Mechanistic studies of the hydroformylation of 1-alkenes using a monodentate phosphorus diamide ligand

Article abstract of DOI:10.1021/om000785h

The mechanism of the rhodium-catalyzed hydroformylation reaction using a monodentate phosphorus diamide ligand has been investigated. The system presents an ideal case to illustrate the basics of hydroformylation. A detailed kinetic study and (in situ) spectroscopic techniques revealed that several of the elementary reaction steps are involved in the hydroformylation rate control. Which step is rate-determining depends strongly on the conditions used. Deuterioformylation showed that alkene coordination followed by hydride migration is irreversible under the conditions studied. The rhodium hydride complex HRhL₂(CO)₂ and several rhodium-acyl complexes were observed during the hydroformylation reaction. The structures of the rhodium-acyl complexes have been characterized using ³¹P, ¹³C, and ^103-Rh NMR spectroscopy. The major rhodium-acyl complex formed, RC(O)RhL₂(CO)₂, has a trigonal-bipyramidal structure with the two phosphorus ligands coordinated in the equatorial plane. The exchange rates of the equatorial and apical carbonyl ligands with dissolved carbon monoxide differ significantly, the equatorial carbon monoxide being much more labile.

Full text of DOI:10.1021/om000785h

430
Organometallics 2001, 20, 430-441
Mech a n istic Stu d ies of th e Hyd r ofor m yla tion of
1-Alk en es Usin g a Mon od en ta te P h osp h or u s Dia m id e
Liga n d
Saskia C. van der Slot,† Paul C. J . Kamer,† Piet W. N. M. van Leeuwen,*,†
J onathan A. Iggo,‡ and Brian T. Heaton‡
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 166,
1018 WV Amsterdam, The Netherlands, and Department of Chemistry, Donnan and
Robert Robinson Laboratories, University of Liverpool, P.O. Box 147, Liverpool, U.K. L69 7ZD
Received September 13, 2000
The mechanism of the rhodium-catalyzed hydroformylation reaction using a monodentate
phosphorus diamide ligand has been investigated. The system presents an ideal case to
illustrate the basics of hydroformylation. A detailed kinetic study and (in situ) spectroscopic
techniques revealed that several of the elementary reaction steps are involved in the hydro-
formylation rate control. Which step is rate-determining depends strongly on the conditions
used. Deuterioformylation showed that alkene coordination followed by hydride migration
is irreversible under the conditions studied. The rhodium hydride complex HRhL2(CO)2 and
several rhodium-acyl complexes were observed during the hydroformylation reaction. The
103
structures of the rhodium-acyl complexes have been characterized using 31P, 13C, and
-
Rh NMR spectroscopy. The major rhodium-acyl complex formed, RC(O)RhL2(CO)2, has a
trigonal-bipyramidal structure with the two phosphorus ligands coordinated in the equatorial
plane. The exchange rates of the equatorial and apical carbonyl ligands with dissolved carbon
monoxide differ significantly, the equatorial carbon monoxide being much more labile.
In tr od u ction
full characterization has not been achieved. HP NMR
spectroscopy is a powerful technique for the character-
ization of complexes formed under high pressures.
However, in situ characterization of complexes present
during catalysis cannot be performed using an HP NMR
tube because of diffusion limitation of the reacting gases.
A continuous supply of gas and optimal mixing is
achieved in an HP NMR flow cell,12 and therefore, this
is an elegant tool for the characterization of hydro-
formylation reaction intermediates. The combination of
(in situ) HP IR and HP NMR spectroscopy can lead to
complete structural analysis of the intermediates present
during catalysis.
The generally accepted, dissociative mechanism pro-
posed by Wilkinson13 is presented in Scheme 1. The five-
coordinate rhodium bisphosphine complexes shown in
this scheme can be resting states of the catalyst, but
rhodium complexes with one or three phosphine ligands
may also be involved.1,6,8 The reactive species in this
mechanism are coordinatively unsaturated complexes
formed by dissociation of CO or phosphine. A free
coordination site for alkene coordination is obtained in
the CO dissociation step of the rhodium hydride complex
2a (steps 1 and 2). After hydride migration (step 3) and
coordination of CO, insertion of CO occurs to give a
rhodium-acyl complex (2g, steps 4 and 5). The unsat-
urated rhodium-acyl complex 2g undergoes either
hydrogenolysis (step 6) or CO coordination (step 7).
