Rhodium-catalyzed hydroformylation and deuterioformylation with pyrrolyl-based phosphorus amidite ligands: Influence of electronic ligand properties

Article abstract of DOI:10.1021/om010760y

The influence of electronic ligand properties on the catalyst performance in the rhodium-catalyzed hydroformylation of alkenes was investigated. Two bidentate phosphorous amidite and phosphinite ligands were also synthesized. The results show that all lin

Full text of DOI:10.1021/om010760y

Organometallics 2002, 21, 3873-3883
3873
Rh od iu m -Ca ta lyzed Hyd r ofor m yla tion a n d
Deu ter iofor m yla tion w ith P yr r olyl-Ba sed P h osp h or u s
Am id ite Liga n d s: In flu en ce of Electr on ic Liga n d
P r op er ties
Saskia C. van der Slot, J osep Duran,† J ordy Luten, Paul C. J . Kamer, and
Piet W. N. M. van Leeuwen*
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 166,
1018 WV Amsterdam, The Netherlands
Received August 17, 2001
The influence of electronic ligand properties on the catalyst performance in the rhodium-
catalyzed hydroformylation of alkenes has been investigated. Two bidentate phosphorus
amidite and phosphinite ligands have been synthesized: 1,1′-biphenyl-2,2′-diyl-bis(dipyr-
rolylphosphoramidite) (3) and 1,1′-biphenyl-2,2′-diyloxy-bis(diphenylphosphinite) (4). Their
monodentate analogues have also been studied: phenyldipyrrolylphosphoramidite (1) and
phenyl diphenylphosphinite (2). These two sets of ligands have very similar steric properties
but the amidites are much stronger π-acceptor ligands. Spectroscopic studies showed that
under hydroformylation reaction conditions the monodentate ligands 1 and 2 form mixtures
of HRhL2(CO)2 and HRhL3(CO) complexes depending on the ligand and rhodium concentra-
tions and the carbon monoxide pressure. Depending on the reaction conditions, the bidentate
ligands 3 and 4 form mixtures of HRh(L∩L)(CO)2 and HRh(L∩L)(L∩L′)(CO), where L∩L′
functions as a monodentate. All ligands have been tested in the hydroformylation reaction
of oct-1-ene. A high π-acidity of the ligand resulted in a high rate of hydroformylation. The
monodentate ligands 1 and 2 showed moderate selectivity for the linear aldehyde. The
catalyst formed with the bidentate phosphorus amidite ligand 3 revealed high regioselectivity
for the linear aldehyde (ratio l/b ) ∼100) at a high rate together with a moderate selectivity
for isomerization (∼7%). Deuterioformylation experiments of 1-hexene showed that the
hydride (deuteride) migration is reversible in the hydroformylation system formed by 3.
Surprisingly, both the linear rhodium-alkyl and the branched rhodium-alkyl complex
undergo â-hydride elimination. Furthermore, the 2-hexylrhodium intermediate regenerates
more often monodeuterated 1-hexene than 2-hexene. The rhodium hydride species formed
this way reacts relatively slowly with the excess of D2 and as a result large amounts of
monodeuterated heptanal (40% D1 vs 60% D2) and monodeuterated 1-hexene are formed.
At higher conversions the latter gives trisdeuterated heptanal as well as bisdeuterated
heptanal.
In tr od u ction
to the steric and electronic ligand properties and, for
bidentate ligands, the bite angle of the ligands.6-12
Many examples of modified phosphine ligands are
known in the literature.1,8,10 These ligands sometimes
form very selective catalysts although the activity in the
hydroformylation reaction is often moderate. Many
examples of phosphite ligands are known that form
active hydroformylation catalysts,1 but these ligands
One of the key issues in hydroformylation research
is the development of new ligands to obtain highly active
and selective catalysts.1-5 Attention has been devoted
* To whom correspondence should be addressed. E-mail: pwnm@
science.uva.nl.
† Current address: Departament de Quimica, Facultat de Ciencies,
Universitat de Girona, Campus de Montilivi, 17071 Girona, Spain.
(1) See for references about phosphorus ligand modified hydroformyl-
ation the following books and reviews and references herein: (a) van
Leeuwen, P. W. N. M.; Casey, C. P.; Whiteker, G. T. Rhodium
Catalysed Hydroformylation; van Leeuwen, P. W. N. M., Claver, C.,
Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2000;
Chapter 4. (b) Agbossou, F.; Carpentier, J . F.; Mortreux, A. Chem. Rev.
1995, 95, 2485. (c) Beller, M.; Cornils, B.; Frohning, C. D.; Kohlpaint-
ner, C. W. J . Mol. Catal. A 1995, 104, 17.
(5) (a) van Leeuwen, P. W. N. M.; Roobeek, C. F. J . Organomet.
Chem. 1983, 258, 343. (b) J ongsma, T.; Challa, G.; van Leeuwen, P.
W. N. M. J . Organomet. Chem. 1991, 421, 121. (c) Van Rooy, A.; Orij,
E. N.; Kamer, P. C. J .; Van den Aardweg, F.; Van Leeuwen, P. W. N.
M. J . Chem. Soc., Chem. Commun. 1991, 1096. (d) Van Rooy, A.; Orij,
E. N.; Kamer, P. C. J .; Van Leeuwen, P. W. N. M. Organometallics
1995, 14, 34. (e) Magee, M. P.; Luo, W.; Hersh, W. H. Organometallics
2002, 21, 362. (f) Breit, B.; Winde, R.; Mackewitz, T.; Paciello, R.;
Harms, K. Chem. Eur. J . 2001, 7, 3106. (g) J ackstell, R.; Klein, H.;
Beller, M.; Wiese, K.-D.; Rottger, D. Eur. J . Org. Chem. 2001, 20, 3871.
(h) Breit, B. Chem. Commun. 1996, 2071.
(2) Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B.
R.; Matter, B. A.; Powell, D. R. J . Am. Chem. Soc. 1999, 121, 63.
(3) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.;
Gavney, J . A., J r.; Powell, D. R. J . Am. Chem. Soc. 1992, 114, 5535.
(4) van der Veen, L. A.; Boele, M. D. K.; Bregman, 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.
(6) Unruh, J . D.; Christenson, J . R. J . Mol. Catal. 1982, 14, 19.
(7) Moser, W. R.; Papile, C. J .; Brannon, D. A.; Duwell, R. A. J .
Mol. Catal. 1987, 41, 271.
10.1021/om010760y CCC: $22.00 © 2002 American Chemical Society
Publication on Web 08/21/2002

3874 Organometallics, Vol. 21, No. 19, 2002
van der Slot et al.
Sch em e 1. Mon od en ta te a n d Bid en ta te
introduce less steric hindrance close to the rhodium
center compared to phosphines, which can lead to
moderate selectivity for the linear aldehyde. Low selec-
tivity but extremely high rates can be obtained by using
bulky monophosphite ligands.5a-d In recent years espe-
cially ligands that are good electron acceptors have been
found to be effective ligands.5e-g Bulky bidentate phos-
phites can give high selectivity to linear aldehydes, but
these catalysts are less active than bulky monophos-
phites.13 Phosphorus amide ligands introduce more
steric bulk close to the phosphorus atom compared to
the electron-withdrawing phosphites because of the high
degree of substitution of nitrogen compared to oxygen.
For dialkylamido phosphorus ligands the ø-values are
typically low, but when suitably substituted the ø-val-
ues14 of phosphorus amidites can be in the same high
range as those of phosphites.5,15 Therefore, a suitable
combination of steric and electronic properties will
result in crowded, electron-poor catalysts, which might
increase both the activity and selectivity of the hydro-
formylation catalyst. In recent years a few examples of
π-acidic phosphorus amides as ligands in the hydro-
formylation reaction have been published.5,16-20 The
results obtained with these ligands showed improved
activity compared to phosphine systems and improved
selectivity compared to phosphite systems. Moloy and
co-workers found that pyrrolyl substituents on the
phosphorus atom lead to highly π-acidic phosphorus
ligands.15 Zio´lkowski and Beller and co-workers inves-
tigated the rhodium-catalyzed hydroformylation reac-
tion using several monodentate N-pyrrolylphosphine
ligands.5g,17 Here we report on novel monodentate and
bidentate phosphorus amidite ligands that contain two
pyrrolyl substituents (1, 3). The performance of the
hydroformylation catalyst based on these ligands will
be compared to that of similar diphenyl-substituted
phosphinite ligands (2, 4). Substitution of the phenyl
substituent for a pyrrolyl substituent has a large
influence on the π-acidity of the ligand. By using these
ligands (1-4), the electronic effect on the catalyst
performance can be investigated with a minimal change
in steric influences. The electronic properties of the
ligands have been determined by comparison of the IR
P h osp h or u s Am id ite a n d P h osp h in ite Liga n d
Str u ctu r es
Ta ble 1. IR F r equ en cies of th e Ca r bon yl Liga n d s
of Va r iou s Rh Cl(CO)L2
a
ligand
νCO (cm-1
)
PPh3
1978b
1991
Ph2P(OPh) (2)
(pyrrolyl)2PPh
P(OPh)3
(pyrrolyl)2POPh (1)
P(pyrrolyl)3
2007c
2016b
2019
2024c
Measured in CH2Cl2. See refs 26 and 27. c See ref 16.
a
b
stretching frequency of the carbonyl ligand in trans-
RhCl(CO)L2 (L ) 1, 2). The solution structures of the
rhodium complexes formed under hydroformylation
conditions were studied and these complexes were
tested in the hydroformylation reaction of alkenes. The
reaction mechanism was studied with use of ligand 3
in the deuterioformylation of 1-hexene.
