E. L.; Jesson, J. P. J. Am. Chem. Soc.
1972, 94, 5271. (b) Ball, R. G.; James, B. R.; Mahajan, D.; Trotter, J. Inorg.
Chem. 1981, 20, 254. (c) Fryzuk, M. D. Can. J. Chem. 1983, 61, 1347. (d)
Trzeciak, A. M.; Ziolkowski, J. J.; Aygen, S.; Eldik, R. v. J. Mol. Catal.
1986, 34, 337. (e) Castellanos-Paez, A.; Castillon, S.; Claver, C.; Van
Leeuwen, P. W. N. M.; De Lange, W. G. J. Organometallics 1998, 17,
2543. (f) Rio, I. d.; De Lange, W. G. J.; Van Leeuwen, P. W. N. M.; Claver, C.
J. Chem. Soc., Dalton Trans. 2001, 1293. (g) Raebiger, J. W.; DuBois, D. L.
Organometallics 2005, 24, 110.
(28) Price, A. J.; Ciancanelli, R.; Noll, B. C.; Curtis, C. J.; DuBois,
D. L.; DuBois, M. R. Organometallics 2002, 21, 4833.
(29) James, B.; Mahajan, D.; Rettig, S. J.; Williams, G. M. Organo-
metallics 1983, 2, 1452.
(30) Rossi, A. R.; Hoffmann, R. Inorg. Chem. 1975, 14, 365.

Article
Organometallics, Vol. 29, No. 5, 2010 1217
These observations are in line with previous studies by our
group on xanthene-based ligands 1a and 1c.4a,14 It should be
noted that the bite angles of these two ligands differ signifi-
cantly and that differences in the effective donor and acceptor
properties of the two ligands might therefore arise from
differences in bite angle. For the highly flexible ferrocene-based
ligands 3, we do not observe this change in bite angle upon
introduction of 2,8-dimethylphenoxaphosphine moieties, and
the increase in carbonyl stretching frequencies of species A-ee
can only be attributed to the lower basicity of the phosphacyclic
moieties in ligand 3c compared to the diphenylphosphine
moieties of ligand 3a. Because of the presence of two carbonyl
ligands and only a single phosphine moiety in the equatorial
plane of complex A-ea, back-donation from the metal to
the carbonyl ligands is expected to be less substantial in
structure A-ea than in structure A-ee. Consequently, the car-
bonyl stretching frequencies of complex A-ea are less sensitive
to variations in the basicity and π-acidity of the phosphine
species.
For (small bite angle) benzene-based ligands 4, the forma-
tion of species A was not observed under hydroformylation
conditions. Instead, the predominant species under catalytic
conditions is bischelate compound B. The geometry of
structure B has previously been described as pseudotetrahe-
dral, because the phosphorus ligands form a tetrahedral
arrangement around the metal. The hydride moiety occupies
one of the faces of this tetrahedron, resulting in a distorted
trigonal-bipyramidal structure for B.27a For symmetrical
ligands 4a and 4c all four phosphorus moieties interchange
rapidly on the NMR time scale by movement of the hydride
from one face to another via the edges of the tetrahedron.
Surprisingly, similar studies on the coordination behavior of
bidentate ligands containing two distinct donor moieties in
structure B have not been reported. The high-pressure NMR
data collected for complex B(4b) suggest that the hydride is
predominantly located trans toward the two phosphacyclic
moieties of the complex. This is in sharp contrast with the
coordination mode of ligands 2b and 3b in related species A,
where the most basic (diphenylphosphine) donor group is
located trans toward the hydride. Here, the electronic pre-
ference for coordination of the least basic phosphine moiety
in the equatorial plane of trigonal-bipyramidal rhodium
structure A is overridden by steric interactions between the
two diphosphine ligands in sterically congested complex
B(4b). Although we have shown that the introduction of
phenoxaphosphine moieties in the ligand structure increases
the steric hindrance between the two donor groups within the
ligand (vide supra), the steric hindrance of the diphosphine
ligand toward the other ligands in transition metal com-
plexes actually decreases due to the planar structure of the
phenoxaphosphine moiety (Table 2). Consequently, the
diphenylphosphine moieties of both diphosphine ligands in
complex B(4b) are preferentially coordinated in the equa-
torial plane of the distorted trigonal-bipyramidal complex,
minimizing steric repulsion between the two bidentate
ligands in this highly congested structure.
the rate of hydroformylation upon introduction of phos-
phacyclic moieties. Both the decreased electron density at
the metal center and reduced steric hindrance around the
rhodium center induced by these cyclic structures should
favor the coordination of the more bulky and more electron-
donating alkene ligand over the electron-withdrawing and
nonbulky carbon monoxide ligand and, so, will shift the
equilibrium that exists between the carbon monoxide com-
plex and the alkene complex toward the latter. Moreover, it
has been shown that electron-withdrawing ligands facilitate
the subsequent rate-limiting hydride migration step.31
Therefore, both the decreased donor strength and reduced
steric bulk of these cyclic ligands enhance the overall activity
of the rhodium catalyst in the hydroformylation reaction.
Using different isosteric diphosphine ligands several
authors have previously shown that the linear over branched
ratio of the aldehyde products formed in the reaction is
influenced by the electronic properties of the ligands used in
the reaction. Casey and co-workers showed that decreasing
the σ-donating properties of the equatorial phosphine moi-
eties in structure A leads to an increase in the l/b ratio, while
decreasing the σ-donating properties of the axial phosphorus
moiety in structure A-ea decreases the l/b ratio.6b,c We have
shown that for wide bite angle ligands the percentage of
linear aldehyde formed is largely independent of the electro-
nic properties of the diphosphine ligand. The observed
increase in l/b ratio for less basic diphosphine ligands was
attributed to an increase in the rate of alkene isomerization
during the hydroformylation process.5a,14 For ligands 1-4,
however, we observe a decrease in both the l/b ratio and the
percentage of linear aldehyde formed when the less basic
phosphacyclic moieties are used instead of diphenylpho-
sphine moieties, despite the increase in the rate of alkene
isomerization. We attribute this decrease in regioselectivity
to the large steric differences between the diphenylphosphine
and 2,8-dimethylphenoxaphosphine moieties, which out-
weigh the smaller electronic differences between the two
moieties.