Hydrogenolysis completes the catalytic cycle with the
Rhodium-catalyzed hydroformylation has been exten-
sively studied.1-8 An important aspect of these studies
has been the unraveling of the reaction mechanism in
order to obtain a better understanding of the outcome
of the overall reaction. Although the reaction kinetics
have been investigated many times, very little has been
reported on the characterization of the intermediates
present during the reaction. Nowadays high-pressure
(HP) spectroscopic techniques are regularly applied to
identify organometallic compounds present under high
pressures.3,8-11 HP IR is most often applied to identify
in situ hydroformylation intermediates, but only infor-
mation about carbonyl ligands has been obtained and
* To whom correspondence should be addressed. E-mail: pwnm@
anorg.chem.uva.nl.
† University of Amsterdam.
‡ University of Liverpool.
(1) Osborn, J . A.; Wilkinson, G.; Young, J . F. Chem. Commun. 1965,
17.
(2) Brown, J . M.; Kent, A. G. J . Chem. Soc., Perkin Trans. 2 1987,
1597.
(3) Moser, W. R.; Papile, C. J .; Brannon, D. A.; Duwell, R. A. J . Mol.
Catal. 1987, 41, 271.
(4) Garland, M.; Bor, G. Inorg. Chem. 1989, 28, 410.
(5) Moasser, B.; Gladfelter, W. L. Organometallics 1995, 14, 3832.
(6) Schmid, R.; Herrmann, W. A.; Frenkling, G. Organometallics
1997, 16, 701.
(7) Matsubara, T.; Koga, N.; Ding, Y.; Musaev, D. G.; Morokuma,
K. Organometallics 1997, 16, 1065.
(8) Die´guez, M.; Claver, C.; Masdeu-Bulto´, A. M.; Ruiz, A.; van
Leeuwen, P. W. N. M.; Schoemaker, G. C. Organometallics 1999, 18,
2107.
(9) Garland, M.; Pino, P. Organometallics 1991, 10, 1693.
(10) Feng, J .; Garland, M. Organometallics 1999, 18, 417.
(11) Castellanos-Pa´ez, A.; Castillo´n, S.; Claver, C.; van Leeuwen,
P. W. N. M.; de Lange, W. G. J . Organometallics 1998, 17, 2543.
(12) Iggo, J . A.; Shirley, D.; Tong, N. C. New J . Chem. 1998, 1043.
(13) Evans, D.; Osborn, J . A.; Wilkinson, G. J . Chem. Soc. A 1968,
3133.
10.1021/om000785h CCC: $20.00 © 2001 American Chemical Society
Publication on Web 01/11/2001

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 431
Sch em e 1. Hyd r ofor m yla tion Rea ction Cycle
regeneration of the coordinatively unsaturated hydride
complex 2b and the production of either the linear or
the branched aldehyde (step 6). Depending on the
electronic and steric properties of the ligands used,
different steps will control the reaction rate and differ-
ent intermediates will be present in the reaction mixture
as the catalyst resting state.
lation rhodium catalysts. Cavalieri d’Oro and co-work-
ers21 investigated the reaction kinetics of a triphe-
nylphosphine-modified catalyst, which were different
from the kinetics of catalysts containing bulky phos-
phite ligands. This study showed that the rate equa-
tion for this system depends on the substrate and
phosphine concentrations. The reaction showed a zero-
order dependency on both CO and H2 concentration. The
selectivity for the linear aldehyde was found to depend
dramatically on the ratio between CO and ligand
concentrations. Wilkinson22 and Brown2 and co-workers
examined the coordination chemistry of several rhodi-
um-phosphine complexes that are potential intermedi-
ates in the hydroformylation reaction using IR and
NMR spectroscopy. Under the conditions studied, HRh-
(PPh3)2(CO)2 was present as the most abundant spe-
cies during the hydroformylation reaction, as detected
by in situ IR and NMR techniques.23,24 Moser investi-
gated the hydroformylation reaction intermediates
formed from a range of para-substituted triphenylphos-
phine ligands.3 From their data they concluded that
when the electron density on rhodium is decreased, the
rate-limiting step shifts from “just after the formation
of HRhL2(CO)2 (L ) PPh3) to just after formation of
RhR(CO)2L2 (L ) (p-ClPh)3P or L ) (p-CF3Ph)3P)”.
Recently we reported on a new group of phosphorus
diamides as ligands for the hydroformylation reaction
Many detailed studies on the reaction kinetics and
the resting state of an unmodified rhodium carbonyl
catalyst using various substrates have been per-
formed.9,10,14,15 These studies showed that the rate-
determining step of the hydroformylation reaction for
this electron-poor rhodium catalyst is hydrogenolysis.