Resu lts a n d Discu ssion
Syn th esis a n d Deter m in a tion of th e Electr on ic
Liga n d P r op er t ies. Phosphorus ligands (1-4) were
prepared by the reaction of either chlorodipyrrolylphos-
phine or chlorodiphenylphosphine and the correspond-
ing alcohol in the presence of a base. The ligands were
purified by crystallization, washing, or column chroma-
tography and stored under an inert atmosphere, al-
though they are relatively stable toward water and
oxygen. To investigate the electronic properties of these
phosphorus amidite and phosphinite ligands, trans-
RhCl(CO)(1)2 (5) and trans-RhCl(CO)(2)2 (6) were syn-
thesized. The carbonyl frequency in the IR spectrum is
a measure of the π-acidity of ligands 1-4.15,23 The IR
frequencies of the carbonyl ligands of the trans-RhCl-
(CO)L2 (L ) 1, 2) complexes are presented in Table 1.
The position of the carbonyl frequency of trans-RhCl-
(CO)(1)2 (5) in the IR spectrum shows that the electronic
character of ligand 1 is similar to that of triphenyl
phosphite. Comparison of the carbonyl frequency of
trans-RhCl(CO)(2)2 (6) with that of the trans-RhCl(CO)-
(8) van der Veen, L. A.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.
Organometallics 1999, 18, 4765.
(9) Cuny, G. D.; Buchwald, S. L. J . Am. Chem. Soc. 1998, 120, 11616.
(10) Wink, D. J .; Kwok, T. J .; Yee, A. Inorg. Chem. 1990, 29, 5006.
(11) Billig, E.; Abatjoglou A. G.; Bryant, D. R., U.S. Patent 4,668,-
651; European Patent. Application 213,639, 1987 (to Union Carbide);
Chem. Abstr. 1987, 107, 7392.
(12) Billig, E.; Abatjoglou, A. G.; Bryant, D. R.; Murray, R. E.;
Maher, J . M., U.S. Patent 4,599,206, 1986 (to Union Carbide); Chem.
Abstr. 1988, 109, 233177.
(13) (a) van Rooy, A.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.;
Goubitz, K.; Fraanje, F.; Veldman, N.; Spek, A. L. Organometallics
1996, 15, 835. (b) Billig, E.; Abatjoglou, A. G.; Bryant, D. R.; Murray,
R. E.; Maher, J . M., U.S. Patent 4,599,206, 1986 (to Union Carbide);
Chem. Abstr. 1988, 109, 233177.
(14) Moasser, B.; Gladfelter, W. L.; Roe, D. C. Organometallics 1995,
14, 3832.
(15) Tolman, C. A. J . Am. Chem. Soc. 1970, 92, 2953
(16) Moloy, K. G.; Petersen, J . L. J . Am. Chem. Soc. 1995, 117, 7696
(17) van der Slot, S. C.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.;
Fraanje, J .; Goubitz, K.; Lutz, M.; Spek, A. L. Organometallics 2000,
19, 2504.
(18) van Leeuwen, P. W. N. M.; Roobeek, C. F. Tetrahedron 1981,
37, 1973.
(19) Francio´, G.; Faraone, F.; Leitner, W. Angew. Chem., Int. Ed.
2000, 39, 1428.
(21) Li Wu, M.; Desmond, M. J .; Drago, R. S. Inorg. Chem. 1979,
18, 679.
(22) Trzeciak, A. M.; Glowiak, T.; Grybek, R.; Zio´lkowski, J . J . Chem.
Soc., Dalton Trans. 1997, 1831.
(20) Clarke, M. L.; Cole-Hamilton, D. J .; Slawin, A. M. Z.; Woollins,
J . D. Chem. Commun. 2000, 2065.
(23) Naili, S.; Mortreux, A.; Agbossou, F. Tetrahedron Asymmetry
1998, 3421.

Rhodium Hydroformylation of Pyrrolyl Amidite Ligands
Organometallics, Vol. 21, No. 19, 2002 3875
Ta ble 2. Sp ectr oscop ic Da ta of Sever a l HRh (L)x(CO)4-x Com p lexes (x ) 2, 3)
b
complex
HRh(1)2(CO)2
HRh(2)2(CO)2
HRh(PPh3)2(CO)2
HRh(P(OPh)3)2(CO)2
HRh(1)3(CO)
δ(1H) (ppm)a
δ(31P) (ppm)a
J
HP (Hz)
J HRh (Hz)
J RhP (Hz)
νCO (cm-1
)
-10.1 (br)
-9.5 (br)
137
145
37
n.d.
n.d.
n.d.
n.d.
219
165
139
2076, 2024
2045, 1970
2042, 1981
2070, 2018
2063
c
c
-9.8 (dq)
-9.4 (dq)
-9.1 (m)
-9.1 (dq)
-10.9 (dq)
134
142
47
109
141
7
3
3
n.d.
3
3
217
169
133
211
240
HRh(2)3(CO)
13
n.d.
8
1943
HRh(PPh3)3(CO)d
2040e
HRh(P(pyrrolyl)3)3(CO)d
HRh(P(OPh)3)3(CO)d
2079e
3
2060e
Measured in toluene-d8 or benzene-d6, dq ) double quartet. Measured in cyclohexane. c See ref 29. See ref 21. e Measured in KBr.
a
b
d
(PPh3)2 complex shows that the electron-donating char-
nal) carbonyl frequencies, indicating that only one
isomer is present (either ee or ea ). While by NMR
spectroscopy often fast equilibration between ee and ea
isomers cannot be halted, two sets of two carbonyl
frequencies are observed in the IR spectrum when two
isomers are present.8 The positions of the carbonyl
frequencies of HRh(1)2(CO)2 are similar to those ob-
served for the ee-coordinated rhodium-bis-triphenyl
phosphite complex and the positions of the carbonyl
frequencies of HRh(2)2(CO)2 are comparable to those of
the ee-coordinated rhodium-bis-triphenylphosphine
complex (see Table 2).25 The J RhP coupling constant of
HRh(1)2(CO)2 was in the same range as those found for
similar phosphite ligands in equatorial positions (219
Hz), while the J RhP coupling constant of HRh(2)2(CO)2
was comparable to those in complexes having arylphos-
phines in equatorial positions (165 Hz).
Depending on the ligand concentration, a significant
amount of HRhL3(CO) (L ) 1, 2) was formed under the
conditions used for the spectroscopic studies. These
hydride complexes have relatively large J PH coupling
constants for a pure cis coordination (7 and 13 Hz
respectively), indicating a slight distortion of the geom-
etry of the trigonal bipyramidal structure.26,27 Again,
the J RhH coupling constant (3 Hz) and the J RhP coupling
constants observed for the rhodium-hydride complexes
are in the same range as those found for similar
phosphine- or phosphite-containing hydride complexes
(Table 2).21 The positions of carbonyl frequencies ob-
served in the high-pressure IR spectra are consistent
with the π-acidity of the ligands.
Analogous to the results observed for the monodentate
ligands 1 and 2, mixtures of two different hydride
complexes were found for the bidentate ligands 3 and 4
depending on the ligand and rhodium concentrations
and the carbon monoxide pressure. In the NMR experi-
ments ([Rh] ) 10 mM), a ligand-to-rhodium ratio of 1
is sufficient to obtain complete conversion to the hydride
complex. For the high-pressure IR experiments, lower
rhodium concentrations are used and therefore higher
ligand concentrations are a prerequisite to obtain
complete conversion to the rhodium-hydride complex.
With use of ligand 3 in these experiments, a ligand-to-
rhodium ratio of 1.5 is enough, but for ligand 4 a ligand-
to-rhodium ratio of 3 is necessary for complete conver-
sion of the rhodium precursor (Rh(acac)(CO)2) to the
acter of ligand 2 is in the same range as that of
triphenylphosphine. Comparison of the carbonyl fre-
quencies of trans-RhCl(CO)(1)2 (5) and trans-RhCl(CO)-
(2)2 (6) shows that ligand 1 is a stronger π-acid than
ligand 2, whereas ligands 1 and 2 will have comparable
steric ligand properties, 1 being slightly smaller than
2. The electronic properties of the bidentate ligands 3
and 4 are very similar to those of ligands 1 and 2,
respectively, because of the similarity of the substitu-
ents used.