Direct comparison of the catalytic activities of benzene-
based ligands 4 to those of the other ligands used in this study
is hampered by the quantitative formation of bischelate
species B under the reaction conditions. For these ligands,
the generation of a catalytically active species will require
dissociation of one of the diphosphine ligands from species
B. Therefore, the catalytic activity of catalyst systems con-
taining ligands 4 is highly dependent on the stability of
species B and the rhodium and ligand concentrations under
experimental conditions. The high-pressure NMR spectro-
scopic results show that, while for ligand 4a species B is
converted to catalytically active species A and dimer C,
compound B is stable for ligands 4b and 4c under identical
conditions. The stability of species B relative to catalytically
active species A is enhanced by both the lower basicity and
reduced steric hindrance of the phenoxaphosphine-contain-
ing ligands 4b and 4c compared to the diphenylphosphine-
based ligand 4a. It is therefore not surprising that for ligand
4c, present in excess, the high stability of structure B under
H2/CO atmosphere leads to an inactive catalyst system for
hydroformylation of 1-octene.
In some instances, the differences in catalytic activity and
selectivity can be correlated to the results obtained from the
in situ spectroscopic studies. Since xanthene-based ligands
1a-1c all exhibit mainly bisequatorial coordination in struc-
ture A, it is not surprising that dissymmetric ligand 1b shows
rates of aldehyde formation and alkene isomerization that
are an average of the values obtained for symmetrical ligands
1a and 1c. Also for ligands 2 and 3 we observe an increase in
(31) (a) Zuidema, E.; Daura-Oller, E.; Carbo, J. J.; Bo, C.;
Van Leeuwen, P. W. N. M. Organometallics 2007, 26, 2234. (b) Zuidema,
E.; Escorihuela, L.; Echelsheim, T.; Carbo, J. J.; Bo, C.; Kamer, P. C. J.;
Van Leeuwen, P. W. N. M. Chem.;Eur. J. 2008, 14, 1843.

1218 Organometallics, Vol. 29, No. 5, 2010
Zuidema et al.
4. Conclusions
co-workers. Solid angles were calculated by projecting the
ligands onto a sphere centered at the rhodium atom. The solid
angles were expressed as the percentage of the sphere covered by
the ligand projection. The solid angle calculations were per-
formed using the Craig algorithm, which takes into account
multiple overlaps in the projection of the ligand onto the sphere.
5.3. Ligand Synthesis. 1-(Bis(diethylamino)phosphino)-2-
bromobenzene (5). To a solution of 1 mL of 1,2-dibromobenzene
(8.3 mmol) in 100 mL of a 1:1:1 mixture of THF, pentane, and
diethyl ether was added dropwise 3.3 mL of a 2.5 M solution of
n-butyllithium (8.25 mmol) at -120 ꢀC. The solution was stirred
for 15 min. Subsequently, 1.75 g (8.3 mmol) of bis(diethyl-
amino)phosphine chloride dissolved in 15 mL of a 1:1:1 mixture
of THF, pentane, and diethyl ether was added dropwise to the
reaction mixture. The resulting mixture was stirred overnight,
allowing the temperature of the mixture to rise to room tem-
perature. The solvents were removed in vacuo. The resulting
yellow oil was dissolved in dichloromethane and purified using
flash column chromatography over neutral alumina. Evapora-
tion of the solvent yields 1.35 g (4.1 mmol, 49%) of a colorless,
We have synthesized a series of new bidentate phosphine
ligands based on 2,20-p-tolyl ether, 1,10-ferrocene, and 1,2-
benzene ligand backbones containing one or two 2,8-dimethyl-
phenoxaphosphine phosphacyclic moieties. The ligands were
employed in the hydroformylation of 1-octene, and their
coordination behavior under hydroformylation conditions
was studied using high-pressure spectroscopic techniques. Both
the catalytic and high-pressure spectroscopic data clearly illus-
trate the distinct electronic and steric properties of the diphenyl-
phosphine- and 2,8-dimethylphenoxaphosphine-modified
diphosphine ligands.
High-pressure spectroscopic studies on ligands incorpo-
rating both a phenoxaphosphine and a diphenylphosphine
moiety (1b-4b) show that phenoxaphosphine moieties are
less basic than diphenylphosphine donor moieties. Although
introduction of phenoxaphosphine moieties increases the
steric hindrance between the two donor moieties of the
diphosphine ligand, it decreases the steric hindrance imposed
by the diphosphine ligand toward other ligands in transition
metal complexes. Both the lower basicity and lower steric
bulk of phenoxaphosphine-modified ligands facilitate car-
bon monoxide-alkene exchange and the key hydride migra-
tion in the hydroformylation catalytic cycle, leading to an
increase in the overall rate of hydroformylation for the
phosphacyclic ligands. The lower regioselectivity in the
hydroformylation of 1-octene observed for these phospha-
cyclic ligands is also consistent with the decreased steric
interaction between the diphosphine ligand and the alkene
substrate during the hydroformylation reaction.