Garland and co-workers9,16 identified a rhodium-acyl
intermediate as the resting state of the catalyst using
in situ IR spectroscopy. van Leeuwen et al.17,18 inves-
tigated hydroformylation catalysts based on bulky
monodentate phosphite ligands. Kinetic studies of the
hydroformylation reaction catalyzed by this electron-
poor rhodium catalyst revealed a zero order in alkene
concentration and an approximate first-order rate de-
pendency on the hydrogen concentration, indicating that
the rate-determining step is the hydrogenolysis.19 In situ
IR studies confirmed this hypothesis by showing that
the rhodium-acyl complex is present as the resting
state of the catalyst during the hydroformylation reac-
tion under these conditions.20
Despite their importance, very few kinetic studies
have been performed on ligand modified hydroformy-
(19) van Rooy, A.; Orij, E. D.; Kamer, P. C. J .; van Leeuwen, P. W.
N. M. Organometallics 1995, 14, 34.
(20) van Rooy, A. Thesis 1995, page 48.
(14) Heil, B.; Marko´, L.; Bor, G. Chem. Ber. 1971, 104, 3418.
(15) Vidal, J . L.; Walker, W. E. Inorg. Chem. 1981, 20, 249.
(16) Liu, G.; Volken, R.; Garland, M. Organometallics 1999, 18, 3429.
(17) van Leeuwen, P. W. N. M.; Roobeek, C. F. J . Organomet. Chem.
1983, 258, 343.
(18) J ongsma, T.; Challa, G.; van Leeuwen, P. W. N. M. J . Orga-
nomet. Chem. 1991, 421, 121.
(21) Cavalieri d’Oro, P.; Raimondi, L.; Pagani, G.; Montrasi, G.;
Gregorio, G.; Andreetta, A. Chim. Ind. 1980, 62, 572.
(22) Yagupsky, G.; Brown, C. K.; Wilkinson, G. J . Chem. Soc. A
1970, 1392.
(23) Morris, D. E.; Tinker, H. B. CHEMTECH 1972, 554.
(24) Bianchini, C.; Man Lee, H.; Meli, A.; Vizza, F. Organometallics
2000, 19, 849.

432 Organometallics, Vol. 20, No. 3, 2001
van der Slot et al.
Ch a r t 1
F igu r e 1. Variation of the rhodium concentration. Condi-
tions: T ) 41 °C, pCO ) pH ) 10 bar, [L] ) 19 mM,
[1-octene] ) 0.59 M. Values o2f the rate are least-squares
fits (r2 g 0.998) of lines from plots of millimoles of aldehyde
vs time over 0-10% conversion.
based on a 1,3,5-N,N′,N′′-trisubstituted biuret struc-
ture.25 These ligands combine the π-acidity of phosphites
with the steric properties of phosphines. This unique
combination of electronic and steric ligand properties
leads to an increase of the hydroformylation activity in
comparison to that of phosphine-based catalyst systems.
Under hydroformylation reaction conditions, these bulky
ligands formed exclusively the equatorial-equatorial
rhodium-hydride complex; although the ligands are
very bulky, the catalyst selectivity for the linear alde-
hyde remained moderate. This observation suggests
that during the catalytic cycle an equatorial-apical
complex might be formed. Here we present a mecha-
nistic study using these phosphorus diamide ligands in
order to understand the catalyst performance. The rate-
limiting step in the hydroformylation reaction of 1-alk-
enes using ligand 1 (Chart 1) was investigated and the
solution structure of the resting state of the catalyst was
studied using (in situ) high-pressure spectroscopic
techniques.
Sch em e 2. Equ ilibr iu m betw een Mon om er ic a n d
Dim er ic Rh od iu m Com p lexes
Ta ble 1. Su bstr a te Con cen tr a tion Dep en d en cy on
th e Hyd r ofor m yla tion Rea ction Ra tea
b
[1-octene] (M)
conversn (%)
kald
l/b (%)
2-octene (%)
0.44
0.59
0.59
0.88
1.18
22
13
20
25
23
152
165
162
183
201
1.7
1.7
1.7
1.9
1.9
2
2
1
2
1
a
Resu lts a n d Discu ssion
Conditions: T ) 41 °C, pCO ) pH ) 10 bar, [Rh] ) 1.9 mM,
[ligand] ) 19 mM. Samples were t2aken between 0 and 10%
b
Kin et ic St u d ies. The rate-controlling steps of the
hydroformylation reaction of 1-octene using HRh(1)2-
(CO)2 (2a ) were determined by a detailed kinetic study.
In this study the concentration dependency of the
reaction rate of all the reactants was investigated. The
initial rate of aldehyde production was determined over
the first 10% of conversion. Earlier research showed that
ligand 1 can form mixtures of two different hydride
complexes (HRh(1)2(CO)2 and HRh(1)(CO)3) under hy-
droformylation conditions.25 The performances of these
two catalysts in the hydroformylation reaction are
different. Therefore, the optimal rhodium and ligand
concentrations were determined using high-pressure IR
spectroscopy to ensure exclusive formation of HRh(1)2-
(CO)2.
