Phosphinoureas: Cooperative ligands in rhodium-catalyzed hydroformylation? on the possibility of a ligand-assisted reductive elimination of the aldehyde

Article abstract of DOI:10.1021/om901028z

We report the synthesis of a novel type of phosphinourea ligand, its coordination chemistry with rhodium, its use in the asymmetric hydroformylation of styrene, and investigations on the hydroformylation reaction mechanism. Complex studies on the 2:1 complex of phosphinourea to [Rh(acac)(CO) ₂] showed that a neutral trans-coordinating complex [Rh(HL-κP)(L-κ²O,P)(CO)] was formed. An anionic O,P-chelating ligand has displaced the anionic acac^- ligand via an acid-base reaction involving the deprotonation of an acidic urea proton, giving Hacac. A second phosphinourea is coordinated as a neutral monodentate ligand and is linked to the chelating anionic ligand via an intramolecular hydrogen bond. The behavior of these supramolecular complexes in the hydroformylation reaction and the possible cooperative role of the ligands in the catalytic cycle were studied both experimentally and by computational methods. High-pressure NMR spectroscopy revealed that the catalytically active rhodium hydride species further consists of two neutral phosphinourea ligands and is in equilibrium with the neutral species [Rh(HL-κP)(L-κ²O,P)(CO)]. This equilibrium is likely an integrated part of a productive hydroformylation cycle involving a ligand-assisted reductive elimination of the aldehyde. DFT calculations revealed that the ligand-assisted mechanism could well be the preferred lower energetic pathway; however, the orientation of the anionic oxygen donor atom in [Rh(HL-κP)(L-κ²O,P)(CO)] prevented us from finding a direct (nonsolvent assisted) transition state to connect the intermediates. We therefore cannot exclude a mechanism where [Rh(HL-κP)(L-κ²O,P)(CO)] is a dormant species outside the productive hydroformylation cycle, although the intermediate associated with this mechanism is higher in energy. Finally, the synthesis of heteroligated complexes was investigated, consisting of two electronically different phosphinoureas, which sets the stage for combinatorial supramolecular ligand approaches in catalysis. Simply mixing two electronically different phosphinoureas with metal precursor [Rh(acac)(CO)₂] resulted in the formation of a heterobidentate ligand. A set of six new phosphinoureas was used to prepare such rhodium complexes in a combinatorial fashion for the asymmetric hydroformylation of styrene, resulting in high conversions and selectivities for the branched product and moderate enantioselectivities.

Full text of DOI:10.1021/om901028z

Organometallics 2010, 29, 2413–2421 2413
DOI: 10.1021/om901028z
Phosphinoureas: Cooperative Ligands in Rhodium-Catalyzed
Hydroformylation? On the Possibility of a Ligand-Assisted Reductive
Elimination of the Aldehyde
Jurjen Meeuwissen,† Albertus J. Sandee,†,‡ Bas de Bruin,† Maxime A. Siegler,§
Anthony L. Spek,§ and Joost N. H. Reek*,†
†Van’t Hoff Institute for Molecular Sciences, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV
Amsterdam, The Netherlands, ‡BASF Nederland B.V. Catalysts, Strijkviertel 67, 3454 PK De Meern,
The Netherlands, and Crystal and Structural Chemistry, Bijvoet Centre for Biomolecular Research,
Faculty of Science, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands
§
Received November 27, 2009
We report the synthesis of a novel type of phosphinourea ligand, its coordination chemistry with
rhodium, its use in the asymmetric hydroformylation of styrene, and investigations on the hydro-
formylation reaction mechanism. Complex studies on the 2:1 complex of phosphinourea to [Rh(acac)-
(CO)2] showed that a neutral trans-coordinating complex [Rh(HL-κP)(L-κ2O,P)(CO)] was formed.
An anionic O,P-chelating ligand has displaced the anionic acac- ligand via an acid-base reaction
involving the deprotonation of an acidic urea proton, giving Hacac. A second phosphinourea is
coordinated as a neutral monodentate ligand and is linked to the chelating anionic ligand via an
intramolecular hydrogen bond. The behavior of these supramolecular complexes in the hydroformy-
lation reaction and the possible cooperative role of the ligands in the catalytic cycle were studied both
experimentally and by computational methods. High-pressure NMR spectroscopy revealed that the
catalytically active rhodium hydride species further consists of two neutral phosphinourea ligands and
is in equilibrium with the neutral species [Rh(HL-κP)(L-κ2O,P)(CO)]. This equilibrium is likely an
integrated part of a productive hydroformylation cycle involving a ligand-assisted reductive elimina-
tion of the aldehyde. DFT calculations revealed that the ligand-assisted mechanism could well be the
preferred lower energetic pathway; however, the orientation of the anionic oxygen donor atom in
[Rh(HL-κP)(L-κ2O,P)(CO)] prevented us from finding a direct (nonsolvent assisted) transition state
to connect the intermediates. We therefore cannotexclude a mechanism where [Rh(HL-κP)(L-κ2O,P)-
(CO)] is a dormant species outside the productive hydroformylation cycle, although the intermediate
associated with this mechanism is higher in energy. Finally, the synthesis of heteroligated complexes
was investigated, consisting of two electronically different phosphinoureas, which sets the stage for
combinatorial supramolecular ligand approaches in catalysis. Simply mixing two electronically
different phosphinoureas with metal precursor [Rh(acac)(CO)2] resulted in the formation of a
heterobidentate ligand. A set of six new phosphinoureas was used to prepare such rhodium complexes
in a combinatorial fashion for the asymmetric hydroformylation of styrene, resulting in high
conversions and selectivities for the branched product and moderate enantioselectivities.
Introduction
an integrated part of the catalyst, strongly partaking in the
reaction mechanism. Among such integrated catalysts are
ligands capable of substrate orientation,1 hemilabile (hybrid)
ligands,2 cooperative ligands,3,4 or a combination these
functionalities. Ligands with functionality for substrate
orientation preorganize the substrate in a favorable way to
Traditionally, “catalysis by ligand design” means that
variation of the steric and electronic properties of ligands is
used to influence the catalytic performance of a catalyst as a
whole. More recently, new developments invoke a more
active role of the ligands, in which they are considered as
(2) (a) Braunstein, P. J. Organomet. Chem. 2004, 689, 3953–3967.
(b) Braunstein, P.; Naud, F. Angew. Chem., Int. Ed. 2001, 40, 680–699.
*To whom correspondence should be addressed. E-mail: j.n.h.reek@
uva.nl. Fax: þ31 (0)20 525 6422. Tel: þ31 (0)20 525 6437.
€
(3) Grutzmacher, H. Angew. Chem., Int. Ed. 2008, 47, 1814–1818.
(1) (a) Breuil, P.-A. R.; Patureau, F. W.; Reek, J. N. H. Angew.
(4) (a) Patureau, F. W.; Kuil, M.; Sandee, A. J.; Reek, J. N. H. Angew.
Chem., Int. Ed. 2008, 47, 3180–3183. (b) Gunanathan, C.; Ben-David, Y.;
Milstein, D. Science 2007, 317, 790–792. (c) Zweifel, T.; Naubron, J.-V.;
Gr€utzmacher, H. Angew. Chem., Int. Ed. 2009, 48, 559–563. (d) Maire, P.;
ꢀ
Chem., Int. Ed. 2009, 48, 2162–2165. (b) Smejkal, T.; Breit, B. Angew.
ꢀ
Chem., Int. Ed. 2008, 47, 311–315. (c) Smejkal, T.; Breit, B. Angew. Chem.,
Int. Ed. 2008, 47, 3946–3949. (d) Gru€nanger, C. U.; Breit, B. Angew. Chem.,
€
€
Int. Ed. 2008, 47, 7346–7349. (e) Borner, A. Chirality 2001, 13, 625–628.
Bottner, T.; Breher, F.; Le Floch, P.; Gr€utzmacher, H. Angew. Chem., Int. Ed.
€
(f) Borner, A. Eur. J. Inorg. Chem. 2001, 327–337. (g) Sawamura, M.; Ito, Y.
2005, 44, 6318–6323. (e) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org.
Chem. 2001, 66, 7931–7944.