1
3
air-sensitive oil: H NMR (CDCl3) δ(ppm), 1.09 (t, JH,H
=
7.3 Hz, 12H), 3.08 (m, 8H), 7.07 (t, 3JH,H = 6.1 Hz, 1H), 7.26 (t,
3JH,H = 6.1 Hz, 1H), 7.47 (bs, 1H), 7.50 (m, 1H); 31P{1H} NMR
(CDCl3) δ(ppm), 95.8 (s); 13C{1H} NMR (CDCl3) δ(ppm), 15.2
(s), 43.9 (d, 2JC,P = 18 Hz), 127.0 (s), 127.3 (s), 129.3 (s), 132.5 (d,
1
2JC,P = 6 Hz), 133.7 (s), 147.6 (d, JC,P = 11 Hz); exact mass
(FABþ) calcd for C14H25N2BrP (M þ H) 331.0939, found
331.0939. Anal. Calcd for BrC14H24N2P: C, 50.77; H, 7.30; N,
8.46. Found: C, 50.86; H, 7.35; N, 8.53.
1-(2,8-Dimethylphenoxaphosphino)-2-(bis(diethylamino)phos-
phino)benzene (6). To a solution of 2.0 g of 1-(bis(diethylamino)-
phosphino)-2-bromobenzene (5) (6.0 mmol) in 40 mL of diethyl
ether was added dropwise 2.8 mL of a 2.35 M solution of
n-butyllithium (6.6 mmol) at -80 ꢀC. The reaction mixture
was stirred at -80 ꢀC for one hour and one additional hour at
0 ꢀC. The dark red solution was cooled to -80 ꢀC, and a solution
of 1.8 g of 2,8-dimethylphenoxaphosphine chloride (6.9 mmol)
in 20 mL of THF was added dropwise. The resulting suspension
was stirred overnight, during which the solution was allowed to
warm to room temperature. The solvents were removed in vacuo,
and the product was purified using flash column chromatogra-
phy over basic alumina (eluent: dichloromethane). Recrystalli-
zation of the solid from cold n-hexane yielded 1.9 g of white
5. Experimental Section
5.1. General Remarks. All reactions were carried out using
standard Schlenk techniques under an atmosphere of purified
argon, unless stated otherwise. Toluene was distilled from sodium,
THF and diethyl ether were distilled from sodium/benzophenone,
and hexanes and n-pentane were distilled from sodium/benzophe-
none/triglyme. Dichloromethane and all deuterated solvents were
distilled from CaH2. Chemicals were purchased from Aldrich
Chemical Company and Acros Chimica. 2,20-Dibromo-p-tolyl
ether,32 2,20-bis(diphenylphosphino)-p-tolyl ether, 2,8-dimethyl-
phenoxaphosphine chloride,8c,33 (2-bromophenyl)(diphenyl)phos-
phine,20,34 1-bromo-10-diphenylphosphinoferrocene,19 bis(diphenyl-
phosphino)-2,7-di-tert-butyl-9,9-dimethylxanthene,8a,b and bis(2,
8-dimethylphenoxaphosphino)-2,7-di-tert-butyl-9,9-dimethylxan-
thene8a,b,14 were synthesized according to previously published
procedures.
1
crystals (3.9 mmol, 65%): H NMR (CDCl3) δ(ppm) 1.16 (t,
3JH,H = 7.2 Hz, 12H), 2.03 (s, 6H), 3.20 (m, 8H), 7.02 (d,
3JH,H = 8.1 Hz, 2H), 7.07 (m, 4H), 7.20 (t, 3JH,H = 7.2 Hz, 1H),
3
7.32 (bd, JH,H = 9.6 Hz, 2H), 7.55 (m, 1H); 31P{1H} NMR
(CDCl3) δ(ppm) -58.9 (d, 3JP,P = 126 Hz), 95.1 (d, 3JP,P = 126
Hz); 13C{1H} NMR (CDCl3) δ(ppm), 15.1 (s), 20.5 (s), 44.1 (s),
1
2
117.8 (s), 119.9 (dd, JC,P = 9 Hz, JC,P = 4 Hz), 128.7 (d,
1JC,P = 17 Hz), 130.2 (m), 130.9 (s), 132.8 (d, 1JC,P = 10 Hz),
2
134.5 (bs), 135.1 (dd, 1JC,P = 32 Hz, JC,P = 5 Hz), 144.1 (s),
144.3 (d, J = 29 Hz), 148.0 (dd, 1JC,P = 31 Hz, 2JP,C = 16 Hz),
153.8 (s). Anal. Calcd for C28H36N2OP2: C, 70.28; H, 7.58; N,
5.85. Found: C, 70.39; H, 7.69; N, 5.74.
5.2. Computational Details. Natural bite angle calculations
were performed with CaChe WorkSystem 6.1 using a modified
MM3 force field. A methodology similar to the one described by
Casey and Whiteker was employed.2 The rhodium-phosphorus
1-(2,8-Dimethylphenoxaphosphino)-2-(dichlorophosphino)benzene
(7). 120 mg of 1-(2,8-dimethylphenoxaphosphino)-2-(bis(die-
thylamino)phosphino)benzene (6) (0.25 mmol) was dissolved
at 0 ꢀC in 1.0 mL of PCl3. The solution was heated to 50 ꢀC for
2 h. The solvents were removed in vacuo, and the resulting solid was
extracted using 2 ꢀ 2 mL of diethyl ether. Evaporation of the
solvents yielded 85 mg of a highly air- and moisture-sensitive white
solid (0.21 mmol, 84%), which was recrystallized from n-hexane.
1H NMR (CDCl3) δ(ppm), 2.27 (s, 6H), 7.10 (d, 3JH,H = 8.7 Hz,
2H), 7.17 (dd, 3JH,H = 8.7 Hz, 4JH,H = 1.8 Hz, 2H), 7.24 (m, 3H),
˚
equilibrium distance was set to 2.315 A, and the corresponding
force constant was set to 1000. The phosphorus-rhodium-
phosphorus angle was set to 180ꢀ, and its force constant was set
to 0.0. Energy minimizations were performed using the block-
diagonal Newton-Raphson method, applying a geometry con-
-1
˚
vergence criterion of 0.0001 kcal A
.
Solid angles22 of the MM3-optimized structures were deter-
mined using the Steric program,23 developed by Taverner and
7.31 (dd, 3JH,H = 7.8 Hz, 4JH,H = 1.5 Hz, 1H), 7.44 (dd, 3JH,H
7.8 Hz, 3JH,H = 7.2 Hz, 1H), 8.19 (m, 1H); 31P{1H} NMR (CDCl3)
=
(32) Okawa, H.; Koyama, H.; Inazu, T.; Yoshino, T. Bull. Chem. Soc.
Jpn. 1970, 43, 1729.