One of the important side reactions obtained under
hydroformylation reaction conditions is the dimerization
reaction depicted in Scheme 2. High rhodium and low
H2 concentrations and low temperatures promote the
formation of these dimeric rhodium complexes.11,26,27 To
investigate the effect of the rhodium concentration on
the reaction rate, we varied this concentration between
0.5 and 4 mM (see Figure 1). The rate of aldehyde
production is linearly proportional to the rhodium
conversion. In units of (mol of aldehyde) (mol of Rh)-1 h-1
averaged over time.
F igu r e 2. Variation of the partial CO pressure. Condi-
tions: T ) 41 °C, [Rh] ) 2 mM, Rh/L ) 1/10, pH ) 10 bar,
[1-octene] ) 0.59 M. Values of the rate are lea2st-squares
fits (r2 g 0.998) of lines from plots of millimoles of aldehyde
vs time over 0-10% conversion.
concentration. This first-order dependency indicates
that dimerization of the rhodium complexes is not
involved under the applied hydroformylation conditions
using ligand 1.
The 1-octene concentration was varied between 0.4
and 1.2 M. The results depicted in Table 1 show a rate
dependency of 0.3. The partial CO pressure was varied
between 10 and 35 bar. The natural logarithm plot of
the reaction rate against the partial CO pressure shows
a negative first-order dependency on the partial CO
pressure (see Figure 2). The partial hydrogen pressure
was varied between 10 and 40 bar (Table 2). The natural
(25) 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.
(26) Evans, D.; Yagupsky, G.; Wilkinson, G. J . Chem. Soc. A 1968,
2660.
(27) J ames, B. R.; Mahajan, D.; Rettig, S. J .; Williams, G. M.
Organometallics 1983, 2, 1452.

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 433
Ta ble 2. CO a n d H2 P a r tia l P r essu r e Dep en d en cy
on th e Hyd r ofor m yla tion Rea ction Ra tea
Sch em e 3
b
pH (bar) pCO (bar) conversn (%) kald
l/b (%) 2-octene (%)
2
10.5
10.6
10.6
10.5
10.0
10.6
11.8
18.1
29.9
34.7
11.5
12.1
20.2
24.3
29.5
38.5
10.6
10.5
10.3
10.5
25
23
20
23
29
18
26
26
24
24
191
183
108
93
77
52
193
263
407
456
1.7
1.8
1.7
1.8
1.7
1.7
1.8
1.8
1.8
1.8
2
3
3
2
2
2
3
3
1
2
a
Conditions: T ) 41 °C, [Rh] ) 1.9 mM, [ligand] ) 19 mM,
[1-octene] ) 0.59 M. Samples were taken between 0 and 10%
b
conversion. In units of (mol of aldehyde) (mol of Rh)-1 h-1
hydroformylation mechanism. Several reaction steps in
the proposed mechanism are involved controlling the
reaction rate. The alkene coordination/hydride migra-
tion and the hydrogenolysis have very similar rates
leading to a combination of the type I and type II
kinetics, as given in eq III, of the overall reaction rate
averaged over time.
Ta ble 3. Liga n d Con cen tr a tion Dep en d en cy on th e
Hyd r ofor m yla tion Rea ction Ra tea
b
[L] (mM)
conversn (%)
kald
l/b (%)
2-octene (%)
19
29
37
49
67
25
28
26
25
22
226
206
191
183
165
1.9
1.9
1.9
1.9
2.0
2
2
2
2
2
G[Rh][H2][alkene]
rate(overall) )
(III)
H[CO] + K[alkene] + [H2]
a
Conditions: T ) 41 °C, pCO ) pH ) 10 bar, [Rh] ) 1.9 mM,
of the approximate form. The overall reaction rate is
strongly dependent on the reaction conditions used. At
high hydrogen pressure, the catalyst will show type I
kinetics: i.e. the alkene coordination/hydride migration
determines the overall rate of reaction. When the alkene
concentration is increased, the catalyst will show type
II kinetics: i.e., the hydrogenolysis determines the
overall reaction rate. Therefore, noninteger reaction
orders given above are valid only within the window of
the concentration and pressure ranges studied. As a
result of these kinetics, both rhodium-hydride and
rhodium-acyl complexes (2a , 3a ) are expected to be
present under the conditions studied.