Chem. Rev. 1992, 92, 857–871.
r
2010 American Chemical Society
Published on Web 05/13/2010
pubs.acs.org/Organometallics

2414 Organometallics, Vol. 29, No. 11, 2010
Meeuwissen et al.
Figure 1. (Left) Displacement ellipsoid plot (50% probability level) of one of the two crystallographically independent molecules of
HL1 (CH hydrogen atoms omitted for clarity). The molecules link into infinite chains through NH O hydrogen bonds. (Right) POV-
3 3 3
Ray picture showing the 1D chain built of molecules of HL1 via N-H O hydrogen bond interactions (2.109 and 2.177 A). P: yellow,
O: red, N: blue, C: black, H: white.
3 3 3
enhance selectivity and activity. Breit and Reek exploited
hydrogen bonding as a tool for substrate orientation. This
strategy was demonstrated to be effective for the highly
selective hydroformylation of unsaturated carboxylic acids
with guanidinium-functionalized phosphines1b,c and for the
hydrogenation of 2-hydroxymethylacrylate using a supra-
molecular heterobidentate ligand5 of LEUPhos1a and a
urea-functionalized phosphine. In these examples hydrogen
bonding between the substrate and the ligand plays a crucial
role to obtain high selectivity. Hemilabile ligands generally
consist of a strongly coordinating donor atom and an
additional weakly coordinating donor atom that can tem-
porarily coordinate to the transition metal to stabilize reac-
Scheme 1. Synthesis of Phosphinourea Ligand HL1
disubstituted urea with a PCl compound.6,7 These ligands all
contain a phosphorus donor atom directly connected to a urea
motif. For example, phosphorus amide HL1 was prepared by
converting N,N-dibenzyl urea into a lithium salt, which was
subsequently reacted with chlorodiphenylphosphine (Scheme 1).
The X-ray structure of HL1 shows that in the solid state the
molecules are linked through hydrogen bonding via the urea
motif (Figure 1).
Addition of two equivalents of HL1 to a solution of
[Rh(acac)(CO)2] in CDCl3 resulted in the formation of a
yellow solution of complex [Rh(HL1-κP)(L1-κ2O,P)(CO)]
(Scheme 2, I).4a,8 Two doublets of doublets resonating at δ
62.9 ppm (1JP-Rh = 135Hz) and atδ 82.0 ppm (1JP-Rh = 129
Hz) are visible in the 31P{1H} NMR spectrum. The large P-P
coupling constant (2JP-P = 326 Hz) indicates that two
inequivalent phosphorus ligands are in mutual trans-positions
(Scheme 2, II). The 1H NMR spectrum shows free acetylace-
tone (Hacac) displaced from the metal complex and that one
urea proton has disappeared. This implies that the acidic urea
proton has been transferred to acac-, yielding Hacac and
an anionic ligand L1-, which presumably coordinates in a
bidentate fashion (κ2P,O), forming a five-membered ring.9
tive intermediates. Cooperative ligands were recently defined
€
3
by Grutzmacher as ligandsthat participate directly in a bond
activation reaction and undergo reversible chemical trans-
formations. Milstein combined hemilabile and cooperative
functionalities in a RuII complex based on a 2-(di-tert-butyl-
phosphinomethyl)-6-(diethylaminomethyl)pyridine “pincer”
ligand for the unprecedented dehydrogenative coupling of
alcohols and amines to yield amides.4c The reversible ligand-
assisted heterolytic splitting of H2 was observed to be a key
step in the mechanism. These examples clearly demonstrate
that the role of the ligand in transition metal catalysis may
be far less static than originally anticipated. In this paper
we report on a new type of phosphinourea ligand that forms
heterobidentate rhodium complexes via self-assembly. The
experimental and computational studies point at a ligand-
assisted (cooperative) mechanism in hydroformylation, imply-
ing an active role of this particular ligand in this reaction.
Interestingly, the approach can be extended to supramolecular
heterobidentate ligands, which sets the stage for combinatorial
approaches. This is demonstrated by the preparation of 14
complexes using combinations of phosphinourea ligands and
their application in the asymmetric hydroformylation reaction.
(6) A similar phosphinourea ligand was reported before; see:
Marchenko, A. P.; Koydan, G. K.; Smaliy, R. V.; Chaykovskaya, A.
A.; Pinchuk, A. M.; Tolmachev, A. A.; Shishkin, O. V. Eur. J. Inorg.
Chem. 2008, 3348–3352.
€
€
(7) For reports on related compounds see: (a) Kuhl, O.; Lonnecke, P.
Inorg. Chem. 2002, 41, 4315–4317. (b) K€uhl, O. Dalton Trans. 2003, 949–952.
(c) K€uhl, O. Coord. Chem. Rev. 2006, 250, 2867–2915. (d) Bhattacharyya, P.;
Slawin, A. M. Z.; Smith, M. B.; Williams, D. J.; Woollins, J. D. J. Chem. Soc.,
Dalton Trans. 1996, 3647–3651. (e) Slawin, A. M. Z.; Wainwright, M.;
Woollins, J. D. J. Chem. Soc., Dalton Trans. 2001, 2724–2730.
Results and Discussion
The general procedure for the preparation of phosphi-
nourea ligands described in this paper consists of two steps:
(1) the conversion of a secondary amine into a N,N-disubstituted
urea, followed by (2) a condensation reaction of the N,N-
(8) For similar metal complexes see: (a) Fuentes, J. A.; Clarke, M. L.;
Slawin., A. M. Z. New J. Chem. 2008, 32, 689–693. (b) Trzeciak, A. M.;
ꢀ
ꢁ
ꢀ
ꢂ
ꢂ
Stepnicka, P.; Mieczynska, A. E.; Ziozkowski, J. J. J. Organomet. Chem.
ꢂ
2005, 690, 3260–3267. (c) Trzeciak, A. M.; Ziozkowski, J. J.; Tadeusz, L.;
Choukroun, R. J. Organomet. Chem. 1999, 575, 87–97.
(5) (a) Wilkinson, M. J.; van Leeuwen, P. W. N. M.; Reek, J. N. H.
Org. Biomol. Chem. 2005, 3, 2371–2383. (b) Breit, B. Angew. Chem., Int.
Ed. 2005, 44, 6816–6825. (c) Sandee, A. J.; Reek, J. N. H. Dalton Trans.
2006, 3385–3391.
(9) A related anionic phosphinoamide ligand has been found to
coordinate to rhodium via a O,P chelate; see: Braunstein, P.; Heaton,
B. T.; Jacob, C.; Manzi, L.; Morise, X. Dalton Trans. 2003, 1396–1401.

Article
Organometallics, Vol. 29, No. 11, 2010 2415
Scheme 2. (I) Reactivity of HL1 with [Rh(acac)(CO)2]; (II) 31P{1H} NMR spectrum of the resulting complex [Rh(HL1-κP)(L1-κ2O,P)-
(CO)] (20 mM in CDCl3)
Additional proof for the urea proton abstraction from one of
the ligands was obtained by proton-coupled 31P NMR spec-
troscopy, showing that the signal resonating at δ 62.9 ppm has
retained the P-H coupling (2JP-H = 14.6 Hz, P with the urea
proton) and thus corresponds to the neutral ligand. The signal
resonating at δ 82.0 ppm lacks this coupling and thus
corresponds to the (deprotonated) anionic ligand.10 The
31P{1H}-1H 2D-NMR spectrum shows that the single urea
proton in the 1H NMR correlates to the signal resonating at δ
62.9 ppm in the 31P{1H} NMR spectrum, providing further
evidence that this peak corresponds to the neutral ligand.10
1
The urea proton signal in the H NMR spectrum of the
neutral ligand in the complex has shifted downfield by ∼1.4 ppm
compared to the free ligand, indicating that this proton is
involved in hydrogen bonding, presumably via the coordinating
oxygen atom of the anionic ligand (Scheme 2).11 The proposed
geometry of [Rh(HL1-κP)(L1-κ2O,P)(CO)] in Scheme 2, I, is
supported by the DFT-optimized structure (benzyl groups
simplified to methyl groups), showing the two P atoms of the
chelating anionic ligand and the neutral ligand in a trans-
geometry (Figure 2).12,14 The anionic (π-donating) oxygen donor
and the π-accepting CO ligand coordinate preferably in mutual
trans-positions for electronic reasons.13 The DFT-optimized
Figure 2. DFT (BP86, SV(P))-optimized structure of [Rh(HL-
κP)(L-κ2O,P)(CO)], HL = 1,1-dimethyl-3-(diphenylphosphino)-
urea (Rh-CO 1.818 A, Rh-P(κP) 2.349 A, Rh-P(κ2O,P)
2.317 A, Rh-O 2.115 A, O HN 1.973 A). CH hydrogen
3 3 3
atoms are omitted for clarity. Rh: green, P: yellow, O: red, N:
blue, C: black, H: white.
structure further reveals a NH O distance of 1.973 A,
3 3 3
indicative of hydrogen bonding.14 The complex was further
characterized by solution IR spectroscopy, showing the coordi-
nated CO ligand (ν(CO) = 1970 cm-1), and mass spectrometry
(m/z calcd for C54H49N4O2P2Rh 950.2386, found 950.2380
(CO ligand dissociated)).