Ca ta lyst Ch a r a cter iza tion Stu d ies. The structures
of the complexes formed under hydroformylation condi-
tions were studied with high-pressure (HP) NMR and
IR spectroscopy. The catalyst was formed by using Rh-
(acac)(CO)2 in the presence of various ligand concentra-
tions at 20 bar of CO/H2 (1/1). The spectroscopic data
of the hydride complexes observed are presented in
Table 2. The monodentate ligands gave two different
complexes under syn-gas pressure: HRhL2(CO)2 and
HRhL3(CO). The ratio between these two complexes
depends on the concentrations of rhodium, ligand, and
carbon monoxide. Higher rhodium and ligand concen-
trations resulted in a larger proportion of HRhL3(CO).
The presence of two complexes is not unusual for these
relatively small monodentate ligands; similar results
have been obtained for triphenylphosphine and triphen-
yl phosphite.24,25
Rhodium complexes of the form HRhLx(CO)4-x (x )
1, 2, 3) generally have a (distorted) trigonal bipyramidal
structure. Detailed NMR and IR studies have shown
that the hydride ligand coordinates at an apical position
of the bipyramid. The phosphorus atoms can coordinate
at either the equatorial positions or the remaining apical
position. The coordination mode of the phosphorus
ligands can be derived from the magnitude of the J PH
NMR coupling constant as has been shown in many
catalyst characterization studies.3,4,31
The hydride resonances observed for the complexes
containing two monodentate ligands (HRhL2(CO)2, L )
1, 2) are broad (see Table 2). The resonance remained
broad when the sample was cooled to -70 °C. Therefore
the structure of these complexes could not be deter-
mined by using the J PH and J RhH coupling constants.
The IR spectra of these complexes showed two (termi-
(24) Breit, B. J . Mol. Catal. 1999, 143, 143.
(25) Pottier, Y.; Mortreux, A.; Petit, F. J . Organomet. Chem. 1989,
370, 333.
(26) Haar, C. M.; Huang, J .; Nolan, S. P.; Petersen, J . L. Organo-
metallics 1998, 17, 5018.
(29) Die´guez, M.; Claver, C.; Masdeu-Bulto´, A. M.; Ruiz, A.; van
Leeuwen, P. W. N. M.; Schoemaker, G. C. Organometallics 1999, 18,
2107.
(27) Li Wu, M.; Desmond, M. J .; Drago, R. S. Inorg. Chem. 1979,
18, 679.
(28) Evans, D.; Osborn, J . A.; Wilkinson, G. J . Chem. Soc. A 1968,
3133.
(30) Buisman, G. J . H.; Vos, E. J .; Kamer, P. C. J .; van Leeuwen, P.
W. N. M. J . Chem. Soc., Dalton Trans. 1995, 409.
(31) Nettekoven, U.; Kamer, P. C. J .; Wildhalm, M.; van Leeuwen,
P. W. N. M. Organometallics 2000, 19, 4596.

3876 Organometallics, Vol. 21, No. 19, 2002
Ta ble 3. Sp ectr oscop ic Da ta of HRh (L)x(CO)4-x Com p lexes (x ) 2, 3, a n d 4)
van der Slot et al.
b
Complex
δ(1H) (ppm)a
δ(31P) (ppm)a
J
HP (Hz)
n.d.
J HRh (Hz)
J RhP (Hz)
νCO (cm-1
)
HRh(3)(CO)2 (7)
DRh(3)(CO)2
HRh(4)(CO)2 (8)
HRh(3)2(CO) (9)
-10.7 (br)
138
n.d.
218
2078, 2026
2067, 2020
2055, 1996
2069
-9.2 (dt)
149
5
6
4
e3
166
-10.5 (br q)c
135 (m)
n.d.e
111 (s)d
DRh(3)2(CO)
HRh(4)2(CO) (10)
2041
2036
-9.6 (dq)c
153 (m)
115 (s)d
129
13
3
n.d.e
HRh(3)2 (11)
-10.9 (dqui)
-10.3 (dqui)
36
36
6
6
203
203
130
a
b
Measured in toluene-d8 or benzene-d6. Measured in cyclohexane. c All phosphorus atoms have similar J HP coupling constants. q )
quartet, qui ) quintet. This chemical shift belongs to the noncoordinated phosphorus atom of (L∩L′). e Not determined.
d
Sch em e 2. Geom etr y of th e Hyd r id e Com p lexes
F or m ed w ith th e Bid en ta te Liga n d s
phosphorus ligand. From these results we conclude that
in the presence of more than 1 equiv of ligand per
rhodium some of the rhodium-hydride complexes HRh-
(L∩L)(CO)2 are converted to hydride complexes of the
form HRh(L∩L)(L∩L′)(CO). One of the bidentate ligands,
denoted L∩L′, coordinates in a monodentate fashion.32
The geometry of this type of rhodium-hydride com-
plexes is depicted in Scheme 2. High-pressure IR studies
showed that the concentrations of HRh(L∩L)(L∩L′)(CO)
and HRh(L∩L)(CO)2 present in solution depend on the
ligand and rhodium concentrations and the carbon
monoxide pressure. HRh(L∩L)(CO)2 and HRh(L∩L)-
(L∩L′)(CO) were always observed together, except when
only 1 equiv of L∩L was used, which led to formation of
the former complex only. Both HRh(3)2(CO) (9) and
HRh(4)2(CO) (10) have relatively large J HP coupling
constants (6 and 13 Hz, respectively), indicating a
distortion of the trigonal bipyramidal structure.13,30 The
complex multiplet observed for the phosphorus reso-
nances of these complexes indicates that the phosphorus
atoms are inequivalent on the NMR time scale.
hydride complex.The spectroscopic data of the observed
hydride complexes with 3 and 4 are presented in Table
3. The hydride resonance of HRh(3)(CO)2 was broad and
remained broad after cooling to -60 °C and the values
of the J RhH and J PH coupling constants remained
unresolved. The NMR data do not lead to detailed
information about the trigonal bipyramidal structure.
The IR spectrum obtained from the NMR sample of
HRh(3)(CO)2 showed two absorption bands for the
carbonyl ligands at 2078 and 2026 cm-1. In view of the
number and position of the carbonyl frequencies, we
conclude that only the ee isomer is present.4 The
hydride resonance of HRh(4)(CO)2 was sharp and all J XH
coupling constants (X ) 103Rh or 31P) were assigned after
selective decoupling of the phosphorus resonance of the
hydride complex. The J RhH and J PH coupling constants
of 4 and 5 Hz, respectively, are indicative of coordination
of both phosphorus atoms in the equatorial plane of the
trigonal bipyramid. The presence of two carbonyl ab-
sorption bands in the IR spectrum obtained from the
NMR sample of HRh(4)(CO)2 indicates that the ee
isomer is the only isomer present. The positions of the
phosphorus resonances and J RhP coupling constants of
HRh(3)(CO)2 and HRh(4)(CO)2 are similar to those
found for the monodentate ligands. As mentioned above,
for both ligands 3 and 4 ligand-to-rhodium ratios higher
than one are required to reach complete conversion to
the rhodium-hydride complexes in the (high-pressure)
IR experiments performed under hydroformylation re-
action conditions ([3] ) ∼1-12 mM). The high-pressure
IR spectra with L/Rh g 1.5 showed a third carbonyl
frequency in the terminal CO region. When the high-
pressure NMR experiments were performed with L/Rh
g 1.5, additional hydride resonances were observed. The
multiplicity of these hydride resonances (double quar-
tets) indicates that three phosphorus atoms are coor-
dinated to the rhodium center. The 31P NMR spectra
showed complex multiplets in the region where coordi-
nated phosphorus atoms resonate together with an
additional singlet, close to the resonance of the free
We synthesized HRh(3)2 (11) by the reaction of Rh-
(acac)(CO)2 and 2 equiv of ligand 3 under hydrogen
pressure. The spectroscopic data of this complex are
depicted in Table 3. Two hydride resonances, double
1
quintets, were observed in the H NMR spectrum, and
two doublets were observed in the 31P NMR spectrum
having identical J RhH, J PH, and J RhP coupling constants,
indicating that two isomers of 11 are present in solution.
The quintet structures indicate fast movement of the
hydride atom with respect to the phosphorus atoms. The
two isomers are probably two diastereomers resulting
from the atropisomerism of the diphenol backbone.
Addition of carbon monoxide (1 bar) to a solution of 11
results in complete conversion to 10. From this we
conclude that under hydroformylation reaction condi-
tions complex 11 is totally absent.
When we used a L/Rh g 2 for ligand 4 a third hydride
resonance (multiplet) at -9.7 ppm was observed in the
1H NMR spectrum. In the 31P NMR spectrum additional
resonances appeared in the region of the coordinated
phosphorus ligand. Selective decoupling of the phos-
phorus resonance at 153 ppm showed a doublet in the
hydride region, having a J RhH coupling constant of 3 Hz.