(33) Herwig, J.; Bohnen, H.; Skutta, P.; Sturm, S.; Van Leeuwen,
P. W. N. M.; Bronger, R. P. J. WO 2002068434, 2002.
(34) Baillie, C.; Xiao, J. Tetrahedron 2004, 60, 4159.
3
3
δ(ppm), -66.8 (d, JP,P = 344 Hz), 160.5 (d, JP,P = 344 Hz);
13C{1H} NMR (CDCl3) δ(ppm), 20.8 (s), 116.5 (d, 1JC,P = 9 Hz),
130.4 (dd, 1JC,P = 9 Hz, 2JC,P = 3.6 Hz), 130.6 (s), 132.3 (s), 133.3

Article
Organometallics, Vol. 29, No. 5, 2010 1219
(d, 1JC,P = 28 Hz), 133.6 (s), 134.8 (s), 134.8 (d, 1JC,P = 31 Hz),
145.0 (d, 1JC,P = 32 Hz), 145.3 (dd, 1JC,P = 31 Hz, 2JC,P = 6 Hz),
146.0 (d, 1JC,P = 31 Hz), 155.6 (s).
and the resulting solid was redissolved in 60 mL of dichlo-
romethane. The mixture was washed using 30 mL of a
degassed 1 M aqueous HCl solution. The organic layer was
separated and dried using MgSO4. The solvents were re-
moved in vacuo, and the resulting solid was recrystallized
from dichloromethane/methanol, yielding 1.83 g of colorless
crystals (2.5 mmol, 70%): 1H NMR (CDCl3) δ(ppm) 1.03 (s,
2-Bromo-20-(2,8-dimethylphenoxaphosphino)p-tolyl Ether (8).
To a solution of 3.0 g of 2,20-dibromo-p-tolyl ether (8.4 mmol) in
50 mL of diethyl ether was added dropwise 3.5 mL of a 2.4 M
solution of n-butyllithium (8.4 mmol) at -80 ꢀC. The solution
was stirred for 30 min. Subsequently, 2.23 g of 2,8-dimethyl-
phenoxyphosphine chloride (8.5 mmol) suspended in 30 mL of
THF was added to the solution. The resulting mixture was
stirred overnight, allowing the temperature to rise to room
temperature. The solvents were removed in vacuo, and the
resulting sticky solid was dissolved in 50 mL of dichloro-
methane. Then 40 mL of degassed water was added to the
mixture. The organic layer was separated, and the water layer
was extracted with 2 ꢀ 20 mL of dichloromethane. The com-
bined organic layers were dried using MgSO4. Removal of the
solvent in vacuo yielded a light yellow solid. The product was
recrystallized from dichloromethane/methanol, yielding 5.2 g of
colorless needles (5.0 mmol, 62%): 1H NMR (toluene-d8)
9H), 1.15 (s, 9H), 1.60 (s, 6H), 2.21 (s, 6H), 7.02 (d, 3JH,H
=
8.2 Hz, 2H), 7.08 (dd, 3JH,H = 8.2 Hz, 4JH,H = 2.4 Hz), 7.21
(d, JH,H = 2.4 Hz), 7.4-7.5 (m, 13H); 31P{1H} NMR
4
(CDCl3) δ(ppm), -70.9 (d, JP,P = 26 Hz), -15.3 (d, JP,P
=
26 Hz); 13C{1H} NMR (CDCl3) δ(ppm) 20.9 (s), 21.4 (s), 34.1
(s), 34.4 (s), 112.3 (s), 115.4 (s), 117.2 (s), 117.8 (s), 118.9 (s),
119.2 (s), 127.3 (s), 128.4 (s), 129.3 (s), 130.9 (s), 131.6 (s),
131.9 (s), 132.5 (s), 132.7 (s), 132.9 (s), 133.5 (s), 134.3 (s),
134.5 (s), 136.0 (s), 136.4 (s), 137.8 (d, 1JC,P = 15 Hz), 144.3
1
1
(s), 153.2 (s), 157.1 (d, JC,P = 23 Hz), 158.3 (d, JC,P = 22
Hz); exact mass (FABþ) calcd for C49H51O2P2 (M þ H)
733.3364, found 733.3362. Anal. Calcd for C49H50O2P2: C,
80.3; H, 6.88. Found: C, 80.25; H, 6.81.
2-(2,8-Dimethylphenoxaphosphino)-20-(diphenylphosphino)-p-
tolyl Ether (2b). To a solution of 1.5 g of 2-bromo-20-(2,8-
dimethylphenoxaphosphino)-p-tolyl ether (8) (3.0 mmol) in 40
mL of THF was added dropwise 1.3 mL of a 2.4 M solution of
n-butyllithium (3.1 mmol) at -80 ꢀC. The resulting solution was
stirred at -80 ꢀC for one hour, after which 0.6 mL of chloro-
(diphenyl)phosphine (3.1 mmol) dissolved in 10 mL of diethyl
ether was added dropwise. The reaction mixture was stirred
overnight, allowing the solution to warm slowly to room
temperature. The solvents were removed in vacuo, and the
resulting solid was redissolved in 40 mL of dichloromethane.