Table 1. Hydroformylation of Styrene Using [Rh(HL1-κP)-
(L1-κ2O,P)(CO)]a
Rh
[mM]
pCO
(bar)
pH2
(bar)
conversion
(%)
branched
(%)
entry
We subsequently investigated the activity of [Rh(HL1-κP)-
(L1-κ2O,P)(CO)] in the hydroformylation of styrene, while
1
2
3
4
5
6
7
1
5
10
15
5
10
10
10
10
10
30
20
10
10
10
10
30
10
20
6.2
26.5
32.3
33.2
62.0
18.7
22.2
95.8
95.3
95.4
95.3
95.4
96.2
96.0
(10) See Supporting Information.
(11) For other examples of supramolecular urea-functionalized li-
gands see: (a) Meeuwissen, J.; Kuil, M.; van der Burg, A. M.; Sandee,
A. J.; Reek, J. N. H. Chem.;Eur. J. 2009, 15, 10272–10279. (b) Sandee,
A. J.; van der Burg, A. M.; Reek, J. N. H. Chem. Commun. 2007, 864–866.
(c) Knight, L. K.; Freixa, Z.; van Leeuwen, P. W. N. M.; Reek, J. N. H.
Organometallics 2006, 25, 954–960. (d) Duckmanton, P. A.; Blake, A. J.; Love,
J. B. Inorg. Chem. 2005, 44, 7708–7710. (e) Yoo, H.; Mirkin, C. A.; DiPasquale,
A. G.; Rheingold, A. L.; Stern, C. L. Inorg. Chem. 2008, 47, 9727–9729.
(12) We have attempted several times to crystallize [Rh(HL1-κP)-
(L1-κ2O,P)(CO)]; however we were not able to obtain crystals suitable
for X-ray analysis.
5
5
a Reaction conditions: Rh = [Rh(acac)(CO)2], [HL1] = 2.1 ꢀ [Rh],
[S] = 1 M, S = styrene, reaction time = 6 h, temperature = 40 °C,
solvent = CH2Cl2.
varying several reaction parameters, such as the rhodium
concentration and the partial pressures of CO and H2
(Table 1). The complex is indeed an active catalyst and
showed in all cases good selectivity for the branched product
(95.3-96.2%). Because it is generally believed that a hydride
species is the actual catalytically active species in hydroformy-
lation, we initially considered [Rh(HL1-κP)(L1-κ2O,P)(CO)]
merely as a precursor complex for a hydride-containing
(13) (a) Tejel, C.; Ciriano, M. A.; del Rıo, M. P.; Hetterscheid,
´
D. G. H.; Tsichlis i Spithas, N.; Smits, J. M. M.; de Bruin, B. Chem.;
Eur. J. 2008, 14, 10932–10936. (b) Tejel, C.; del Río, M. P.; Ciriano, M. A.;
ꢂ ꢀ
Reijerse, E. J.; Hartl, F.; Zalis, A.; Hetterscheid, D. G. H.; Tsichlis i Spithas;
de Bruin, B. Chem.;Eur. J. 2009, 15, 11878–11889.
(14) The corresponding anionic N-bound chelate is 16 kcal.mol-1
higher in energy than the anionic O-bound chelate.

2416 Organometallics, Vol. 29, No. 11, 2010
Meeuwissen et al.
Figure 3. (I) High-pressure 31P{1H} NMR spectrum of [Rh(HL1-κP)(L1-κ2O,P)(CO)] at 10 bar of CO/H2. (II) 1H NMR spectrum at
20 bar of CO/H2. (III) 31P{1H} NMR spectrum at 20 bar of CO/H2. Conditions: [Rh] = 40 mM, temp = 25 °C, solvent = CD2Cl2. The
precursor complex [Rh(HL1-κP)(L1-κ2O,P)(CO)] is represented by the red diamonds.
species that is formed under hydroformylation conditions.15
An increase of the rhodium concentration from 1 to 15 mM
led to an increase of the conversion from 6.2 to 33.2%
(compare entries 1-4). However, when the rhodium concen-
tration exceeds 5-10 mM, the conversion does not increase
further presumably due to formation of inactive carbonyl-
bridged dimeric rhodium species.16 An increase of the partial
CO pressure from 10 to 30 bar gave a decrease of the
conversion from 26.5 to 18.7% (compare entries 2 and 6).
Such a trend is normally observed in this reaction;15 however,
an increase in the partial H2 pressure from 10 to 30 bar led to a
remarkably strong increase of the conversion from 26.5 to
62.0% (compare entries 2 and 5). This result indicates that an
increase of the partial H2 pressure promotes the formation of
the catalytically active hydride species from precursor com-
plex [Rh(HL1-κP)(L1-κ2O,P)(CO)], which is likely still pre-
sent in the reaction mixture under these conditions.
Accordingly, we studied the anticipated precursor com-
plex [Rh(HL1-κP)(L1-κ2O,P)(CO)] under hydroformylation
conditions via high-pressure NMR spectroscopy to observe
the formation of the hydride species. The 31P{1H} NMR
spectrum of the precursor complex in CD2Cl2 under a pres-
sure of syngas (10 bar) shows that the precursor complex
giving rise to the signals resonating at δ 63.0 and 81.9 ppm
remains present in solution (Figure 3, I, red diamonds). The
other major signals visible in the spectrum are a doublet
resonating at δ 49.8 ppm (JRh-P=163 Hz, blue squares), a
broad signal resonating at δ 51 ppm (black triangle), and a
doublet resonating at δ 58.1 ppm (JRh-P = 166 Hz, green
circle) (Figure 3, I). One of these signals most likely belongs
to the carbonyl-bridged dimeric rhodium species, as the
NMR experiment is done at relatively high concentrations.
In line with this, the color of the solution in the NMR tube
turned to orange under syngas pressure, which is typical
for this dirhodium species.16 The 1H NMR spectrum shows
a single hydride signal (triplets of doublets) resonating at
δ -9.74 ppm (1JH-Rh = 6.7 Hz, 2JH-P = 18 Hz), which indi-
cates that the hydride species is present in the solution
carrying two equivalent ligands (Figure 3, II).17 The rela-
tively small H-P coupling constant reveals that the two
ligands are predominantly coordinated in a bis-equatorial
geometry.18 The typical signals in the NMR spectra indicate
that the identical coordinated ligands are both neutral. To
convert precursor complex [Rh(HL1-κP)(L1-κ2O,P)(CO)]
into the hydride species, it reacts with a dihydrogen molecule
to form a rhodium hydride and to transform the anionic
chelating ligand into a neutral ligand. Increasing the pressure
in the NMR tube to 20 bar of syngas decreased the amount of
precursor complex in the solution in favor of the other
species, which indicates that the composition of the reaction
mixture is dependent on the pressure applied (Figure 3, III).