The IR spectrum did not show any additional terminal
or bridging carbonyl bands. These results suggest that
the structure of the additional hydride complex might
be formulated as HRh(4)2, similar to HRh(3)2 (vide
supra, 11). The ratio between HRh(4)2(CO) (10) and this
(32) Hughes, R. O.; Young, D. A. J . Am. Chem. Soc. 1981, 103, 6636.

Rhodium Hydroformylation of Pyrrolyl Amidite Ligands
Organometallics, Vol. 21, No. 19, 2002 3877
Ta ble 4. Resu lts of th e Hyd r ofor m yla tion of
1-Octen e w ith Liga n d s 1, a n d 2a
Ta ble 5. Resu lts of th e Hyd r ofor m yla tion of
1-Octen e w ith Liga n d s 3 a n d 4a
time conversion
linear 2-octenes
tim conversion
linear 2-octenes
ligand L/Rh (h)
(%)
l/b
TOFb
(%)
(%)
ligand L/Rh (h)
(%)
l/b
TOFb
(%)
(%)
1
5
0.084
35
32
30
3
32
35
3.1 10 × 103
3.7 12 × 103
3.5 4 × 103
2.3 15
61
68
66
50
56
68
18
14
16
27
22
2
3
1.5 0.10
33
20
46
23
24
20
21
107 10 × 103 86
11
7
13
15
5
10 0.074
50 0.26
3
10
50
3
0.054
0.27
0.55
1.5
110 10 × 103 92
98 5 × 103
104 1 × 103
5 424
9 246
6 240
86
85
80
85
83
2
5
5.1
5c 2.8
2.6 n.d.d
4c
50c 2.0
2.3 219
5
2.5
5
3
50
3.0
a
Conditions used: T ) 80 °C, [Rh] ) 0.2 mM, pCO ) pH ) 10
2
b
a
bar, [1-octene] ) 0.64 M in toluene. Initial TOF in mol aldehyde
Conditions used: T ) 80 °C, [Rh] ) 0.2 mM, pCO ) pH ) 10
2
d
b
(mol Rh‚h)-1
.
c [Rh] ) 0.5 mM. The TOF could not be determined
bar, [1-octene] ) 0.64 M in toluene. Initial TOF in mol aldehyde
because of incomplete catalyst formation.
(mol Rh‚h)-1 c 3 h instead of 1 h catalyst formation time was used.
.
ensure complete catalyst formation, we increased the
rhodium and ligand concentrations 2.5-fold ([Rh] ) 0.5
mM). In this instance the catalyst did not show any
activity during the first 2.5 h. After 2.5 h, the reaction
started, indicating that the catalyst formation time of
1 h was not enough to obtain complete conversion of the
rhodium precursor. Again, a high percentage of 2-octenes
was found with use of these higher concentrations.
When we used a catalyst formation time of 3 h instead
of 1 h, the catalyst activity decreased, indicating pre-
mature decomposition of the catalyst. By increasing the
ligand-to-rhodium ratio to 50 and using 1-h catalyst
formation time, the catalyst activity did not change
during the run and 2-octene formation was low, thus
catalyst formation is complete with high ligand and
rhodium concentrations.The hydroformylation reaction
of 1-octene with the bidentate ligands 3 and 4 was also
performed at various ligand-to-rhodium ratios. The
results obtained are depicted in Table 5. A ligand-to-
rhodium ratio of 1.5 was the minimum value for ligand
3 with this low rhodium concentration (0.2 mM). Use
of smaller amounts of ligand resulted in a high percent-
age of 2-octenes and low hydroformylation activity,
indicating incomplete catalyst formation. The catalyst
formed with pyrrolyl-based ligand 3 was highly active
in the hydroformylation reaction and a very high
regioselectivity for the linear aldehyde was observed,
linear/branched ratios typically being on the order of
100. The percentage of 2-octenes formed during the
reaction ((11%) is rather high, which results in a
moderate overall catalyst selectivity ((85%). The cata-
lyst is inactive for the hydroformylaton of 2-octene and
thus the linearity remains very high even at prolonged
reaction times. The combination of both high hydro-
formylation and isomerization activity is reminiscent of
the results observed for bulky (bidentate) phosphite
ligands.13 This similarity can be explained by the
electron-withdrawing character of these phosphorus
amidite ligands. Generally, electron-withdrawing phos-
phorus ligands show an increase of the reaction rate as
a result of the facile carbon monoxide dissociation and
stronger alkene coordination.4,7,36 The linearity observed
with this bidentate phosphorus amidite ligand (3) is
much higher than that observed for bidentate phosphite
ligands having similar backbones.13 A decrease of the
hydroformylation activity together with a constant
selectivity was observed when the concentration of
ligand 3 was increased. The decrease in hydroformyla-
tion activity is probably caused by coordination of
additional hydride complex remained constant (1:1)
upon increasing the ligand concentration and removing
carbon monoxide. Attempts to isolate HRh(4)2 were
unsuccessful.
Hyd r ofor m yla tion of Alk en es. Ligands 1-4 were
applied in the rhodium-catalyzed hydroformylation of
1-octene. The catalysts were prepared in situ from Rh-
(acac)(CO)2, the required amount of ligand, CO, and H2
at 80 °C. As already mentioned in the previous section,
several rhodium-hydride complexes can be formed
under hydroformylation reaction conditions depending
on the ligand and rhodium concentrations and the
carbon monoxide pressure. The ligand and rhodium
concentrations used in catalysis are much lower than
those used in the NMR spectroscopic experiments.
Therefore, the proportion of HRhLx(CO)2 (x ) 2 for L )
1, 2 and x ) 1 for L ) 3, 4) in the reaction mixture used
in catalysis is significantly higher than that used in the
spectroscopic experiments. The hydroformylation reac-
tion has been performed with several ligand-to-rhodium
ratios to investigate the effect of the ligand concentra-
tion on the catalyst activity and selectivity.
The results of the hydroformylation reaction of 1-octene
for systems based on 1 and 2 are presented in Table 4.
The activity and selectivity obtained with ligand 1 are
in the range of those observed with bulky phosphite
ligands.5 At the rhodium concentration used, the ligand-
to-rhodium ratio of 5 is the minimum amount of ligand
needed; the use of lower L/Rh ratios resulted in a large
amount of 2-octene and we observed that the catalyst
activity increased during the reaction. These results
point to incomplete catalyst formation with L/Rh < 5.
Increasing the ligand-to-rhodium ratio from 10 to 50
results in a large decrease of the activity. This negative
dependency of the reaction rate on ligand concentration
can be explained by the formation of less reactive
complexes.29,33-35
Ligand 2 was tested in the hydroformylation reaction
by using the same reaction conditions as ligand 1. With
use of these low ligand and rhodium concentrations, the
catalyst based on 2 was hardly active and a high
percentage of 2-octenes was formed (Table 4). The large
amount of 2-octenes indicates that by using low phos-
phinite ligand concentrations small amounts of Rh-
(acac)(CO)2 are present in solution that can be converted
to HRh(CO)4, which is a good isomerization catalyst. To
(33) Brown, C. K.; Wilkinson, G. J . Chem. Soc. A 1970, 2753.
(34) Evans, D.; Yagupsky, G.; Wilkinson, G. J . Chem. Soc. A 1968,
2660.
(35) Hjortkjaer, J .; Toromanova-Petrova, P. J . Mol. Catal. 1982, 14,
19.
(36) J ongsma, T.; Challa, G.; van Leeuwen, P. W. N. M. J . Orga-
nomet. Chem. 1991, 421, 121.

3878 Organometallics, Vol. 21, No. 19, 2002
van der Slot et al.
irreversible.45 Reversible formation of 2-alkylrhodium
species may result in 1-deuterio-2-alkene (or a 1-deu-
terio-1-alkene) and a rhodium hydride.39 Thus, after a
moderate preference for the linear alkyl in the alkene
insertion step, the backward reaction of the 2-alkyl-
rhodium species to alkenes may lead to an overall
selectivity for linear aldehyde. In these systems the
formation of 2-alkenes is indicative of this type of
kinetics, implying a lower intrinsic preference for 1-alkyl
formation than the l/b ratio might suggest.
Sch em e 3. Deu ter iofor m yla tion In volvin g
Isom er iza tion a n d F a st Exch a n ge of Rh H a n d
Excess D2
The rhodium hydride formed may either react with a
new molecule of 1-alkene, thus giving a monodeuterated
aldehyde, or undergo a fast exchange with the excess
of D2 present in the system, regenerating a rhodium
deuteride. This situation is outlined in Scheme 3.
However, Casey has shown that in the BISBI system
insertion to 2-alkylrhodium is indeed reversible, giving
1-deuterio-1-alkene for 75% at 33 °C and 5-6 bar (the
remainder forms branched, bisdeuterated aldehyde).