The mixture was washed using 30 mL of a degassed 1 M aqueous
HCl solution. The organic layer was separated and dried using
MgSO4. The solvents were removed in vacuo, and the resulting
solid was recrystallized from dichloromethane/methanol, yield-
ing 1.3 g of colorless crystals (2.1 mmol, 73%): 1H NMR
(CDCl3) δ(ppm) 2.05 (s, 3H), 2.20 (s, 3H), 2.25 (s, 6 H), 6.22
(dd, 3JH,H = 8.0 Hz, 3JH,H = 4.5 Hz, 1H), 6.38 (d, 3JH,H = 4.5
Hz, 1H), 6.43 (d, 3JH,H = 4.2 Hz, 1H), 6.65 (dd, 3JH,H = 4.2 Hz,
4JH,H = 2.2 Hz, 1H), 6.82 (dd, 3JH,H = 8.5 Hz, 4JH,H = 2.2 Hz),
3
δ(ppm), 2.11 (s, 3H), 2.26 (s, 6H), 2.31 (s, 3H), 6.43 (d, JH,H
= 12.1 Hz, 1H), 6.52 (m, 2H), 6.94 (d, 3JH,H = 7.9 Hz, 2H), 7.07
(d, 3JH,H = 7.9 Hz, 2H), 7.15 (d, 3JH,H = 7.9 Hz, 2H), 7.42 (bd,
3JH,H = 12.1 Hz), 7.46 (bs, 1H); 31P{1H} NMR (toluene-d8)
δ(ppm), -64.8 (s); 13C{1H} NMR (toluene-d8) δ(ppm), 20.7 (s),
20.8 (s), 21.1 (s), 114.0 (s), 117.2 (s), 117.6 (s), 117.9 (s), 129.3 (s),
130.3 (s), 130.6 (d, 1JC,P = 26 Hz), 131.0 (s), 131.9 (s), 132.9 (d,
1JC,P = 12 Hz), 133.2 (s), 133.4 (s), 134.0 (s), 134.5 (s), 136.0 (s),
136.5 (s), 152.1 (s), 154.9 (s), 156.5 (d, 1JC,P = 16 Hz); exact mass
(FABþ) calcd for C28H25BrO2P (M þ H) 503.0776, found
503.0780. Anal. Calcd for C28H24BrO2P: C, 66.81; H, 4.81.
Found: C, 66.68; H, 4.90.
4-Bromo-2,7-di-tert-butyl-9,9-dimethyl-5-diphenylphosphino-
xanthene (9). This compound was synthesized using a modified
literature procedure.8a,b To a solution of 5.0 g of 4,5-dibromo-
2,7-di-tert-butyl-9,9-dimethylxanthene (10.4 mmol) in 120 mL
of THF was added dropwise 4.3 mL of a 2.4 M solution of
n-butyllithium (10.4 mmol) at -80 ꢀC. The solution was stirred
for 2 h. Subsequently, 1.9 mL of chlorodiphenylphosphine
(10.5 mmol) in 25 mL of hexanes was added dropwise to the
solution. The resulting mixture was stirred at -80 ꢀC for one
hour and was subsequently allowed to warm to room tempera-
ture overnight. The mixture was hydrolyzed using a 1.0 M
solution of HCl in water and extracted three times with 60 mL
of dichloromethane. The combined organic layers were dried
using MgSO4. Removal of the solvent in vacuo yielded a light
yellow solid. The product was recrystallized from dichloro-
methane/methanol, yielding 3.7 g of colorless needles (6.4 mmol,
62%): 1HNMR(CDCl3) δ(ppm), 1.15(s, 9H), 1.32 (s, 9H), 1.75(s,
6.94 (dd, 3JH,H = 8.5 Hz, 4JH,H = 2.2 Hz, 1H), 7.03 (d, 3JH,H
=
8.1 Hz, 2H), 7.09 (dd, 3JH,H = 8.1 Hz, 4JH,H = 1.8 Hz, 2H), 7.26
(dd, 3JH,H = 10.5 Hz, 4JH,H = 1.8 Hz, 2H), 7.35-7.6 (m, 10H);
31P{1H} NMR (CDCl3) δ(ppm), -66.3 (s), -15.9 (s); 13C{1H}
NMR (CDCl3) δ(ppm), 20.7 (s), 21.1 (s), 117.1 (s), 117.4 (s),
117.6 (s), 119.3 (s), 127.8 (s), 128.7 (s), 128.9 (s), 130.9 (s), 131.6
(s), 131.9 (s), 132.5 (s). 132.8 (s), 132.9 (s), 133.5 (s), 134.3 (s),
134.5 (s), 136.0 (s), 136.4 (s), 137.1 (d, 1JC,P = 11 Hz), 154.6 (s),
156.4 (d, 1JC,P = 26 Hz), 158.3 (d, 1JC,P = 17 Hz); exact mass
(FABþ) calcd for C40H35O2P2 (M þ H) 609.2112, found
609.2117. Anal. Calcd for C40H34O2P2: C, 78.93; H, 5.63.
Found: C, 78.58; H, 5.54.
3
6H), 6.54 (dd, JP,H = 4.4 Hz, 4JH,H = 2.4 Hz), 7.35 (m, 13H);
31P{1H} NMR (CDCl3) δ(ppm), -13.5 (s); 13C{1H} NMR
(CDCl3) δ(ppm), 31.2 (s), 31.4 (s), 32.3 (s), 34.6 (s), 35.5 (s),
121.6 (s), 123.1 (s), 125.0 (d, 1JP,C = 15 Hz), 128.5 (s), 128.7 (d,
2,20-Bis(1,10-dimethylphenoxaphosphino)-p-tolyl Ether (2c).
To a solution of 2.0 g of p-tolyl ether (10 mmol) and 3.5 mL of
TMEDA (23 mmol, TMEDA = N,N,N0,N0-tetramethylethyl-
enediamine) in 50 mL of ether was added 9.25 mL of a 2.5 M
solution of n-butyllithium (23 mmol) at 0 ꢀC. The solution was
allowed to warm to room temperature and was stirred over-
night. The resulting suspension was cooled to -80 ꢀC, and
6.1 g of 2,8-dimethylphenoxyphosphine chloride (23 mmol) in
100 mL of toluene was added. The mixture was stirred over-
night, allowing the temperature of the solution to rise to room
temperature. The solvents were removed in vacuo. The resulting
sticky solid was dissolved in 100 mL of dichloromethane. The
mixture was quenched with 60 mL of degassed water. The water
layer was extracted with 5 ꢀ 60 mL of dichloromethane, and the
combined organic layers were dried using MgSO4. The solvent
was removed in vacuo. The product was recrystallized from
dichloromethane/ethanol, yielding 4.0 g of colorless crystals
(6.1 mmol, 61%): 1H NMR (CDCl3) δ(ppm) 2.10 (s, 6H), 2.20
2JP,C = 28 Hz), 129.1 (s), 131.1 (s), 134.0 (s), 134.8 (s), 136.4 (d,
2
1JP,C = 11 Hz), 145.0 (s), 145.6 (s), 146.8 (s), 150.1 (d, JP,C
=
15 Hz); exact mass (FABþ) calcd for C35H39BrOP (M þ H)
587.1908, found 587.1919. Anal. Calcd for C35H38BrOP: C, 71.79;
H, 6.55. Found: C, 71.58; H, 6.59.