These results are in accordance with studies by Trzeciak and
ꢀ
ꢁ
ꢀ
Stepnicka on a related bis-ligated rhodium complex contain-
ing a neutral and an anionic O,P-chelating ligand.8b They
found that with this type of complexes the hydride species is
generated under hydroformylation conditions and also car-
ries two neutral monodentate ligands.8b Depressurization
of the solution in the NMR tube and changing to an
atmosphere of argon regenerated the precursor complex
[Rh(HL1-κP)(L1-κ2O,P)(CO)] almost completely. These re-
sults imply that the precursor complex and the hydride
species are in equilibrium, and the distribution of the species
involved in this equilibrium is influenced by the pressure
applied, in particular the partial pressure of H2. Indeed, an
increased hydroformylation activity of the precursor com-
plex was observed by increasing the partial pressure of H2,
which is now well explained by the increase in the concentra-
tion of the active species (Table 1).
The NMR studies indicate a relatively fast establishment
of the equilibrium between the hydride species and the
precursor complex [Rh(HL1-κP)(L1-κ2O,P)(CO)], which
suggests that this complex is active in the hydroformylation
reaction without a significant incubation time. To verify the
absence of an incubation time, the hydroformylation of
styrene using this complex was monitored via IR spectros-
copy in combination with GC chromatography. Indeed, the
kinetic profile shows immediate activity after pressurizing
the reaction mixture with syngas, and there is no observable
incubation time (Figure 4).10 Furthermore, the positive
order in alkene indicates that this system follows type I
kinetics and that the observed positive order in H2 in Table 1
is not related to type II kinetics, but merely to the equilibrium
between the precursor complex and the hydride species.15
(15) Van Leeuwen, P. W. N. M.; Claver, C., Eds. Rhodium Catalyzed
Hydroformylation; Kluwer Academic Publishers: Dordrecht, The
Netherlands, 2000.
(16) Chan, A. S. C.; Shieh, H.-S.; Hill, J. R. J. Chem. Soc., Chem.
Commun. 1983, 688–689.
(17) Attempts to obtain the proton-coupled 31P NMR spectrum in
order to correlate couplings with the 1H NMR spectrum failed, because
of too large broadening of the peaks.
(18) Buisman, G. J. H.; van der Veen, L. A.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M. Organometallics 1997, 16, 5681–5687.

Article
Organometallics, Vol. 29, No. 11, 2010 2417
seems that these static gas-phase DFT calculations, which
disregard solvent-assisted (de)protonation, are of no further
help to truly discriminate between these two possible mechan-
isms. Nonetheless, it is noteworthy that intermediate G**,
which occurs as a key intermediate in the ligand-assisted
mechanism, lies 16.3 kcal mol-1 lower in energy than the
intermediate G*, which is the key intermediate in the pre-
equilibrium mechanism. If a low-barrier (solvent assisted)
route from F to G exists, the ligand-assisted mechanism is
the preferred route for the reductive elimination of the
aldehyde. In that case, the equilibrium between precursor
complex A and hydride species B is integrated in the catalytic
cycle and a new cooperative ligand-assisted mechanism of
hydroformylation is operative.
Figure 4. Kinetic profile of the hydroformylation of styrene using
[Rh(HL1-κP)(L1-κ2O,P)(CO)]. Reaction conditions: [Rh] = 10
mM, Rh = [Rh(acac)(CO)2], [HL] = 2.1 mM, [S] = 0.2 M,
S = styrene, temp = 40 °C, pressure = 20 bar of 1:1 CO/H2,
solvent = CH2Cl2. The reaction was monitored in situ by way
of the typical vibrations of the aldehyde product that arise in the
IR spectrum with time. The intensity of the vibrations was
calibrated by an independent determination of the conversion at
the end of the reaction by means of GC (GC analysis after 23 h:
conversion = 99%).
Wehave demonstrated thatthe formation of[Rh(HL1-κP)-
(L1-κ2O,P)(CO)] involves a deprotonation of an acidic urea
proton of oneof the two phosphinourea ligands. Interestingly,
this potentially allows the controlled formation of a hetero-
bidentate ligand using two phosphinoureas having different
acidity of the urea proton. For this reason, phosphoramidite
phosphinourea HL2 was prepared via a direct condensation
reaction of N,N-dibenzylurea with a phosphorochloridite of
(S)-2,20-bisnaphthol in the presence of an excess of triethyla-
mine base. This new ligand was fully characterized by 1H, 13C,
and 31P NMR and mass spectrometry.10 By using equimolar
equivalents of HL1, HL2, and [Rh(acac)(CO)2] a metal com-
plex was formed that shows in the 31P{1H} NMR spectrum
two doublets of doublets (δ 63.3 and 170.9 ppm), indicating
that a heteroligated complex has been formed with a trans-
geometry of the two phosphorus donor atoms (Scheme 3, I).
The 31P{1H} NMR spectrum of the new complex is very
different from [Rh(HL1-κP)(L1-κ2O,P)(CO)] (δ 62.9 and
82.0 ppm). In particular the downfield signal at 170.9 ppm,
which belongs to HL2, indicates that this ligand forms the
anionic chelate instead of HL1 (Scheme 3, II). The 1H NMR
spectrum shows that acetylacetone (Hacac) is displaced from
the metal complex and thatoneurea proton isabstracted from
HL2. Proton-coupled 31P NMR and 31P{1H}-1H 2D-NMR
Although the equilibrium between the precursor com-
plex and the hydride species is established, the position of
this equilibrium in the catalytic cycle is still unclear. We
considered therefore two conceivable pathways, for which
we performed comparative DFT calculations starting from
a model precursor complex A: [Rh(HL-κP)(L-κ2O,P)(CO)],
L = 1,1-dimethyl-3-(diphenylphosphino)urea (Scheme 4).
The initial steps in the catalytic cycle are identical in both
pathways. Starting the catalytic cycle from the hydride
species B, the acyl species F is formed via a series of steps
involving the association of styrene to form C, migratory
insertion of styrene to give D, association of CO to form
E, and another migratory insertion step to form the acyl
species F.19 At this point, two different mechanisms are
conceivable, which places the equilibrium between A and B
within the hydroformylation cycle or as a equilibrium be-
tween a dormant state A and intermediate B.
Mechanism 1:. A pre-equilibrium mechanism. An inter-
molecular reaction involving the oxidative addition of H2
dihydride to F leads to species G*, which subsequently
regenerates B via reductive elimination of the aldehyde. In
this mechanism species A is in fact a dormant species in a pre-
equilibrium outside the productive catalytic cycle.
Mechanism 2:. A ligand-assisted mechanism analogous to
the reaction of phosphinourea ligands with [Rh(acac)(CO)2]
in Scheme 2, in which an acidic urea proton is transferred
from a phosphinourea ligand in F to the acyl group via the
metal center after an intramolecular acid-base reaction. In
this pathway, an intramolecular oxidative addition of a
phosphinourea ligand forms intermediate G**, after which
reductive elimination of the aldehyde regenerates species A.
In this case the equilibrium between A and B is part of the
catalytic cycle.
evidenced that the signal resonating at δ 63.3 ppm (1JP-Rh
=
131 Hz), corresponding to HL1, correlates to the urea proton
still visible in the 1H NMR spectrum. This result supports that
the signal resonating at δ 170.9 ppm (1JP-Rh = 190 Hz),
corresponds to HL2 as an anionic chelating ligand.10 So, the
ligand carrying the most acidic urea proton donates a proton to
the acac- ligand in an acid-base reaction and forms the anionic
chelate. This result confirms that by using two electronically
different phosphinoureas a heteroligated complex is formed
selectively. This provides a new tool in the area of combinatorial
catalysis, as a large number of assembled catalysts can be
prepared from a subset of well-chosen phosphinourea ligands.20
To demonstrate this new tool, we prepared rhodium com-
plexes using six phosphinoureas (HL1-HL6) in a combinatorial
(20) For other examples of catalysts made in combinatorial fashion in
rhodium-catalyzed hydroformylation see: (a) Goudriaan, P. E.; Kuil,
M.; Jiang, X.-B.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Dalton
Trans. 2009, 1801–1805. (b) Kuil, M.; Goudriaan, P. E.; Kleij, A. W.; Tooke,
D. M.; Spek, A. L.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Dalton Trans.
So far, wewere unable to find a simple (nonsolvent assisted)
intramolecular transition state to connect species F with G**,
because the urea proton from the phosphinourea in F pointsin
the opposite direction from the urea oxygen (Figure 5). It thus
€
2007, 2311–2320. (c) Slagt, V. F.; Roder, M.; Kamer, P. C. J.; van Leeuwen, P.