2
[1-octene] ) 0.59 M. Samples were taken between 0 and 10%
conversion. In units of (mol of aldehyde) (mol of Rh)-1 h-1
b
averaged over time.
logarithm plot of the reaction rate against the partial
hydrogen pressure shows a large positive (0.8) depen-
dency on the hydrogen pressure. We studied the effect
of the ligand concentration on the reaction rate by
varying this concentration between 19 and 67 mM. The
results show a rate dependency of -0.3 (Table 3).
The positive reaction order found for the alkene
concentration and the negative reaction order found for
the CO pressure and ligand concentration are similar
to the results found for the triphenyl phosphine based
catalyst.21 The rate equation for this type of kinetics,
as proposed by van Leeuwen et al.28 as type I kinetics,
is
Deu ter iofor m yla tion Rea ction . To determine the
reversibility of the alkene coordination and hydride
migration, we performed a deuterioformylation experi-
ment.29,30 If rhodium-alkyl formation by hydride mi-
gration to the coordinated alkene is irreversible, then
the aldehydes will contain deuterium labels on the
aldehyde carbon atom and the â-carbon atom. Revers-
ible rhodium-alkyl formation will result in R-deutera-
tion (see Scheme 3). If alkene coordination and hydride
migration are both reversible and fast, then deuterium
will be incorporated into the alkene and the number of
deuterium labels present in the aldehyde molecule will
differ in part from 2. The deuterioformylation was
performed under the same conditions as the hydro-
formylation experiments (T ) 41 °C, [Rh] ) 1.9 mM,
A[Rh][alkene]
rate(type I) )
(I)
B[CO] + C[alkene] + [L]
Rate-influencing reactions are CO or ligand dissociation,
alkene coordination, and hydride migration. The resting
state of the catalyst fitting this kinetics is the rhodium-
hydride complex 2a . Type I kinetics is observed for most
ligand-modified rhodium catalysts. The large positive
reaction order found for the hydrogen pressure suggests
that the hydrogenolysis step is also important in
determining the overall reaction rate. The rate equation
fitting these results which were proposed as Type II
kinetics is28
Rh/L ) 1/10, [1-hexene] ) 0.81 M, pCO ) pD ) 10 bar).
2
1-Hexene was used for the deuterioformylation instead
of 1-octene to facilitate comparison with literature data
1
2
of the H and H NMR spectra.29 The rates and selectiv-
ity for 1-hexene are comparable with those of 1-octene.
The production of aldehydes was monitored by gas
D[Rh][H2]
rate(type II) )
(II)
E[CO] + F[H2] + [L]
(28) van Leeuwen, P. W. N. M.; Casey, C. P.; Whiteker, G. T. In
Rhodium Catalysed Hydroformylation; van Leeuwen, P. W. N. M.,
Claver, C., Ed.; Kluwer Academic: Dordrecht, The Netherlands, 2000;
Chapter 4.
(29) Casey, P. C.; Petrovich, L. M. J . Am. Chem. Soc. 1995, 117,
6007.
Coordinatively saturated rhodium-acyl complexes such
as 3a will be present in the reaction mixture as resting
states of the catalyst.
The kinetic data obtained for 2a show that the rate-
determining step of the hydroformylation reaction using
this catalyst cannot be reduced to one single step of the
(30) Lazzaroni, R.; Uccello-Barretta, G.; Benetti, M. Organometallics
1989, 8, 2323.

434 Organometallics, Vol. 20, No. 3, 2001
van der Slot et al.
Ta ble 4. Absor p tion s Obta in ed fr om in Situ IR
Exp er im en ts
chromatography. The number of deuteriums incorpo-
rated into 1-hexene, heptanal, and 2-methylhexanal was
monitored by gas chromatography/mass spectroscopy.
The activity and selectivity for the deuterioformylation
of 1-hexene were similar to those of the hydroformyla-
tion of 1-hexene. Gas chromatography/mass spectros-
copy showed hardly any deuterium enrichment at low
conversions of 1-hexene. At higher conversion (g95%),
a slight deuterium enrichment of n-hexene (∼3%) was
observed, which is due to the increasing concentration
of the less reactive 2-alkenes at high conversions. Two
deuterium labels were incorporated in the aldehyde
molecules. This suggests that â-hydride elimination and
alkene dissociation occur only in very few occasions. The
linear and branched aldehydes were separated from the
reaction mixture by distillation and column chromatog-
raphy after complete conversion of 1-hexene. The deu-
a
νCO (cm-1
)
pCO (bar) pH (bar) [1-octene] (M) disappearingb appearingb
2
7
7
0.2
2070 (2a )
2018 (2a )
2085 (w)
2077 (sh)
2028 (w)
2010 (sh)
2001 (m)
1991 (m)
1967 (w)
-
2085 (w)
2077 (sh)
2028 (w)
2010 (sh)
2001 (m)
1991 (m)
1967 (w)
7
7
32
7
0.2
0.6
-
2070 (2a )
2018 (2a )
1
terium distribution was determined using H and 2H
All experiments were performed in cyclohexane at 40 °C.
a
b Appearing absorption bands are obtained for carbonyl frequencies
of complexes that are formed during the hydroformylation reaction;
disappearing absorption bands are obtained for carbonyl frequen-
cies of complexes that are converted to other complexes.