The present system contains a ligand that is much more
electron poor and the reaction was studied at higher
temperatures. Lazzaroni39 has found that hydroformyl-
ation at room temperature is characterized by an
irreversible insertion, but he reported that at 100 °C
incorporation of deuterium in 1-hexene was observed.
The deuterioformylation with ligand 3 was performed
under the same conditions as the hydroformylation
experiments of 1-octene (T ) 80 °C, [Rh] ) 0.2 mM,
additional ligand resulting in an increasing amount of
hydride complex 9 as was shown in the spectroscopic
experiments.
When ligand 4 is used in the hydroformylation reac-
tion of 1-octene, at least 3 equiv of ligand and 3 h of
catalyst formation time are required to obtain complete
conversion of the rhodium precursor to the rhodium-
hydride complex. A longer catalyst formation time (5
h) led to a decrease of the activity, indicating some
catalyst decomposition. The use of higher rhodium and
ligand concentrations did not have any effect. The
activity and selectivity observed with ligand 4 were
moderate and comparable to that observed for phos-
phine ligands having similar electron-donating proper-
ties.4,37 Comparison of the hydroformylation results of
ligands 3 and 4 shows that increasing the electron-
withdrawing capacity of the ligand, with only a small
change in steric hindrance, results in improved activity
and selectivity.
The combination of high activity and high regioselec-
tivity obtained for 1-octene with HRh(3)(CO)2 prompted
us to test this catalyst also in the hydroformylation
reaction of styrene. In this experiment the optimal
ligand-to-rhodium ratio is 1.5 (at [Rh] ) 0.2 mM). With
styrene as the substrate at 80 °C, the selectivity for the
linear aldehyde found amounted to 46% (l/b ) 0.84),
which is relatively high for styrene. The turnover
frequency of 2800 mol aldehyde (mol Rh‚h)-1 is high as
well, specially when compared to phosphine-based
systems (Xantphos: 724 m‚m-1‚h-1, l/b ) 0.88).4,13
Deu ter iofor m yla tion w ith Liga n d 3. The results
of the hydroformylation with ligand 3 showed a high
selectivity to nonanal (l/b ratio ∼100) and approximately
10% isomerization of 1-octene to 2-octene. The latter is
formed via reversible insertion and deinsertion of a
2-octylrhodium species. The question arose whether the
formation of linear aldehyde also involved reversible
insertion of alkene into the rhodium hydride bond.
Deuterioformylation experiments can give detailed in-
formation about the reversibility of the first steps in the
hydroformylation reaction.38,39 Casey has shown38 that
the linear aldehyde formed with BISBI under mild
conditions of deuterioformylation contains two deute-
rium labels, one at the aldehyde carbon atom and the
other at the â-carbon atom (Scheme 3), which indicates
that insertion of 1-alkenes to give 1-alkylrhodium is
L/Rh ) 1.5, [1-hexene] ) 0.81 M, pCO ) pD ) 10 bar).
2
1-Hexene was used for the deuterioformylation studies
1
to facilitate comparison with literature data of the H
and 2H NMR spectra.38 The rate of reaction for the
hydroformylation of 1-hexene and the percentage of
2-hexenes formed were similar to those observed for
1-octene. The initial linear-to-branched ratio was very
high for 1-hexene, viz. 161. The side product 2-hexene
has an extremely low activity for hydroformylation
under these conditions and thus the linearity remained
high. The aldehyde production was monitored by gas
chromatography. The number of deuterium atoms in-
corporated into 1-hexene, 2-hexene, and heptanal was
monitored by gas chromatography/mass spectroscopy.
After complete conversion of 1-hexene the deuterium
contents at each position in heptanal were determined
2
by H NMR spectroscopy. The activity and selectivity
for the deuterioformylation of 1-hexene were the same
as those of the hydroformylation of 1-octene. The data
of the deuterium distribution observed with gas chro-
matography/mass spectroscopy have been collected in
Table 6. The table gives the composition of the system,
(38) Casey, P. C.; Petrovich, L. M. J . Am. Chem. Soc. 1995, 117,
6007.
(39) Lazzaroni, R.; Uccello-Barretta, G.; Benetti, M. Organometallics
1989, 8, 2323.
(40) The absorption band of C(O)H should be at 1733 cm-1 (cyclo-
hexane).
(41) van der Veen, L. A.; Kamer, P. C. J .; van Leeuwen, P. W. N.
M. Angew. Chem., Int. Ed. 1999, 38, 336.
(42) Elsevier, C. J . J . Mol. Catal. 1994, 92, 285.
(43) Atherton, M. J .; Fawcett, J .; Hill, A. P.; Holloway, J . H.; Hope,
E. G.; Russell, D. R.; Saunders, G. C.; Stead, R. M. J . J . Chem. Soc.,
Dalton Trans. 1997, 1137.
(44) Keim, W.; Maas, H. J . Organomet. Chem. 1996, 514, 271.
(45) Reversible insertion without alkene-rhodium dissociation would
also lead to formation of D2 aldehydes, but as yet there is no evidence
for an irreversible alkene coordination. In fact this latter would mean
that alkene complexation is rate determining, but not necessarily
regioselectivity determining; kinetic measurements cannot distinguish
between these two possibilities.1a
(37) Hughes, R. O.; Unruh, J . D. J . Mol. Catal. 1981, 14, 71.

Rhodium Hydroformylation of Pyrrolyl Amidite Ligands
Organometallics, Vol. 21, No. 19, 2002 3879
Ta ble 6. Deu ter iu m Distr ibu tion Obser ved in th e Deu ter iofor m yla tion of 1-Hexen ea
sample
(conversion, %)
time
(s)
l/b
ratio
total %
1-hexene
total %
heptanal
no. of deuterium
% 1-hexene
Dn
% 2-hexene
Dn
% heptanal
Dn
atoms, n
1 (10)
2 (21)
3 (30)
85
172
253
90
79
70
9
20
28
0
1
2
3
0
1
2
3
0
1
2
3
4
0
1
2
3
4
0
1
2
3
4
0
1
2
3
4
84
5.3
1
0
73
5
0.5
0.4
0.1
0
0.6
0.7
0.1
0
0.9
0.8
0.1
0
0.2
3.6
5.2
0
1.6
7.3
10.0
1.2
2.6
9.6
13.7
2.1
0
3.7
11.4
15.3
2.6
0
3.9
13.3
19.3
3.8
0
161
132
1
0
62.6
6.0
1.0
0.2
0.2
54.3
9.1
1.4
0.13
0.07
46.7
9.3
1.5
0.6
0
0
4 (35)
5 (42)
326
409
133
128
78
65
58
0
33
40
90
0.9
0.9
0.2
0
0
1.1
1.1
0.22
0
0
6 (100)
7,200
0
0
0
0
1.9
2.4
2.8
2.0
0.9
6.4
24.2
34.8
17.2
7.3
0
a
Conditions used: T ) 80 °C, [Rh] ) 0.2 mM, pCO ) pD ) 10 bar, [1-hexene] ) 0.81 M, in benzene. Conversion is that of 1-hexene.
2
Data are calibrated for a total of 100% for 1-hexenes, 2-hexenes, and heptanal.
Sch em e 4. Deu ter iofor m yla tion In volvin g Isom er iza tion a n d Slow Exch a n ge of Rh H a n d Excess D2
educt, and products at a range of conversion levels,
normalized to mole percent to demonstrate the amount
of incorporated deuterium in each component at the
respective conversion of 1-hexene. The amounts of the
compounds found were corrected for the natural abun-
dance of 13C.
sumed that the intermediary rhodium-hydride reacts
rapidly with the excess of D2 to regenerate a rhodium-
deuteride species (and HD). At low conversion this
would have given rise to D2 aldehydes only. Since this
is not the case, the rate of hydroformylation must be
higher than (or at least approximately equal to) the rate
of H-D exchange between D2 and RhH. To confirm this
surprising outcome, we decided to measure this rate
independently. Indeed, at the initial 1-hexene concen-
tration, hydroformylation (and insertion-deinsertion)
is about 10 times faster than H-D exchange (vide infra,
in separate section) with this catalyst. At 10% conver-
sion of 1-hexene (to heptanal and 2-hexene, sample 1)
we see indeed incorporation of deuterium in hexenes.
The commonly accepted mechanism would suggest that
deuterated 2-hexene would be produced. The amount
of deuterium “lacking” in heptanal (3.5 mol, normalized
to 100) was not recovered as D1-2-hexene, wich con-
tained only 0.7 equiv of deuterium, but it was recovered
in 1-hexene (5.3 mol %). Thus, reversible formation of
1-hexene is the main pathway for the deinsertion
Sample 1, taken after 10% conversion of 1-hexene,
shows that 60% of the heptanal formed contains two
deuterium atoms (5.2 mol % in Table 6) as one would
expect for a deuteriohydroformylation, but 40% contains
only one deuterium atom (3.5%). Monodeuterated alde-
hyde is formed, we propose, by a cycle that starts with
rhodium hydride, because in the initial period of the
experiment the final step involving D2 will always
deliver one deuterium atom to heptanal. Rhodium
hydride is formed via reversible insertion-deinsertion
of hexenes as is evidenced by the deuterium-enriched
hexenes that were observed. The hydroformylation
cycle, started by rhodium hydride and completed by
reaction of the acylrhodium species with D2, thus leads
to monodeuterated product. Usually, it is tacitly as-

3880 Organometallics, Vol. 21, No. 19, 2002
van der Slot et al.