4-(2,8-Dimethylphenoxaphosphino)-5-(diphenylphosphino)-
2,7-di-tert-butyl-9,9-dimethylxanthene (1b). This compound
was synthesized using a modified literature procedure.8a,b To
a solution of 2.0 g of 4-bromo-2,7-di-tert-butyl-9,9-dimeth-
yl-5-diphenylphosphinoxanthene (9) (3.5 mmol) in 50 mL
of THF was added dropwise 1.5 mL of a 2.4 M solution of
n-butyllithium (3.6 mmol) at -80 ꢀC. The resulting solution
was stirred at -80 ꢀC for one hour, after which 1.0 g of 2,8-
dimethylphenoxaphosphine chloride (3.8 mmol) suspended
in 25 mL of THF was added dropwise. The reaction mixture
was stirred overnight, allowing the solution to warm slowly
to room temperature. The solvents were removed in vacuo,

1220 Organometallics, Vol. 29, No. 5, 2010
Zuidema et al.
(bs, 12H), 5.94 (dd, 3JH,H = 8.1 Hz, 4JH,H = 5.3 Hz, 2H), 6.36
(d, 4JH,H = 5.3 Hz, 2H), 6.72 (d, 3JH,H = 8.1 Hz, 2H), 7.0-7.8
(m, 12H); 31P{1H} NMR (CDCl3) δ(ppm), -62.3 (s); 13C{1H}
NMR (CDCl3) δ(ppm), 20.7 (s), 21.1 (s), 117.6 (s), 117.9 (s),
temperature. The solvents were removed in vacuo. Then 50 mL of
dichloromethane was added, and the mixture was quenched using
60 mL of degassed water. The water layer was extracted with 3 ꢀ
10 mL of dichloromethane, and the combined organic layers were
dried using MgSO4. The solvent was removed in vacuo. The product
was recrystallized from dichloromethane/methanol, yielding1.8gof
white crystals (3.7 mmol, 94%): 1H NMR (CDCl3) δ(ppm), 2.18 (s,
6H), 6.96 (m, 1H), 7.00 (bs, 1H), 7.05 (bs, 1H), 7.09 (bd, 4H),
7.14 (m, 2H), 7.28 (m, 1H), 7.40 (bd, 4H), 7.42 (bs, 6H); 31P{1H}
118.5 (s), 130.5 (s), 130.8 (s), 131.8 (s), 132.7 (s), 133.1 (d, 1JP,C
=
11 Hz), 136.4 (s), 136.9 (s), 155.3 (s), 147.3 (d, 1JP,C = 15 Hz);
exact mass (FABþ) calcd for C42H37O3P2 (M þ H) 651.2218,
found 651.2221. Anal. Calcd for C42H36O3P2: C, 77.53; H, 5.58.
Found: C, 77.41; H, 5.34
3
1-(2,8-Dimethylphenoxaphosphino)-10-(diphenylphosphino)ferro-
cene (3b). To a solution of 2.7 g of 1-bromo-10-diphenylpho-
sphinoferrocene (6.0 mmol) in 40 mL of diethyl ether was added
dropwise 2.8 mL of a 2.4 M solution of n-butyllithium (6.7 mmol)
at -80 ꢀC. The mixture was stirred for 1 h. To this mixture was
added dropwise a solution of 1.9 g of 2,8-dimethylphenoxapho-
sphine chloride (7.2 mmol) in 30 mL of toluene. The solution was
stirred overnight, allowing the temperature of the mixture to rise to
room temperature. The solvents were removed in vacuo. Then
50 mL of dichloromethane was added, and the mixture was
quenched using 20 mL of degassed water. The organic layer was
separated, and the water layer was washed with 2 ꢀ 10 mL of
dichloromethane. The combined organic layers were dried over
MgSO4. The solvent was removed in vacuo. The yellow solid
was washed with 2 ꢀ 20 mL of pentane. The compound was
recrystallized from dichloromethane/methanol, yielding 1.9 g
of orange-red crystals (3.1 mmol, 52%): 1H NMR (C6D6)
δ(ppm), 2.07 (s, 6H), 3.93 (s, 2H), 4.08 (s, 2H), 4.29 (s, 4H), 4.21
(s, 2H), 6.88 (bd, 3JH,H = 8.2 Hz, 2H), 7.14 (d, 3JH,H = 8.2 Hz,
2H), 7.20 (bs, 2H), 7.52 (m, 10 H); 31P{1H} NMR (CDCl3)
δ(ppm), -66.7 (s), -16.6 (s); 13C{1H} NMR (C6D6) δ(ppm),
20.6 (s), 72.2 (s), 72.7 (d, 1JC,P = 15 Hz), 74.0 (d, 1JC,P = 15 Hz),
77.5 (d, 1JC,P = 9 Hz), 82.7 (d, 1JC,P = 21 Hz), 117.7 (s), 120.0 (s),
NMR (CDCl3) δ(ppm), -11.7 (d, JP,P = 162 Hz), -62.9 (d,
3JP,P = 162 Hz); 13C{1H} NMR (CDCl3) δ(ppm), 20.8 (s), 117.7
(s),118.6(s),128.9(s),129.3(s),129.8(s),131.7(s),133.4(d,1JC,P
=
51 Hz), 133.8 (d, 1JC,P = 7 Hz), 134.4 (s), 135.1 (s), 135.5 (s), 137.8
(bs), 153.7 (s); exact mass (FABþ) calcd for C32H27OP2 (M þ H)