NMR spectroscopy. We were not able to detect alde-
hydes containing the deuterium on the R-carbon atom
1
2
using H NMR spectroscopy. The H NMR spectrum of
the mixture of the linear and branched aldehydes shows
deuterium labels exclusively at the aldehyde carbon and
â-carbon atoms. Similar to the results of Casey and
Petrovich29 obtained for the bidentate phosphine ligands,
a deuterium label at the R-carbon was not observed
within the detection limits of 2H NMR spectroscopy.
Therefore, we conclude that for this ligand, under the
conditions studied, hydride migration is virtually an
irreversible process. Probably, the alkene dissociation
is much faster than the hydride migration. For a small
percentage of molecules that do undergo â-hydride
elimination, the alkene dissociates before hydride mi-
gration occurs again. Both linear and branched rhodium-
alkyl complexes are formed irreversibly, and therefore
the regioselectivity of the reaction is determined in the
hydride migration step.
order rate dependency on the rhodium concentration
observed in the kinetic experiments.
The kinetic experiments indicate the presence of
both coordinatively saturated rhodium-hydride and
rhodium-acyl complexes during the hydroformylation.
The spectroscopic data of several rhodium-hydride
complexes obtained with ligand 1 are depicted in Table
5. Rhodium-hydride complexes containing two phos-
phorus ligands can have either equatorial-equatorial
(ee) or equatorial-apical (ea) coordinated phosphorus
ligands.31 The formation of the ea isomer is not likely
since complexes with ea coordinated phosphorus ligands
were never observed under hydroformylation conditions
using ligand 1.25 Immediately after addition of 1-octene,
the strong absorption band of nonanal (1734 cm-1
)
appeared in the IR spectrum, proving that the hydro-
formylation reaction has started. The amount of 2a
dropped considerably upon addition of 1-octene. How-
ever, comparing the intensity of the carbonyl frequencies
of 2a before and after addition of 1-octene showed that
2a was still present in low concentration during the
hydroformylation reaction. Seven new absorptions ap-
peared in the terminal carbonyl region, indicating that
complex 2a was converted (in part) to several new
carbonyl-containing rhodium complexes. Only one rhod-
ium-hydride complex was observed during the hydro-
formylation reaction, the rhodium-hydride complex
containing two phosphorus diamide ligands (Table 5).
The kinetic experiments showed that the hydro-
genolysis (step 6, Scheme 1) is a relatively slow step in
the hydroformylation reaction and rhodium-acyl com-
plexes are expected to be present during the hydro-
formylation reaction. Several possible structures of
rhodium-acyl complexes are depicted in Figure 4. Since
the catalyst does not show high selectivity for the linear
aldehyde, it can be concluded that both the linear and
branched rhodium-acyl complexes are formed (3a ,b).
Coordination of one of the phosphorus atoms at an
apical position leading to ea coordinated complexes is
In Situ IR Stu d ies. In the previous section, we
showed that several elementary reaction steps of the
hydroformylation reaction influence the overall reaction
rate. On the basis of these results rhodium-hydride and
rhodium-acyl complexes were proposed to be present
as potential “resting” states during the hydroformylation
reaction. The complexes present were investigated using
in situ spectroscopic techniques. The results obtained
in the in situ HP IR experiments are presented in Table
4. In a typical experiment the hydride complex 2a was
formed in situ, and after addition of 1-octene the
hydroformylation reaction was monitored using a rapid
scan IR technique (7 scans/s). The difference spectra
obtained (Figure 3) show negative absorption bands for
the carbonyl frequencies of complexes that are converted
(in part) to other complexes after addition of the
substrate. Positive absorption bands are obtained for
carbonyl frequencies of complexes that are formed
during the hydroformylation reaction. The rhodium
hydride complex 2a (νCO ) 2070, 2018 cm-1; Table 5) is
the only rhodium complex present in all experiments,
both before addition and after complete conversion of
1-octene. The difference IR spectra show absorption
bands in the terminal carbonyl region only. No absorp-
tions of bridging carbonyls are observed, which confirms
that inactive rhodium dimers or clusters are not present
during the reaction, as already concluded from the first-
(31) 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.