W. N. M.; Reek, J. N. H. J. Am. Chem. Soc. 2004, 126, 4056–4057. (d) Kuil,
M.; Goudriaan, P. E.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Chem.
Commun. 2006, 4679–4681. (e) Slagt, V. F.; van Leeuwen, P. W. N. M.;
Reek, J. N. H. Chem. Commun. 2003, 2474–2475. (f) Slagt, V. F.; van
Leeuwen, P. W. N. M.; Reek, J. N. H. Angew. Chem., Int. Ed. 2003, 42,
5619–5623. (g) Breit, B.; Seiche, W. Angew. Chem., Int. Ed. 2005, 44,
1640–1643. (h) Reetz, M. T.; Li, X. Angew. Chem., Int. Ed. 2005, 44,
2962–2964.
(19) Species outside the productive cycle are of minor importance to
discriminate between the two pathways and are therefore not discussed
here.

2418 Organometallics, Vol. 29, No. 11, 2010
Meeuwissen et al.
Scheme 3. (I) Reactivity of HL1 and HL2 with [Rh(acac)(CO)2]; (II) 31P{1H} NMR Spectrum of the Resulting Complex [Rh(HL1-
κP)(L2-κ2O,P)(CO)] (20 mM in CDCl3)
Scheme 4. DFT-Optimized Structures of the Two Mechanistic Pathways in the Hydroformylation of Styrene Using a Phosphinourea
Metal Complex via Mechanism 1 (A outside the productive cycle) and Mechanism 2 (A inside the productive cycle)a
a Reported energies are relative to species A in kcal mol-1. Nonproductive species have been omitted for clarity.
fashion, leading to 14 catalysts in total (6 homocombinations,
8 heterocombinations). These complexes were investigated in
the asymmetric hydroformylation of styrene.21 Besides HL1,
allotherligandsHL2-HL6 contain atleastone chiralelement
in their ligand structure (Figure 6). HL4 and HL6 have a urea
motif based on (R,R)-2,5-diphenylpyrrolidine, while HL2,
HL3, and HL5 contain a BINOL-based backbone. Ligand
HL6 contains both chiral elements.
The hydroformylation experiments were carried out at
40 °C in dichloromethane as a solvent at 20 bar of syngas
pressure, and samples were analyzed after 18 h of reaction.
Generally all catalysts gave good conversion (64.1-100%)
and selectivity (85.7-96.6%) for the branched product
(Table 2). The enantioselectivity displayed by the catalyst
depends strongly on the ligands used (ee’s between 0 and
46.4%). The homocombinations of the BINOL-based phos-
phoramidites HL2, HL3, HL5, and HL6 all lead to poor
selectivities, which suggest that this type of backbone is not
an effective chiral element in the these complexes for the
reaction studied. On the other hand, the homocombination
of HL4 was the most selective catalyst, providing an ee of
46.4%. Interestingly, this ligand contains the chirality on the
urea motif only, which is remote from the metal center.
Experiments in which the metal/ligand ratio was varied
showed that the use of three equivalents of HL4 gave
practically the same ee of 45.9%, while using only one
equivalent of HL4 caused the ee to decrease to 38.2%
(compare entries 4, 5, and 6). Ligand HL6, which contains
a chiral center at the urea motif next to the BINOL back-
bone, gave poor selectivity (ee 6.8%). The heteroligated
(21) For selected examples for the asymmetric hydroformylation of
styrene using traditional bidentate ligands see: (a) Sakai, N.; Mano, S.;
Nozaki, K.; Takaya, H. J. Am. Chem. Soc. 1993, 115, 7033–7034. (b) Yan,
Y.; Zhang, X. J. Am. Chem. Soc. 2006, 128, 7198–7202. (c) Robert, T.;
€
Abiri, Z.; Wassenaar, J.; Sandee, A. J.; Romanski, S.; Neudorfl, J.-M.;
Schmalz, H. G.; Reek, J. N. H. Organometallics 2010, 29, 478–483.

Article
Organometallics, Vol. 29, No. 11, 2010 2419
Figure 5. DFT (BP86, SV(P))-optimized structures of intermediates F (acyl species), G*(oxidative addition of H2), and G** (ligand-
assisted intermediate). CH hydrogen atoms are omitted for clarity. Rh: green, P: yellow, O: red, N: blue, C: black, H: white.
mylation reaction may well proceed via a new mecha-
nism involving ligand cooperativity. This mechanism in-
volves the reversible heterolytic splitting of H2 over the metal
center and an anionic phosphinourea fragment, and the
ligand-assisted reductive elimination of the aldehyde. Com-
putational studies support this ligand-assisted mechanism,
but so far we cannot exclude a mechanism in which a
dormant state and a rhodium hydride are in equilibrium
via a reversible heterolytic splitting of H2 over the metal
center, connecting the dormant state with the productive
hydroformylation cycle via a pre-equilibrium. The selective
formation of heterobidentate ligands of these phosphinour-
Figure 6. Structures of phosphinourea ligands HL1-HL6.
eas is potentially a valuable strategy for the creation of cata-
lyst libraries in a combinatorial fashion for high-throughput
screening due to the straightforward ligand synthesis. The
first results in this direction demonstrate that the applica-
tion of a small ligand library already leads to catalysts that
showed a high regioselectivity but so far moderate enantios-
electivity. Surely, if larger libraries of this type are con-
structed, more selective catalysts will be uncovered.
Table 2. Asymmetric Hydroformylation of Styrene Using
[Rh(HL-KP)(L-K2O,P)(CO)]a
entry
ligand
HL1
HL2
HL3
HL4
HL4
HL4
HL5
HL6
HL1 þ HL2
HL1 þ HL3
HL1 þ HL5
HL1 þ HL6
HL4 þ HL2
HL4 þ HL3
HL4 þ HL5
HL4 þ HL6
conversion (%)
branched (%)
ee (%)
1
2
3
4
73.1
100
95.6
90.9
85.7
96.5
96.6
96.4
93.9
94.5
93.8
93.2
94.3
93.6
94.2
94.0
95.1
94.2
3.2 (S)
0
80.6
74.7
99.1
64.1
100
46.4 (S)
38.2 (S)
45.9 (S)
3.4 (R)
6.8 (S)
4.5 (S)
0
2.7 (S)
5.4 (S)
24.5 (S)
10.8 (S)
25.0 (S)
16.9 (S)
5c
6d
7
Experimental Section
General Considerations. Unless stated otherwise, reactions
were carried out under an inert atmosphere of nitrogen or argon
using standard Schlenk techniques. THF was distilled from
sodium benzophenone ketyl; CH2Cl2 was distilled from CaH2,
and toluene was distilled from sodium under nitrogen. Triethy-
lamine was distilled from CaH2 under nitrogen. With the
exception of the compounds given below, all reagents were
purchased from commercial suppliers and used without further
purification. The following ligand precursor compounds were
synthesized according to published procedures: phosphoro-
chloridite of (S)- and (R)-2,20-bisnaphthol,22 phosphorochlor-
idite of (R)-3,30-dimethyl-2,20-bisnaphthol,23 and N,N-dibenzy-
lurea.24 NMR spectra were measured on a Varian Mercury (1H:
300 MHz, 31P{1H}: 121.5 MHz, 13C{1H}: 75.5 MHz) or a Varian
Inova spectrometer (1H: 500 MHz, 31P{1H}: 202.3 MHz,
13C{1H}: 125.7 MHz) using dry solvents at room temperature
unless stated otherwise. Chemical shifts are reported in ppm and
are given relative to TMS (1H and 13C{1H}) or H3PO4 (31P{1H})
as external standards. High-pressure NMR experiments were
conducted using a glass NMR tube (New Era Enterprises, NE-
HP5-M). 2D-NMR spectra were measured using a HMQC
8
9
100
96.7
79.7
100
10
11
12
13b
14b
15b
16
100
91.8
71.7
98.3
100
a Reaction conditions: [Rh] = 10 mM, Rh = [Rh(acac)(CO)2], [HL]
= 21 mM, [S] = 1 M, S = styrene, reaction time = 18 h, temperature =
40 °C, pressure = 20 bar CO/H2 (1:1), solvent = CH2Cl2. b Reaction
time = 16 h. c [HL] = 10 mM. d [HL] = 30 mM.
complexes were prepared from phosphorus amides HL1 and
HL4 in combination with the more acidic phosphoramidites
HL2, HL3, HL5, and HL6. Unfortunately, these complexes
did not yield more selective catalysts than the homocombi-
nations in most examples. The most selective heterocombi-
nations were HL4 þ HL2 and HL4 þ HL5, giving ee’s of 24.5
and 25.0%, respectively.