(Scheme 4). Interestingly, for BISBI the observation was
the same qualitatively, but little attention was paid to
this, because the absolute numbers were so much
smaller.38
Also in samples 2-5 it is seen that the proportion of
D1-heptanal is much higher than that of D1-2-hexene
and that more deuterated 1-hexene is formed. Mono-
deuterated 1-hexene can form from reversible insertion
via 1-hexylrhodium and reversible insertion via 2-hexyl-
rhodium. Reversible insertion via 1-hexylrhodium, which
is peculiar in that usually 1-alkylrhodium formation is
regarded as irreversible, leads eventually to an extra
deuterium in the â-position of heptanal formed. Revers-
ible insertion of 1-hexene via 2-hexylrhodium leads,
after hydroformylation, eventually to heptanal contain-
ing deuterium in the R-position. Reversible insertion of
1-hexene via 2-hexylrhodium is also peculiar, since
usually the intermediate 2-hexylrhodium is thought to
lead to 2-hexene, the isomerization product.1a Since we
have not established the position of deuterium in D1-
1-hexene, at this point we can only conclude that a high
proportion of deuterated 1-hexene is formed via revers-
ible insertion, either route being unexpected under these
conditions. Scheme 4 outlines the routes. At least 40%
of the insertion is reversible. Note that rhodium hydride
will also participate in nonproductive reversible inser-
tion-deinsertion reactions.
F igu r e 1. Difference IR spectrum obtained after addition
of H2 to a solution of Rh(3)(CO)2 and Rh(3)2(CO) under
carbon monoxide pressure at 80 °C.
1-hexene before forming 2-hexene or 2-hexene does
insert into the RhH and RhD bonds but never manages
to carry out a migratory insertion of carbon monoxide.
From sample 6 heptanal was separated from the
reaction mixture by distillation. The deuterium content
at the aldehyde carbon, R- and â-positions, was deter-
2
2
mined with H NMR spectroscopy. The H NMR spec-
trum of the linear aldehyde showed deuterium at the
R-carbon atoms (12-15%), â-carbon atoms (90-92%),
and the aldehyde carbon (95-97%). The 1H NMR
spectrum of sample 6 showed 3-5% heptanal having a
proton at the aldehyde carbon atom. D3-Heptanal is
formed via hydroformylation of deuterium-enriched
1-hexene. D2-Heptanal is formed both by hydroformyl-
ation of D0-hexene via the rhodium-deuteride complex
(Scheme 4; for the sake simplicity not shown) and
hydroformylation of D1-hexene via the rhodium-hy-
dride complex.
At higher conversions D3-heptanal is formed from D1-
1-hexenes. The ratios D1:D2:D3 for the increments of
heptanal formed from the isotopic mixture of hexenes
present closely follow the preference of 40:60 as found
in sample 1, expressing the ratio of the rates and
concentrations of rhodium-hydride and rhodium-deu-
teride.
The proportion of D0-heptanal in samples 2-6 is
somewhat high in view of the small amount of HD that
has been formed, which requires further discussion.
After 2 h complete conversion was reached and sample
6 was analyzed. From the total deuterium balance we
see that, on average, two deuterium atoms have been
built into heptanal, as is to be expected more or less
since the exchange of rhodium-hydride and D2 is slow!
The residual 2-hexene also contains on average two
deuterium atoms, indicating that hexene has partici-
pated in more than one reaction sequence. Since we
started with approximately 70 mmol of D2 and 16 mmol
of 1-hexene, the resulting 2-hexene contains 3.2 mmol
of deuterium and from this we calculate that the final
ratio of HD:D2 is roughly 1:20. At high conversions of
1-hexene the rate of hydroformylation drops and the
rate of the RhH/D2 exchange becomes more important.
A large isotope effect, as we have observed before in the
deuterioformylation of cyclohexene,36 may cause a much
faster reaction for HD than D2, which leads to a high
proportion of D0-heptanal.
In the IR spectrum we only observed the C(O)D
absorption band (1722 cm-1 in pentane) and not the
absorption of C(O)H.36,40 This shows that D1-heptanal
contains its deuterium atom predominantly (> 95%) at
the aldehyde carbon atom. It is formed from nondeu-
terated 1-hexene and rhodium-hydride. In the final
product we observe 18% of D3-heptanal and 7.8% of D4-
heptanal; as a relative percentage of heptanals this
amounts to 27.4% of the heptanals formed. The deute-
rium content at the R-carbon is 12-15%, which stems
from reversible formation of 2-hexylrhodium from rhod-
ium-deuteride and 1-hexene. This accounts for only half
of the heptanal containing more than two deuterium
atoms. Thus, the remainder of the extra deuterium
atoms reside on the â-carbon atoms. This can only form
via reversible 1-hexylrhodium formation from 1-hexene
and rhodium-deuteride. Thus, the linear alkyl forma-
tion must also be partly reversible for this catalyst
system.
Exch a n ge betw een Rh D a n d H2. We studied the
H/D exchange rate in the absence of substrate using IR
spectroscopy. The IR absorptions of the carbonyl ligands
of a rhodium-deuteride and a rhodium-hydride com-
plex differ significantly for rhodium complexes contain-
ing two phosphorus ligands in the equatorial plane
(Table 3). For the investigation of the H/D exchange
rate, we synthesized DRh(3)(CO)2 in situ from Rh(acac)-
Hydroformylation of 2-hexene has not been taken into
account, as it was shown in separate experiments that
the rate of hydroformylation of internal alkenes (2-
octene) is several orders of magnitude lower for this
catalyst. Thus in the present experiment the formation
of 2-hexene should be considered as irreversible as far
as heptanal formation is concerned. The enrichment
observed for the resulting 2-hexene means that either
1-hexene goes back and forth a couple of times enriching

Rhodium Hydroformylation of Pyrrolyl Amidite Ligands
Organometallics, Vol. 21, No. 19, 2002 3881
the carbon monoxide insertion is controlling the reaction
rate under the conditions studied and thereby inducing
the high linear-to-branched ratio observed for this
catalyst.
Hyd r ofor m yla tion Rea ction of In ter n a l Alk en es.
The hydroformylation catalyst formed with ligand 3
showed high isomerization rates together with a high
preference for the linear aldehyde. The deuterioformyl-
ation experiments showed that the rate of the carbon
monoxide insertion for branched rhodium-alkyl com-
plex is slow compared to that of the linear alkyl complex.
This combination of catalyst properties might be ideal
for the hydroformylation of internal alkenes to linear
aldehydes. The reaction conditions for the hydroformyl-
ation of internal alkenes are different from the condi-
tions used for 1-alkenes. Often high temperatures are
used to enhance the isomerization reaction, together
with low syn-gas pressures to facilitate the carbon
monoxide dissociation and alkene coordination.41 The
hydroformylation of 2-octene was performed at 120 °C,
[Rh] ) 1.0 mM, Rh/L ) 1/1.5, [2-octene] ) 0.64 M, pCO
F igu r e 2. Kinetic data for the H/D exchange. (A) Expo-
nential decay of νCO at 2020 cm-1 vs time. (B) Logarithmic
plot of the decay of the relative absorption at 2020 cm-1 vs
time.
(CO)2 and 1.5 equiv of ligand 3 under 20 bar of CO/D2
(1/1). Besides DRh(3)(CO)2, a small amount of DRh(3)2-
(CO) (9) was formed in the reaction mixture. After
complete conversion to the deuteride complex, the D2
gas was removed by flushing several times with 5 bar
of carbon monoxide. The H/D exchange was initiated
by adding 10 bar of hydrogen to DRh(3)(CO)2 under 10
bar of carbon monoxide. The exchange process was
monitored by using rapid scan high-pressure IR spec-
troscopy at 80 °C (1.3 spectra‚s-1). The difference IR
spectrum presented in Figure 1 shows that the rhodium-
deuteride complexes DRh(3)(CO)2 and DRh(3)2(CO)
) pH ) 2.5 bar. After 15 h 3% conversion of 2-octene to
2
aldehydes was observed, showing a linear-to-branched
ratio of only 1. The rate of hydroformylation remained
very low. Thus, the selectivities are comparable with
those of J ackstell et al.,5g using monodentate pyrrolyl
phosphorus ligands, but the rates are lower.
(negative peaks at νCO ) 2067, 2041, and 2020 cm-1
)
are quantitatively converted into the rhodium-hydride
complex HRh(3)(CO)2 (positive peaks at νCO ) 2078 and
2026 cm-1). The presence of small amounts of HRh(3)2-
(CO) (9) could not be determined because of overlap with
one of the absorptions of the deuteride complex.