489.1537, found 489.1534. Anal. Calcd for C32H26OP2: C, 78.68; H,
5.36. Found: C, 78.58; H, 5.31.
Bis(2,8-dimethylphenoxaphosphino)benzene (4c). To a solu-
tion of 120 mg of 2,20-dibromo-p-tolyl ether (0.34 mmol) in
10 mL of diethyl ether was added dropwise 1.0 mL of a 1.4 M
solution of tert-butyllithium (1.4 mmol) at -80 ꢀC. The solution
was stirred for one hour at -80 ꢀC and subsequently allowed to
warm slowly to room temperature. The solution was stirred at
room temperature for two additional hours. The resulting
yellow solution was added dropwise over a period of one hour
to a solution of 100 mg of 1-(2,8-dimethylphenoxaphosphino)-
2-(dichlorophosphino)benzene (7) (0.25 mmol) in 15 mL of
diethyl ether at room temperature. The resulting suspension
was stirred overnight at room temperature. Solvents were
removed in vacuo, and the resulting solid was dissolved in
20 mL of dichloromethane. The mixture was quenched using
20 mL of a degassed 1 M aqueous HCl solution. The organic
layer was separated, and the aqueous layer was washed using 2 ꢀ
10 mL of dichloromethane. The combined organic layers
were dried using MgSO4, and subsequently, the mixture was
filtered over a plug of dry silica. Removal of the solvent yielded
0.15 g of a white solid. The product was recrystallized from
dichloromethane/methanol, yielding 103 mg of colorless crys-
tals (0.2 mmol, 79%): 1H NMR (CDCl3) δ(ppm), 2.23 (s,
1
1
128.6 (s), 131.8 (s), 132.7 (d, JC,P = 12 Hz), 133.9 (d, JC,P
=
20 Hz), 135.4 (s), 135.8 (s), 139.7 (d, 1JC,P = 11 Hz), 154.8 (s); exact
mass (FABþ) calcd for C36H31FeOP2 (M þ H) 597.1200, found
597.1200. Anal. Calcd for C36H30FeOP2: C, 72.50; H, 5.07. Found:
C, 72.42; H, 5.15.
1,2-Bis(1,10-dimethylphenoxaphosphino)ferrocene (3c). To a
solution of 3 g of ferrocene (16 mmol) and 5.6 mL of TMEDA
(36.8 mmol) in 50 mL of ether was added 16 mL of a 2.5 M
solution of n-butyllithium (40 mmol) at 0 ꢀC. The solution was
allowed to come to room temperature and was stirred overnight.
The resulting suspension was cooled to -80 ꢀC, and 9.8 g of 2,8-
dimethylphenoxyphosphine chloride (37.3 mmol) in 100 mL of
toluene was added. The mixture was stirred overnight, allowing
the temperature of the solution to rise to room temperature. The
solvents were removed in vacuo. The resulting dark brown oil
was dissolved in 50 mL of dichloromethane. The mixture was
quenched using 60 mL of degassed water. The water layer was
extracted with 5 ꢀ 60 mL of dichloromethane, and the combined
organic layers were dried using MgSO4. The solvent was re-
moved in vacuo. The product was recrystallized from dichloro-
methane/ethanol, yielding 4.46 g of red crystals (7.0 mmol,
44%): 1H NMR (CDCl3) δ(ppm), 2.35 (s, 12H), 3.83 (bd,
3
12H), 7.01 (bs, 4H), 7.07 (d, JH,H = 8.2 Hz, 2H), 7.09 (s,
4H), 7.13 (bd, 3JH,H = 8.2 Hz, 2H), 7.20 (bt, 3JH,H = 8.1 Hz,
4H); 31P{1H} NMR (CDCl3) δ(ppm), -59.5 (s); 13C{1H} NMR
(CDCl3) δ(ppm), 20.8 (s), 117.7 (s), 119.0 (s), 127.9 (s), 131.5 (s),
1
1
133.0 (s), 133.1 (d, JC,P = 5.0 Hz), 134.5 (d, JC,P = 38 Hz),
135.1 (d, JC,P = 17 Hz), 145.9 (bs), 153.7 (s); exact mass
1
(FABþ) calcd for C34H29O2P2 (M þ H) 531.1643, found
531.1659. Anal. Calcd for C34H28O2P2: C, 76.97; H, 5.32.
Found: C, 76.86; H, 5.27.
X-ray crystal structure determination of ligand 4c: C34H28O2P2,
fw = 530.50, colorless block, 0.30 ꢀ 0.30 ꢀ 0.30 mm3, monoclinic,
˚
˚
C2/c (no. 15), a = 25.2531(18) A, b = 7.8060(4) A, c = 17.0051(8)
A, β= 126.849(3)ꢀ, V= 2682.4(3) A , Z=4, Dcalc = 1.314 g/cm3,
3
˚
˚
μ = 0.193 mm-1; 34 055 reflections were measured at a tempera-
ture of 150(2) K up to a resolution of (sin θ/λ)max = 0.65 A-1 on
˚
a Nonius KappaCCD diffractometer with rotating anode and
˚
3
3JH,P = 2.0 Hz, 4H), 4.08 (s, 4H), 7.07 (d, 4 H, JH,H = 8.1
graphite monochromator (λ = 0.71073 A). The intensities were
obtained with Eval1435 using an accurate description of the crystal
form and the diffraction geometry. An absorption correction based
on multiple measured reflections was applied36 (0.84-0.94 correc-
tion range). A total of 3086 reflections were unique (Rint = 0.019).
The structure was solved with Direct Methods and refined with
SHELXL-9737 on F2 of all reflections. Non-hydrogen atoms were
refined freely with anisotropic displacement parameters. All hy-
drogen atoms were located in the difference Fourier map and
refined freely with isotropic displacement parameters. A total of
Hz), 7.17 (dd, 3JH,H = 8.1 Hz, 4JH,H = 1.8 Hz, 4H), 7.38 (bdd,
3JH,P = 10.9 Hz, 4JH,H = 1.8 Hz, 4H); 31P{1H} NMR (CDCl3)
δ(ppm), -66.64 (s); 13C{1H} NMR (CDCl3) δ(ppm), 20.0 (s),
71.3 (s), 72.2 (d, 1JC,P = 15 Hz), 82.1 (d, 1JC,P = 19 Hz), 117.8
1
(s), 119.4 (s), 132.1 (s), 133.1 (d, JC,P = 12 Hz), 135.0 (d,
1JC,P = 38 Hz), 154.2 (s); exact mass (FABþ) calcd for
C38H33FeO2P2 (M þ H) 639.1305, found 639.1309. Anal. Calcd
for C38H32FeO2P2: C, 71.49; H, 5.05. Found: C, 71.37; H, 5.11.