Conclusions
(22) Buisman, G. J. H.; van der Veen, L. A.; Klootwijk, A.; de Lange,
W. G. J.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Vogt, D.
Organometallics 1997, 16, 2929–2939.
(23) Huttenloch, O.; Laxman, E.; Waldmann, H. Chem.;Eur. J.
2002, 8, 4767–4780.
(24) Shi, F.; Smith, M. R.; Maleczka, R. E. Org. Lett. 2006,
1411–1414.
In summary, we report the synthesis of new phosphinour-
ea ligands, the preparation of supramolecular homo- and
heteroligated rhodium complexes thereof, and their use in
the asymmetric hydroformylation of styrene. The hydrofor-

2420 Organometallics, Vol. 29, No. 11, 2010
Meeuwissen et al.
pulse sequence without gradient. High-resolution mass spectra
(HRMS) were recorded at the department of mass spectrometry
at the University of Amsterdam using FABþ ionization on a
JEOL JMS SX/SX102A four-sector mass spectrometer with
3-nitrobenzyl alcohol as the matrix. IR spectra were recorded on
a Thermo Nicolet Nexus 670 FT-IR apparatus. High-pressure
IR were carried out in a stainless steel 50 mL autoclave equipped
with INTRAN windows (ZnS), a mechanical stirrer, a tempera-
ture controller, and a pressure indicator.
dissolved in 10 mL of THF was added dropwise to the reaction
mixture, which was stirred for 12 h. The product mixture
was filtered over a basic alumina plug. Solvents were evaporated
in vacuo. The product was redissolved in CH2Cl2 and filtered
over a silica plug. The solvents were evaporated in vacuo to
obtain the product. The ligands HL2, HL3, HL5, and HL6 were
obtained in a yield of 20-30% after filtration over silica.
HL2 ((S)-1,1-dibenzyl-3-(dinaphtho[2,1-d:10,20-f][1,3,2]dioxa-
1
phosphepin-4-yl)urea). White powder. H NMR (CD2Cl2, 500
Synthesis of (2R,5R)-2,5-Diphenylpyrrolidine-1-carboxamide.
(2R,5R)-2,5-Diphenylpyrrolidine (1.12 g, 1 equiv, 5 mmol) was
placed in a round-bottom flask. Then 200 mL of distilled water
was added, and the resulting suspension was vigorously stirred.
A 2.75 mL portion of a 2 M aqueous HCl solution (5.5 mmol, 1.1
equiv) was added dropwise to the reaction mixture. After
stirring for 1 h the reaction mixture turned clear and a solution
of 7.5 mmol of KOCN (0.61 g, 1.5 equiv) in distilled water was
added. The reaction mixture was next stirred for 48 h while the
product precipitated from the reaction mixture. The product
was filtered from the turbid reaction mixture and washed
3ꢀ with water and 3ꢀ with hexanes. Yield: 0.93 g (70%) as a
white powder. 1H NMR (300 MHz, CDCl3): δ 1.74 (d, 2H, CH2,
J = 6.3 Hz), 2.48 (br, 2H, CH2), 4.31 (s, 2H, NH2), 5.04 (br, 1H,
CH), 5.46 (br, 1H, CH), 7.2-7.5 (ArH, 10H). 13C{1H} NMR
(75.5 MHz, CDCl3): δ 31.342, 33.380, 62.157, 125.422, 125.923,
126.910, 128.107, 128.657, 129.336, 142.390, 143.846, 157.288
(NH2CON). HRMS (FABþ): m/z calcd for C17H19N2O
([MH]þ) 267.1497, obsd 267.1501.
General Procedure for Phosphinourea Synthesis (phosphorus
amide). N,N-Disubstituted urea (2 mmol, 1 equiv) was dried via
coevaporation with toluene (3ꢀ) and dissolved in 40 mL of
THF. After that the vigorously stirred reaction mixture was
cooled to 0 °C, and 2.2 mmol (1.1 equiv) of n-BuLi was added
dropwise. After 10 min 2 mmol of Ph2PCl (1 equiv) was added
dropwise to the vigorously stirred reaction mixture. The reac-
tion mixture was allowed to heat to room temperature and
stirred for 1 h. The reaction mixture was filtered over basic
alumina, which washed with 2 ꢀ 10 mL of THF. The solvents
from the collected filtrate were evaporated in vacuo. The product
was redissolved in 20 mL of CH2Cl2 and filtered over a silica
plug. The solvents of the collected filtrate were evaporated
in vacuo to obtain the product.
MHz): δ 4.325 (br d, 2H, CH2), 4.553 (br s, 2H, CH2), 5.687
(d, 1H, NH, 2JH-P = 6.5 Hz), 7.1-7.6 (m, 18H, ArH), 7.8-8.1
(m, 4H, ArH). 31P{1H} NMR (CD2Cl2, 121.5 MHz): δ 144.459.
31P NMR (CD2Cl2, 121.5 MHz): δ 144.459 (d, 2JP-H = 6.0 Hz).
13C{1H} NMR (CD2Cl2, 125.7 MHz): δ 50.300 (CH2), 121.420
(CH), 121.759 (CH), 123.606 (Cquat), 124.257 (Cquat), 124.301
(Cquat), 125.117 (CH), 125.271 (CH), 125.322 (CH), 126.366
(CH), 126.480 (CH), 126.608 (CH), 126.715 (CH), 127.330
(CH), 127.689 (CH), 128.246 (CH), 128.478 (CH), 128.518
(CH), 128.820 (CH), 129.056 (CH), 130.076 (CH), 130.718
(CH), 131.319 (Cquat), 131.762 (Cquat), 132.668 (Cquat), 132.762
(Cquat), 146.845 (Cquat), 146.875 (Cquat), 148.500 (Cquat), 157.185
2
(d, NHCON, JC-P = 15 Hz). HRMS (FABþ): m/z calcd
for C35H28O3N2P ([MH]þ) 555.1838, obsd 555.1832. Solution
IR (10 mM, CDCl3): ν = 3408 cm-1 (NHfree band), 1646 cm-1
(COurea band).
HL3 ((R)-3-(dinaphtho[2,1-d:10,20-f][1,3,2]dioxaphosphepin-4-
yl)-1,1-dimethylurea). White powder. 1H NMR (CDCl3, 500
MHz): δ 2.86 (s, 6H, CH3), 5.56 (d, 1H, NH, 2JH-P = 6.0 Hz),
7.2-7.6 (m, 8H, ArH), 7.9-8.1 (m, 4H, ArH). 31P{1H} NMR
(CDCl3, 202.3 MHz): δ 144.298. 13C{1H} NMR (CDCl3, 125.7
MHz): δ 36.424 (CH3), 121.542 (CH), 122.121 (CH), 123.933
(Cquat), 123.950 (Cquat), 124.323 (Cquat), 124.365 (Cquat), 125.151
(CH), 125.266 (CH), 126.441 (CH), 126.483 (CH), 126.908 (CH),
126.976 (CH), 128.317 (CH), 128.490 (CH), 128.515 (CH),
128.127 (CH), 129.852 (CH), 130.750 (CH), 131.302 (Cquat),
131.739 (Cquat), 132.820 (Cquat), 147.022 (Cquat), 147.054
(Cquat), 148.605 (Cquat), 156.906 (d, NHCON, 2JC-P = 15 Hz).
HRMS (FABþ): m/z calcd for C23H20N2O3P ([MH]þ) 403.1212,
obsd 403.1216.