Con clu sion s
Introduction of the electron-withdrawing pyrrolyl
substituents resulted in a very active catalyst for the
hydroformylation of 1-octene. Especially the bidentate
pyrrolyl-containing ligand formed a catalyst that showed
high activity together with a high regioselectivity for
the linear aldehyde and moderate amount of 2-octenes.
Investigation of the deuterioformylation reaction mech-
anism with use of the pyrrolyl-based bidentate ligand
3 showed that the hydride migration is a reversible step
under the conditions studied for both linear and branched
alkylrhodium species. Branched rhodium-alkyl com-
plexes prefer exclusively â-hydride elimination to carbon
monoxide insertion resulting in high l/b ratios, and
formation of 2-alkenes and, surprisingly, regeneration
of deuterated 1-alkenes. All linear rhodium-alkyl spe-
cies eventually undergo carbon monoxide insertion
leading to highly linear aldehydes.
The exponential decay of the strongest carbonyl
absorption is presented in Figure 2a (νCO ) 2020 cm-1).
The maximum absorbance of this frequency was taken
as 100%. The natural logarithm of the relative absorp-
tion of the carbonyl frequency at 2020 cm-1 versus time
is presented in Figure 2b. The decay of the carbonyl
frequency of the deuteride complex shows first-order
kinetics in the rhodium concentration and the pressure
of H2. The calculated slope of the line through the data
depicted in Figure 2b is the negative of the H/D
exchange rate. On the basis of these results, a H/D
exchange rate of 1140 h-1 was calculated. The r2
coefficient of the least-squares fit of the line is 0.999.
Although the H/D exchange reaction is very fast indeed,
the initial rate of hydroformylation of 10 × 103 mol
aldehyde (mol Rh‚h)-1 is still an order of magnitude
higher. Therefore, we conclude that up to 70% conver-
sion H/D exchange is not significant and rhodium-
hydride will react with the next moleule of alkene before
rhodium-deuteride is formed. At high conversion the
rate of hydroformylation is lower, because of the first-
order dependence in the alkene concentration, and
therefore some H/D exchange will occur and small
amounts of heptanal having a proton at the aldehyde
carbon will be formed.
Summarizing the results observed in the deuterio-
formylation experiments, we conclude that the linear
and branched rhodium-alkyl complexes undergo â-hy-
dride elimination and that the hydride migration is in
part a reversible step under the conditions studied. On
the basis of the large amount of deuterium in 1-hexene
originating from both linear and branched alkyl com-
plexes, we conclude that carbon monoxide insertion is
slower than alkene coordination and dissociation, hy-
dride migration, and â-hydride elimination. Probably
Exp er im en ta l Section
Gen er a l In for m a tion . All preparations were carried out
under an atmosphere of argon with standard Schlenk tech-
niques. Solvents were distilled from sodium/benzophenone. All
glassware was dried by heating under vacuum. Column
chromatography was performed with silica gel 60 230-400
mesh obtained from Merck or activated neutral alumina 50-
200 µm obtained from Acros Organics. All chemicals were
azeotropically dried before use. The alkene was filtered over
neutral alumina to remove peroxides. The routine NMR
spectra were recorded on a Varian Mercury-VX (300 MHz)
spectrometer (1H, 31P, 13C). The high-pressure NMR experi-
ments were recorded on a Bruker DRX (300 MHz) spectrom-
eter. Chemical shifts are given in ppm referenced to TMS or
H3PO4 (external). IR spectra were recorded on a Nicolet 510
FT-IR spectrometer.
High-pressure NMR experiments were performed in a 10
mm (o.d.) sapphire NMR tube described by Elsevier et al.42 In

3882 Organometallics, Vol. 21, No. 19, 2002
van der Slot et al.
a typical experiment 5 mg (0.019 mmol) of Rh(acac)(CO)2 and
1, 1.5, 2, or 5 equiv of ligand were dissolved in 1.5 mL of
toluene-d8 or benzene-d6. The solution was brought into the
argon-flushed tube. The tube was flushed with syn-gas (CO/
H2 ) 1/1) and put under a pressure of 20 bar. The tube was
heated in the NMR machine to 80 °C and the complex
formation was monitored in time.
The suspension was diluted with 100 mL of THF and the salts
were filtered off. The product was obtained as a colorless oil
after vacuum distillation (bp 88 °C, 2.7 mmHg). Yield 51%;
1H NMR (CDCl3) δ 6.5 ppm (m, 4 H, CHCHNP), 7.2 ppm (m,
4 H, CHCHNP); 31P{1H} NMR (CDCl3) δ 104.8 ppm; 13C NMR
2
(CDCl3) δ 122.9 ppm (d, J PC ) 17 Hz, CHCHNP), 114.0 ppm
3
(d, J PC ) 5 Hz, CHCHNP).
High-pressure IR experiments were performed in a 50-mL
autoclave (SS 316) equipped with IRTRAN windows (ZnS,
transparent up to 700 cm-1, i.d. 10 mm, optical path length )
0.4 mm), a mechanical stirrer, a temperature controller, and
a pressure transducer. In a typical experiment 5 mg of Rh-
(acac)(CO)2 and 1.5, 3, 5, or 10 equiv of ligand were dissolved
in 15 mL of cyclohexane under argon. The solution was brought
into the autoclave and the autoclave was pressurized with 15
bar of CO/H2 (1/1). The autoclave was placed in the infrared
spectrometer and heated to 80 °C. IR spectra were recorded
while the sample was stirred.
P h en yld ip yr r olylp h osp h or u sa m id ite (1). A solution of
1.58 g (17 mmol) of phenol dissolved in 10 mL of toluene was
added dropwise to a solution of 3.34 g (17 mmol) of chlorodipyro-
llylphosphine and 10 mL of triethylamine (excess) dissolved
in 50 mL of toluene at room temperature. The triethylamine‚
HCl salt was filtered off after 1 h of stirring at room
temperature. The solvent was removed in a vacuum. The
product was dissolved in toluene and filtered over neutral
alumina and crystallized from pentane at -30 °C. At room
temperature the product was obtained as a colorless oil. Yield
1
3.49 g (80%); H NMR (CDCl3) δ 6.5 (m, 4 H, CHCHNP), 7.0
(m, 3 H, o, p-PhOP), 7.2 (m, 4 H, CHCHNP), 7.4 ppm (m, 2 H,
m-PhOP); 31P{1H} NMR (CDCl3) δ 108.4 ppm; 13C NMR
(CDCl3) δ 113 (d, 3J PC ) 5 Hz, CHCHNP), 120 (d, 2J PC ) 9 Hz,
Hydroformylation experiments were performed in a stain-
less steel (SS 316) autoclave (196 mL). The autoclave was
stirred mechanically and equipped with a substrate reservoir,
a pressure transducer, a thermocouple, and a sampling device.
As shown before5d no mass transfer limitations occur. In a
typical experiment Rh(acac)(CO)2 and the ligand were dis-
solved in 15 mL of toluene and brought into the autoclave.
After flushing the autoclave with CO/H2 (1/1), the autoclave
was put under a pressure of 15 bar. The autoclave was heated
at 80 °C, and after 1 h the substrate solution (2 mL of 1-octene,
1 mL of decane, and 2 mL of toluene) was charged into the
reservoir and added to the reaction mixture by overpressure.
The alkene was filtered over neutral alumina to remove
peroxides. During the reaction several samples were taken at
approximately 0.5 bar pressure drops corresponding with
∼15% increments of conversion and immediately quenched by
adding an excess of P(O-n-Bu)3, to deactivate hydroformylation
or isomerization active rhodium species. The total pressure
drop never exceeded 2 bar. The samples were analyzed by GC
with decane as internal standard.
3
CHCHNP), 122.0 (d, J PC ) 5 Hz, o-PhOP), 125 (s, p-PhOP),
2
131 (s, m-PhOP), 154 ppm (d, J PC ) 11 Hz, ipso-PhP); FAB/
MS m/e 256.08, Anal. Calcd for C14H13N2OP: C, 65.62; H, 5.11;
N, 10.93. Found: C, 65.27; H, 4.94; N, 10.90.
P h en yl Dip h en ylp h osp h in ite (2). A solution of 1.41 g (15
mmol) of phenol dissolved in 20 mL of toluene was added
dropwise to a solution of 2.69 mL (15 mmol) of chlorodiphen-
ylphosphine and 10 mL of triethylamine (excess) dissolved in
100 mL of toluene at room temperature. The triethylamine‚
HCl salt was filtered off after 1 h of stirring at room
temperature. The solvent was removed in a vacuum. The
1
product was obtained as colorless oil. Yield 3.62 g (87%); H
NMR (CDCl3) δ 7.2 (m, 1 H, p-PhOP), 7.3 (m, 2 H, o-PhOP),
7.4 (m, 2 H, m-PhOP), 7.5 (m, 6 H, o, p-PhP), 7.8 ppm (m, 4
H, m-PhP); 31P{1H} NMR (CDCl3) δ 111.8 ppm; 13C NMR
2
(CDCl3) δ 119 (d, J PC ) 11 Hz, o-PhP), 123 (s, p-PhOP), 129
3
(d, J PC ) 6 Hz, m-PhP), 130.8 (s, p-PhP), 131.1 (s, m-PhP),
1
2
141 (d, J PC ) 17 Hz, ipso-PhP), 158 ppm (d, J PC ) 10 Hz,
ipso-PhOP); FAB/MS m/e 278.09.