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 435
F igu r e 3. Difference spectrum of the in situ IR experiment.
Ta ble 5. NMR a n d IR Da ta HRh (1)x(CO)4-x (x ) 1-3) Com p lexes
b
compd
δ(1H)a (ppm)
δ(31P)a (ppm)
δ(13CO)a (ppm)
coupling constants (Hz)
ν )
CO (cm-1
HRh(1)(CO)3
HRh(1)2(CO)2 (2a )
-10.3
104
110
ndc
J RhH e 3, J RhP ) 177, J PH ) 15
J RhH e 3, J RhP ) 181, J CP ) 22,
J RhC ) 58, J CH ) 16, J PH e 3
J RhH e 3, J RhP ) 169, J PH ) 13,
J CP ) 13, J RhC ) 52, J CH ) 36
2095, 2045, 2008
2070, 2018
-10.6 (broad)
194d
HRh(1)3CO
-11.1
117
200
2019
a
d
All NMR experiments were performed in toluene-d8. b Measured in cyclohexane. c Not determined. The equatorial and apical carbonyl
ligands are in fast exchange under the conditions studied.
1600-1700 cm-1. Therefore, the absorption of rhodium-
acyl species could not be observed in the IR spectra.
Rhodium hydride complex 2a was the only rhodium
complex observed during the hydroformylation reaction
when the partial hydrogen pressure was increased to
32 bar (Table 4). Higher hydrogen pressures facilitate
the hydrogenolysis step, and the simplified type I kinetic
equation becomes valid (vide supra). The absence of
terminal carbonyl bands in the difference IR spectra
observed under high partial hydrogen pressure indicates
that under these conditions only the rhodium-hydride
complex 2a is present. The amount of 2a decreased and
the intensity of the additional carbonyl bands increased
with increasing alkene concentration. This indicates
that at higher alkene concentration the contribution of
the hydrogenolysis step as the rate-controlling step
increases, finally resulting in type II kinetics. The shift
of rate control, depending on the conditions used, shows
that the rate expression of this catalyst system is
strongly dependent on the conditions used.
F igu r e 4. Possible structures of rhodium acyl complexes.
also possible. The ea coordination mode has not been
observed for the hydride complex (2a ), but coordination
of the larger acyl moiety at one of the apical positions
can lead to distortion of the trigonal-bipyramidal struc-
ture or coordination of a phosphorus ligand trans to the
acyl ligand. The new carbonyl bands obtained in the IR
spectrum probably belong to several of the rhodium-
acyl complexes, as depicted in Figure 4. The strong
amide bands of the ligand in this region obscure the
expected rhodium-acyl absorption band10,20 around
Stoich iom etr ic Rea ction s. The hydroformylation
reaction was investigated in a stepwise manner by the
subsequent reactions of 2a with alkene, CO, and H2 in
an attempt to characterize the complexes formed during
the hydroformylation reaction. In the absence of H2,
complex 2a , alkene, and CO can undergo all hydro-
formylation reaction steps except the hydrogenolysis
step and rhodium-acyl complexes may be formed. The
IR data obtained after the subsequent reaction steps are

436 Organometallics, Vol. 20, No. 3, 2001
van der Slot et al.
Ta ble 6. IR F r equ en cies Obta in ed in th e
Stoich iom etr ic Rea ction of 2a , 1-Octen e, CO, a n d
H2
conditionsa
νCO (cm-1
)
2a under 5 bar of CO
2a + 25 equiv of 1-octene
under 5 bar of CO
2080, 2017
2085 (w), 2077 (sh), 2028 (w),
2010 (sh), 2001 (m),
1991 (m), 1967 (w)
2080, 2017
2a + 25 equiv of 1-octene
under 10 bar of CO/H2 (1/1)
a
All reactions were performed in cyclohexane at room temper-
ature.
presented in Table 6. The carbonyl frequencies obtained
after the reaction of 2a , 1-octene, and CO were identical
with those obtained in the in situ IR experiments
obtained during the hydroformylation reaction, except
for the absence of the carbonyl frequency due to the
aldehyde. Subsequent addition of hydrogen showed
aldehyde formation, and after all the 1-octene was
converted, 2a was the only complex present again.
Therefore, we conclude that the complexes formed in
the stoichiometric reaction of 2a , alkene, and CO are
equal to the most abundant complexes present during
the hydroformylation reaction at low H2 pressure. This
indicates that in this instance the stoichiometric reac-
tion is an elegant and reliable method to study the
intermediates of the hydroformylation reaction in a
stationary system.