HL5 ((R)-1,1-dibenzyl-3-(2,6-dimethyldinaphtho[2,1-d:10,20-f]-
[1,3,2]dioxaphosphepin-4-yl)urea). White powder. 1H NMR
(CD2Cl2, 500 MHz): δ 2.48 (s, 3H, CH3), 2.58 (s, 3H, CH3), 4.3
(br, 2H, CH2), 4.5 (br, 2H, CH2), 5.72 (d, 1H, NH, 2JH-P = 6.9
Hz), 7.1-7.9 (m, 20H, ArH). 31P{1H} NMR (CD2Cl2, 202.3
MHz): δ 143.787. 13C NMR (CD2Cl2, 75.5 MHz): δ 17.178,
17.422, 50.336, 123.418, 123.450, 124.297, 124.371, 124.961,
125.143, 125.249, 125.379, 126.480, 126.508, 127.407 (br),
127.593, 127.693, 128.765, 129.687, 130.075, 130.169, 130.190,
130.347, 131.110, 131.354, 131.373, 131.479, 11.500, 131.554,
136.708 (br), 146.218, 146.271, 147.520, 147.543, 157.206
HL1 (1,1-dibenzyl-3-(diphenylphosphino)urea). Yield: 47%,
white powder. 1H NMR (CDCl3, 500 MHz): δ 4.58 (s, 4H,
2
CH2), 5.04 (d, 1H, NH, JH-P = 6.9 Hz), 7.0-7.6 (m, 20H,
ArH). 31P{1H} NMR (CDCl3, 202.3 MHz): δ 25.518. 31P NMR
(CD2Cl2, 202.3 MHz): δ 25.518 (d, 2JP-H = 6.7 Hz). 13C{1H}
NMR (CDCl3, 125.7 MHz): δ 51.465 (CH2), 127.498 (CH),
127.739 (CH), 128.528 (CH), 128.579 (CH), 128.958 (CH),
129.207 (CH), 130.996 (CH), 131.171 (CH), 137.732 (Cquat),
139.713 (Cquat), 139.837 (Cquat), 158.264 (d, NHCON, 2JC-P
=
2
(d, NHCON, JC-P = 16 Hz). HRMS (FABþ): m/z calcd for
15 Hz). HRMS (FABþ): m/z calcd for C27H26N2OP ([MH]þ)
425.1783, obsd 425.1790. Solution IR (10 mM, CDCl3): ν =
3412 cm-1 (NH band), 1650 cm-1 (CO band).
C37H32O3N2P ([MH]þ) 583.2151, obsd 583.2154.
HL6 ((2S,5S)-N-((11bR)-dinaphtho[2,1-d:10,20-f][1,3,2]dioxa-
phosphepin-4-yl)-2,5-diphenylpyrrolidine-1-carboxamide). White
powder. 1H NMR (CD2Cl2, 500 MHz): δ 1.77 (s, 2H, CH2), 2.48
(br, 2H, CH3), 4.78 (s, 1H, CH), 5.25 (d, 1H, NH, 2JH-P = 5.5
Hz), 5.56 (s, 1H, CH), 7.0-7.6 (m, 18H, ArH), 7.8-8.2 (m, 4H,
ArH). 31P{1H} NMR (CD2Cl2, 202.3 MHz): δ 145.497. 13C{1H}
NMR (CD2Cl2, 75.5 MHz): δ 31.015 (CH2), 33.410 (CH2),
62.056 (CH), 121.446 (CH), 122.121 (CH), 123.872 (Cquat),
123.974 (Cquat), 124.016 (Cquat), 124.994 (CH), 125.043 (CH),
125.357 (CH), 126.229 (CH), 126.767 (CH), 126.904 (CH),
127.734 (CH), 128.314 (CH), 128.515 (CH), 128.871 (CH),
129.417 (CH), 130.410 (CH), 131.213 (Cquat), 131.478 (Cquat),
132.634 (Cquat), 132.737 (Cquat), 141.253 (Cquat), 143.421 (Cquat),
146.912 (Cquat), 146.942 (Cquat), 148.428 (Cquat), 154.932
HL4 ((2R,5R)-N-(diphenylphosphino)-2,5-diphenylpyrrolidine-
1-carboxamide). Yield: 75%, white powder. 1H NMR (300 MHz,
CD2Cl2): δ 1.81 (d, 2H, CH2, J = 5.7 Hz), 2.56 (s, 2H, CH2), 4.82
(d, 1H, NH, 2JH-P = 7.5 Hz), 5.10 (br, 1H, CH), 5.56 (br, 1H,
CH), 6.7-7.5 (ArH, 20H). 31P{1H} NMR (125.7 MHz, CD2Cl2):
δ 23.53. 13C{1H} NMR (75.5 MHz, CD2Cl2): δ 31.289, 33.400,
62.417, 125.381, 126.036, 126.733, 128.249, 128.331, 128.493,
128.584, 128.620, 129.302, 130.126, 130.406, 131.203, 131.499,
139.716, 139.785, 139.902, 140.007, 142.891, 144.204, 155.569 (d,
2
NHCON, JC-P = 18 Hz). HRMS (FABþ): m/z calcd for
C29H28N2OP ([MH]þ) 451.1939, obsd 451.1935.
General Procedure for Phosphinourea Synthesis (phospho-
ramidite). N,N-Disubstituted urea (2 mmol, 1 equiv) was dried
via coevaporation with toluene (3ꢀ), dissolved in 40 mL of
THF, and cooled to 0 °C, and an excess of dry triethylamine was
added. After 10 min 2 mmol of phosphorchloridite (1 equiv)
2
(d, NHCON, JC-P = 15 Hz). HRMS (FABþ): m/z calcd for
C37H30O3N2P ([MH]þ) 581.1994, obsd 581.1999.
General Procedure for Complex Synthesis. To a Schlenk filled
with 1 mmol (2 equiv) of phosphinourea ligand and 0.5 mmol

Article
Organometallics, Vol. 29, No. 11, 2010 2421
(129 mg, 1 equiv) of [Rh(acac)(CO)2] was added 10 mL of
CH2Cl2. Instantaneously CO gas was released from the clear
yellow reaction mixture. After five minutes of vigorous stirring,
solvents were evaporated in vacuo. Hacac (acetylacetone) pre-
sent as a side-product was removed via coevaporation with
toluene (3ꢀ). The complexes were formed in high purity (>95%
based on 31P{1H} NMR) and were used without further puri-
fication.
samples were prepared and analyzed directly after the reaction
by GC chromatography using an Interscience Focus GC
equipped with a Supelco beta dex 225 column.
X-ray Crystallography of HL1. C27H25N2OP, Mw = 424.46,
colorless rod, 0.32 ꢀ 0.14 ꢀ 0.10 mm3, triclinic, P1 (No. 2), a=
9.3963(4) A, b=10.0852(4) A, c=20.8958(12) A, R=92.021(2)°,
o
3
ꢂ
β = 100.289(2)°, γ = 108.447(2) , V = 2203.99(18) A , Z = 4,
D
exptl=1.279 g/cm3, μ(Mo KR)=0.147 mm-1, T=110 K. X-ray
Complex [Rh(HL1-κP)(L1-κ2O,P)(CO)]. Yellow-orange pow-
der. 1H NMR (CD2Cl2, 500 MHz): δ 4.24 (s, 6H, CH2), 4.81 (s,
2H, CH2), 6.4 (d, 1H, NH, 2JH-P=16.5 Hz), 7.0-7.8 (m, 40H,
data were collected (θ(max)=27.5°; total =57 422; unique =
10124) on a Nonius KappaCCD with graphite-monochromated
Mo KR radiation (λ = 0.71073 A) under program control of
COLLECT.25 The program PEAKREF26 was used to deter-
mine the cell dimensions. Data reduction was done using the
program EVALCCD.27 Absorption correction based on face-
indexing was done with SADABS.28 The structure was solved
with SHELXS-9729 and refined with SHELXL-97.29 Hydrogen
atoms were introduced at calculated positions and refined riding
on their carrier atoms with U(iso) = 1.2 ꢀ Ueq of the attached C
atom. The H atoms attached to N2A and N2B were located
from a difference map and refined with a distance restraint
[d(N-H) - 0.88 A]. Final R = 0.0415 for 7278 reflections
with I > 2 σ(I), wR2 = 0.0948 for 10 124 reflections, S = 1.02,
ArH). 31P{1H} NMR (CD2Cl2, 202.3 MHz): δ 62.9 (dd, 1JRh-P
=
135 Hz, 2JP-P=326 Hz), 82.0 (dd, 1JRh-P=129 Hz, 2JP-P=326
Hz). 31P NMR (CDCl3, 202.3 MHz): δ 62.9 (dddt, 1JRh-P=135
Hz, 2JP-P=326 Hz, 2JP-H=14.6 Hz, 3JAr-H=11.7 Hz), 82.0 (ddt,
1JRh-P = 129 Hz, 2JP-P = 326 Hz, 3JAr-H = 11 Hz). 13C{1H}
NMR (CD2Cl2, 125.7 MHz): δ 49.605 (CH2), 50.406 (CH2),
126.877 (CH), 127.492 (CH), 127.573 (CH), 127.703 (CH),
127.796 (CH), 128.072 (CH), 128.158 (CH), 128.306 (CH),
128.391 (CH), 128.731 (CH), 129.846 (CH), 130.541 (CH),
131.301 (CH), 131.402 (CH), 132.934 (CH), 133.057 (CH),
133.625 (Cquat), 134.003 (Cquat), 136.948 (Cquat), 138.682 (Cquat),
138.755 (Cquat), 139.218 (Cquat), 155.424 (d, Cquat, 2JC-P=7.4 Hz,
N
param = 565. The final difference map is essentially featureless
2
2
(NHCON), 175.5 (dd, Cquat, JC-Rh =27 Hz, JC-P =8.8 Hz,
with excursions in the range -0.36 to 0.43 e/A3. The structure
was validated with PLATON/CHECK.30 The cif file of L1 is
supplied in the Supporting Information.