1-(2,8-Dimethylphenoxaphosphino)-2-(diphenylphosphino)benzene
(4b). To 1.34 g of (2-bromophenyl)(diphenyl)phosphine (3.9 mmol)
in 40 mL of diethyl ether was added dropwise 1.6 mL of a 2.5 M
solution of n-butyllithium (4.0 mmol) at -80 ꢀC. After 15 min, 1.1 g
of 2,8-dimethylphenoxaphosphine chloride (4.2 mmol) suspended
in 20 mL of diethyl ether was added. The solution was stirred
overnight, allowing the temperature of the mixture to rise to room
(35) Duisenberg, A. J. M.; Kroon-Batenburg, L. M. J.; Schreurs,
A. M. M. J. Appl. Crystallogr. 2003, 36, 220.
(36) Sheldrick, G. M. SADABS, Area-Detector Adsorption Correc-
tion.
(37) Sheldrick, G. M. SHELXS-97, Program for Crystal Structure
Refinement.

Article
Organometallics, Vol. 29, No. 5, 2010 1221
174 parameters were refined with no restraints. R1/wR2 [I >
2σ(I)]: 0.0306/0.0806. R1/wR2 [all reflns]: 0.0340/0.0830. S =
chamber. In a typical experiment, Rh(acetylacetonate)(CO)2
(4 mg, 15 μmol) and 10 equivalents of ligand (150 μmol) were
dissolved in 15.0 mL of cyclohexane under an inert atmosphere.
The IR autoclave was placed under vacuum and subsequently
purged three times using 22 bar of a 1:1 mixture of H2 and CO. The
catalyst solution was transferred to the autoclave via syringe. The
autoclave was purged an additional three times and subsequently
pressurized to 17 bar. The autoclave was heated to 40 ꢀC. Spectra
were collected at regular intervals until the catalytically active
species was formed. Hydrogen-deuterium exchange was per-
formed by flushing the autoclave three times using 30 bar of
CO and subsequently pressurizing the autoclave using 20 bar of
D2/CO, 1:1.
High-Pressure NMR Spectroscopy. The high-pressure NMR
spectroscopic experiments were performed in a 10 mm sapphire
high-pressure NMR tube. In a typical experiment, Rh-
(acetylacetonate)(CO)2 (6 mg, 24 μmol) and 2.2 equivalents of
ligand were dissolved in 2.0 mL of toluene-d8. The solution was
introduced into the NMR tube under an inert atmosphere using
Schlenk techniques. The tube was purged three times using 20
bar of CO and subsequently pressurized using 20 bar of H2/CO,
1:1. 1H and 31P{1H} NMR spectra were collected at regular
intervals until changes in the spectra were no longer observed.
Fluxional behavior in the final products was analyzed using
variable-temperature NMR spectroscopy (temperature range
-85 to þ80 ꢀC).
3
˚
1.034. Residual electron density between -0.24 and 0.37 e/A .
Geometry calculations, drawings, and checking for higher sym-
metry were performed with the PLATON package.38
5.4. High-Pressure Studies. Hydroformylation of 1-Octene.
The hydroformylation experiments were performed in a stain-
less steel (SS 316) autoclave (196 mL). The autoclave was stirred
mechanically and equipped with a separate reservoir, a pressure
transducer, and a thermocouple. In a typical experiment,
Rh(acetylacetonate)(CO)2 (5 mg, 20 μmol) and 10 equivalents
of ligand were dissolved in 15.0 mL of toluene under an inert
atmosphere. The autoclave was placed under vacuum for 30 min
and subsequently purged three times using 15 bar of H2/CO. The
catalyst solution was introduced into the autoclave via a syringe.
The autoclave was purged an additional three times using 15 bar
of H2/CO and then pressurized to 15 bar and heated to 80 ꢀC.
The solution was stirred at this temperature for 3 h. A mixture of
1-octene (2 mL, 12.7 mmol), decane (internal standard, 1.0 mL,
5.1 mmol), and 2 mL of toluene was introduced to the separate
reservoir and purged three times with 20.0 bar of H2/CO. The
substrate solution was introduced into the autoclave using
overpressure, and the autoclave was pressurized to a total
pressure of 20.0 bar. The hydroformylation reaction was
stopped after a pressure drop of approximately 0.5 bar using a
solution of tributylphosphite in toluene. The autoclave was
cooled rapidly using an ice-water bath and depressurized.
The experiments were performed in duplicate. The conversion,
regioselectivity, and amount of isomerized alkenes were deter-
mined by GC-analysis of the reaction mixture.
Acknowledgment. This work was supported by The
Netherlands Organization for Scientific Research
(NWO/CW). E.Z. would like to acknowledge R. Bouwer
for the synthesis of ligand 2c.
High-Pressure Infrared Spectroscopy. The high-pressure in-
frared spectroscopic experiments were performed in a stainless
steel (SS 316) 50 mL autoclave equipped with IRTRAN win-
dows (ZnS, transparent up to 700 cm-1, 10 mm i.d., optical
path length = 0.4 mm), a mechanical stirrer, a temperature
controller, a pressure transducer, and a separate 50 mL second
Supporting Information Available: 1H, 31P{1H}, and 13C{1H}
NMR spectra of water- and oxygen-sensitive intermediate 7 and
a CIF file giving details of the single-crystal X-ray structure
determination of ligand 4c. This material is available free of
charge via the Internet at http://pubs.acs.org.
(38) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7.