The concentrations used in the deuterioformylation experi-
ment were similar to those used in the hydroformylation
experiments. Instead of toluene, benzene was used as solvent
to facilitate the separation of aldehyde and solvent. The
autoclave was charged with 1 mg of Rh(acac)(CO)2 and 3 mg
of 3 in 15 mL of benzene and pressurized with 15 bar of CO/
D2 (1/2). The autoclave was heated to 80 °C and after 1 h 2
mL (16 mmol) of 1-hexene in 2 mL of benzene and 1 mL of
decane were charged into the reservoir and added to the
reaction mixture by overpressure of carbon monoxide. The
autoclave was pressurized up to 20 bar with carbon monoxide.
During the reaction several samples were taken and quenched
immediately by adding an excess of P(O-n-Bu)3, to deactivate
hydroformylation- or isomerization-active rhodium species.
The samples were analyzed by GC. The deuterium contents
in the substrate and products during the reaction were
determined by using gas chromatography/mass spectrometry.
After 2 h, complete conversion was reached and the aldehydes
1,1′-Bip h en yl-2,2′-d iyl-b is(d ip yr r olylp h osp h or a m id -
ite) (3). A solution of 2.8 g (15 mmol) of 2,2′-dihydroxy-1,1′-
biphenyl in 10 mL of THF was added dropwise to a solution
of 6.0 g (30 mmol) of chlorodipyrrolylphosphine and 10 mL of
triethylamine (excess) in 50 mL of THF at room temperature.
The triethylamine‚HCl salts were filtered off after 1 h of
stirring at room temperature and the solvent was removed
under vacuum. The white solid was dissolved in toluene and
filtered over neutral alumina (Rf ) 1). The product precipitated
1
when the toluene was concentrated. Yield 68%; mp 95 °C, H
NMR (CDCl3) δ 6.2 (m, 8 H, CHCHNP), 6.7 (m, 8 H,
3
3
CHCHNP), 6.9 (d, 2H, J HH ) 7 Hz, o-Ph), 7.2 (d, 2H, J HH
)
7 Hz, m-Ph-Ph), 7.3 ppm (m, 4H, m,p-Ph); 31P{1H} NMR
(CDCl3) δ 109.4 ppm; 13C NMR (CDCl3) δ 112.5 (d, J PC ) 4
2
3
3
Hz, CHCHNP), 119.5 (d, J PC ) 12 Hz, o-Ph), 121.6 (d, J PC
5 Hz, CHCHNP), 124.7 (s, p-Ph), 129.7 and 132.0 (s, m-Ph),
130.2 (d, J PC ) 4 Hz, o-Ph-Ph), 151.1 (d, J PC ) 10 Hz, ipso-
Ph); FAB/MS m/e 511. Anal. Calcd for C28H24N4O2P2: C, 65.88;
H, 4.74; N, 10.98. Found: C, 65.91; H, 4.66; N, 10.85.
)
2
were distilled. 1H and H NMR spectroscopic data of the linear
3
2
aldehydes were identical with those obtained by Casey et al.38
Chlorodipyrrolylphosphine and the monodentate and bi-
dentate phosphinite ligands 2 and 4 were synthesized accord-
ing to literature procedures.16,43,44 The syntheses are given
below, because no detailed information about the synthesis,
purification, and characterization of these compounds was
given in the literature. The trans-RhCl(CO)L2 complexes were
prepared according literature procedures.26
1,1′-biph en yl-2,2′-diyloxy-bis(diph en ylph osph in ite) (4).
A solution of 1.8 g (10 mmol) of 2,2′-dihydroxy-1,1′-biphenyl
in 7 mL of THF was added dropwise to a solution of 4.4 g (20
mmol) of chlorodiphenylphosphine and 7 mL of triethylamine
(excess) in 35 mL of THF at room temperature. The triethyl-
amine‚HCl salts were filtered off after 1 h of stirring at room
temperature and the solvent was removed under vacuum. The
white solid was dissolved in toluene and washed with water.
Ch lor od ip yr r olylp h osp h in e. A solution of 13.9 mL (0.2
mol) of pyrrole and 100 mL of triethylamine (excess) in 50 mL
of THF was added dropwise to a solution of 8.7 mL (0.1 mol)
of phosphorus trichloride in 200 mL of THF at 0 °C.
Triethylamine‚HCl precipitated directly upon addition. The
reaction mixture was stirred overnight at room temperature.
1
Yield 4.0 g (73%); mp 77 °C; H NMR (CDCl3) δ 7.3-7.1 ppm
(all aromatic H); 31P{1H} NMR (CDCl3) δ 111.9 ppm; 13C NMR
2
(CDCl3) δ 118 (d, J PC ) 16 Hz, o-PhP), 123 (s, p-PhOP), 128
3
(d, J PC ) 7 Hz, m-PhP), 129 (s, m-PhOP), 129.5 (s, p-PhP),

Rhodium Hydroformylation of Pyrrolyl Amidite Ligands
Organometallics, Vol. 21, No. 19, 2002 3883
3
3
130 (d, J PC ) 22 Hz, o-PhOP), 131 (d, J PC ) 4 Hz, o-PhOP),
132 (s, m-PhOP), 141 (d, 1J PC ) 17 Hz, ipso-PhP), 155 ppm (d,
2J PC ) 7 Hz, ipso-PhOP); FAB/MS m/e 554.16.
tr a n s-Rh Cl(CO)(2)2 (6). A solution of 0.1 g (0.4 mmol) of
ligand 2 dissolved in 3 mL of dichloromethane was added
dropwise to a solution of 38 mg (0.1 mmol) of [RhCl(CO)2]2
dissolved in 5 mL of dichloromethane. Carbon monoxide
evolved immediately upon addition of the ligand. The yellow
solution was stirred for 1 h at room temperature. The solvent
was reduced to 1 mL. The product precipitated as a yellow
oil. Yield 112 mg (78%); 1H NMR (CDCl3) δ 7.1 (m, 2 H,
p-PhOP), 7.2 (m, 8 H, o,m-PhOP), 7.4 (m, 12 H, o,p-PhP), 7.8
ppm (m, 8 H, m-PhP); 31P{1H} NMR (CDCl3) δ 123 ppm (J RhP
) 143 Hz); 13C NMR (CDCl3) δ 121 (s, o-PhP), 124 (s, p-PhP),
tr a n s-Rh Cl(CO)(1)2 (5). A solution of 0.1 g (0.4 mmol) of
ligand 1 dissolved in 3 mL of dichloromethane was added
dropwise to a solution of 38 mg (0.1 mmol) of [RhCl(CO)2]2
dissolved in 5 mL of dichloromethane. Carbon monoxide
evolved immediately upon addition of the ligand. The yellow
solution was stirred for 1 h at room temperature. The solvent
was reduced to 1 mL. The product precipitated immediately
when 10 mL of pentane was added. Yield 50 mg (37%); mp
133 °C dec; 1H NMR (CDCl3) δ 6.4 (m, 8 H, CHCHNP), 7.0
(m, 6 H, o, p-PhOP), 7.1 (m, 8 H, CHCHNP), 7.3 ppm (m, 4 H,
m-PhOP); 31P{1H} NMR (CDCl3) δ 111.4 ppm (J RhP ) 190 Hz);
13C NMR (CDCl3) 13C NMR (CDCl3) δ 113 (CHCHNP), 122
(CHCHNP), 124 (o-PhOP), 126 (s, p-PhOP), 131 (s, m-PhOP),
128-133 (m,p-PhP, o,m-PhOP), 136 (dt, J CP ) 26 Hz, J RhC
)
3 Hz, ipso-PhP) 155 ppm (s, ipso- PhOP), the carbonyl carbon
atom was not observed; IR (CH2Cl2) νRh(CO) ) 1991 cm-1; FAB/
MS m/e [RhCl(2)2 - CO] 694.1.
Ack n ow led gm en t. We are indebted to Dr. J . Nijhoff
for useful and stimulating discussions.
152 (ipso-PhOP), 180 ppm (m, Rh(CO)); IR (CH2Cl2) νRh(CO)
)
2018 cm-1. Anal. Calcd for C29H26ClN4O3P2Rh: C, 51.31; H,
3.86; N, 8.25. Found: C, 50.82; H, 4.21; N, 8.16.
OM010760Y