NCON), 192 (br d, Cquat
,
1JC-Rh ≈ 73 Hz, Rh-CO). HRMS
(FABþ): m/z calcd for C54H49N4O2P2Rh ([MH]þ) 950.2386, obsd
950.2380 (CO ligand dissociated). Solution IR (10 mM, CDCl3):
ν=3381 cm-1 (NHassoc. band), 1970 cm-1 (Rh-CO band), 1663
cm-1 (COurea band).
DFT Calculations. The geometry optimizations were carried
out with the Turbomole program31 coupled to the PQS Baker
optimizer.32 All geometries were optimized at the BP8633 level
using the SV(P) basis set34 on all atoms (small-core
pseudopotential35 on rhodium). Optimized geometries of the
species are supplied as a pdf file in the Supporting Information.
Complex [Rh(HL2-κP)(L2-κ2O,P)(CO)]. Yellow-orange pow-
der. 1H NMR (CD2Cl2, 500 MHz): δ 4.2 (m, 4H, CH2), 4.56 (s,
3H, CH2), 5.88 (s, 1H, CH2), 7.2-8.2 (m, 45H, ArHþNH).
31P{1H} NMR (CD2Cl2, 202.3 MHz): δ 149.4 (dd, 1JRh-P=213
Hz, 2JP-P=674 Hz, 170.0 (dd, 1JRh-P = 189 Hz, 2JP-P = 672
Hz). HRMS (FABþ): m/z calcd for C70H54N4O6P2Rh ([MH]þ)
1211.2574, obsd 1211.2566 (CO ligand dissociated).
Acknowledgment. This research is financially sup-
ported by BASF The Netherlands and the Dutch Ministry
of Economic Affairs by way of SenterNovem. J. W. H.
Peeters is acknowledged for the mass spectrometry experi-
ments, J. M. Ernsting for assistance with the 2D-NMR
experiments, and Z. Abiri for providingN,N-dibenzylurea.
Complex [Rh(HL1-κP)(L2-κ2O,P)(CO)]. Yellow-orange pow-
der. 1H NMR (CD2Cl2, 500 MHz): δ 3.76 (d, 1H, CH2, J = 15.5
Hz), 4.07 (d, 1H, CH2, J = 15 Hz), 4.2 (br, 4H, CH2) 4.46 (d, 1H,
CH2, J=15 Hz), 4.95 (d, 1H, CH2, J=15.5 Hz), 6.33 (d, 1H, NH,
2JH-P=17 Hz) 6.8-8.0 (m, 42H, ArH). 31P{1H} NMR (CD2Cl2,
1
2
Supporting Information Available: This material is available
free of charge via the Internet at http://pubs.acs.org.
202.3 MHz): δ 63.3 (dd, JRh-P = 131 Hz, JP-P = 451 Hz),
170.9 (dd, 1JRh-P=190 Hz, 2JP-P=451 Hz). 31P NMR (CD2Cl2,
1
2
202.3 MHz): δ 63.3 (dddt, JRh-P = 131 Hz, JP-P = 451 Hz,
2JP-H = 13.8 Hz, 3JAr-H=12.5 Hz) 170.9 (dd, 1JRh-P=190 Hz,
2JP-P= 451 Hz). HRMS (FABþ): m/z calcd for C62H52O4N4P2Rh
([MH]þ) 1081.2519, obsd 1081.2521 (CO ligand dissociated).
General Procedure for Hydroformylation Experiments. The
hydroformylation experiments were carried out in a stainless
steel autoclave (volume 150 mL) with an insert for five 2 mL
glass reaction vials. The reaction vials were dried in an oven
(135 °C), capped with a septum, and flushed with inert gas
(argon) before use. In each reaction vial the catalyst was made
in situ by charging the reaction vial with the appropriate
amounts of metal precursor and ligands before it was flushed
again with inert gas. The substrate (styrene) was filtered freshly
over basic alumina to remove peroxide impurities. The substrate
was dissolved in CH2Cl2 at the appropriate molarity and was
injected into the reaction vials. The insert containing the reac-
tion vials was placed in the autoclave, which was closed thor-
oughly. Before starting the catalytic reaction, the prepared
autoclave was purged three times with 10 bar of syngas
(CO/H2 = 1:1) and then pressurized to the appropriate pres-
sure. The pressure in the autoclave during the reaction was
monitored by either an analog or digital readout. The reac-
tion was stopped by cooling the autoclave with ice, and the
autoclave was subsequently slowly depressurized. Reaction
(25) COLLECT; Nonius BV: Delft, The Netherlands, 1999.
(26) Schreurs, A. M. M. PEAKREF; Utrecht University: Utrecht, The
Netherlands, 2005.
(27) Duisenberg, A. J. M.; Kroon-Batenburg, L. M. J.; Schreurs, A.
M. M. J. Appl. Crystallogr. 2003, 36, 220–229.
(28) Sheldrick, G. M. SADABS 2006/1; University of Gottingen:
€
€
Gottingen, Germany, 2006.
(29) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112–122.
(30) Spek, A. L. Acta Crystallogr. Sect. D 2009, 65, 148–155.
€
(31) (a) Ahlrichs, R.; Bar, M.; Baron, H.-P.; Bauernschmitt, R.;
Bocker, S.; Ehrig, M.; Eichkorn, K.; Elliott, S.; Furche, F.; Haase, F.;
€
€
€
€
Haser, M.; Hattig, C.; Horn, H.; Huber, C.; Huniar, U.; Kattannek, M.;
€
Kohn, A.; Kolmel, C.; Kollwitz, M.; May, K.; Ochsenfeld, C.; Ohm, H.;
€
Schafer, A.; Schneider, U.; Treutler, O.; Tsereteli, K.; Unterreiner, B.;
€
von Arnim, M.; Weigend, F.; Weis, P.; Weiss, H. Turbomole Version 5;
Theoretical Chemistry Group, University of Karlsruhe, 2002. (b) Treutler, O.;
Ahlrichs, R. J. Chem. Phys. 1995, 102, 346–354.
(32) (a) PQS version 2.4; Parallel Quantum Solutions: Fayetteville, AR,
2001 (the Baker optimizer is available separately from PQS upon request).
(b) Baker, J. J. Comput. Chem. 1986, 7, 385–395.
(33) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100. (b) Perdew,
J. P. Phys. Rev. B 1986, 33, 8822–8824.
€
(34) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
2571–2577.
(35) (a) Turbomole basis set library, Turbomole Version 5. (b) Andrae,
D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta
1990, 77, 123–141.