NMR Stu d y of th e Com p lexes Obta in ed in th e
Stoich iom etr ic Rea ction . High-pressure (HP) NMR
spectroscopy is a powerful technique to determine the
precise structure of complexes present during the hy-
droformylation reaction. Higher sample concentrations
and longer collection times are needed for NMR studies.
The stoichiometric hydroformylation of 1-hexene using
NMR spectroscopy was investigated using the same
procedure as that for the IR experiments. The experi-
ments were performed in a high-pressure NMR flow cell
as described by Iggo and co-workers.12 The advantages
of an HP flow cell instead of an HP NMR tube32 are the
continuous supply of reactants and optimal mixing of
the reactants (minimization of diffusion problems).
Homogeneously catalyzed reactions can be monitored
using this flow cell, and stable intermediates can be
characterized using different NMR techniques.
Rhodium complex 2a was prepared in situ in the
NMR cell from the rhodium precursor (Rh(acac)(CO)2)
and 5 equiv of ligand 1 by heating at 80 °C for 1 h under
20 bar of CO/H2 (1/1) (spectrum a, Figure 5). Complex
2a is the only rhodium complex obtained according to
the 31P NMR spectrum. The hydrogen gas was removed
by bubbling CO through the solution for 30 min at 253
K. Approximately 25 equiv of 1-hexene was injected into
the NMR cell at 253 K, and the solution was warmed
slowly to room temperature. When the temperature was
raised, the 31P NMR spectrum started to broaden and
both the 1H NMR and 31P NMR spectra showed that
the resonance due to the hydride complex decreased in
intensity. To trap all complexes present, the solution
was cooled immediately to 253 K. This results in NMR
spectrum b as presented in Figure 5. The spectrum at
this temperature showed an additional (broad) doublet
(δ 109.5 ppm, J RhP ) 205 Hz) upfield from the doublet
F igu r e 5. Overview of the 31P{1H} NMR spectra of the
stepwise reaction of 2a , 1-hexene, and CO with H2: (a) 31P-
{1H} NMR spectrum of 2a under 5 bar of CO; (b) 31P{1H}
NMR spectrum of 2a in the presence of 5 bar of CO and 25
equiv of 1-hexene; (c) 31P{1H} NMR spectrum after com-
plete conversion of 2a ; (d) 31P{1H} NMR spectrum after
addition of 5 bar of CO/H2 (1/1). All spectra were recorded
at 253 K.
of 2a (spectrum b, Figure 5). Complex 2a was converted
almost completely to the new compounds by warming
the solution to room temperature again. The addition-
ally formed doublet is only visible at 253 K (spectrum
c, Figure 5). A broad resonance appeared downfield from
the doublet, indicating that probably more than one
complex was present (a major and a minor complex).
The 31P NMR spectrum at room temperature showed a
very broad peak at the position of the free ligand (64
ppm), indicating that the coordinated phosphorus ligands
are in exchange with the free phosphorus ligand. The
1
hydride resonance in the H NMR spectrum has disap-
peared, and no aldehyde resonance was observed. We
were not able to obtain a sharp 31P resonance of the
doublet formed under these conditions because of the
temperature limits of the NMR flow cell. When the gas
flow was changed from 5 bar of CO to 5 bar of CO/H2
(1/1), the broad doublet in the phosphorus NMR spec-
trum disappeared and the characteristic doublet due to
the hydride complex 2a was observed (spectrum d,
1
Figure 5). The H NMR spectrum showed the reappear-
ance of the hydride resonance at -10.6 ppm together
with an aldehyde resonance at 9.3 ppm. Repeating this
sequence using this NMR sample gave the same results,
indicating the retention of activity for the rhodium-
catalyzed reaction. These results confirm that the
complex formed in the absence of H2 is an intermediate
of the hydroformylation reaction.
The new complexes could also be prepared in an NMR
tube, and they are stable for several days under 3 bar
of CO at 193 K. The 31P NMR spectrum showed again
a very broad resonance at the position of the free ligand
at room temperature. When the sample was cooled to
253 K, the 31P NMR spectrum showed the appearance
of the broad doublet at 109.5 ppm (J RhP ) 205 Hz)
together with a broad resonance downfield from this
doublet. The spectrum broadened when the temperature
was decreased. The 31P NMR spectrum obtained at 193
K will be discussed in more detail later.
(32) Elsevier, C. J . J . Mol. Catal. 1994, 92, 285.

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 437
F igu r e 6. Variable-temperature 13C{1H} NMR spectra obtained after the reaction of 2a , 1-hexene, and CO: (1a) 13C{1H}
NMR spectrum obtained at 253 K; (1b) simulation of the 13C{1H} NMR spectrum of 1a; (2a ) 13C{1H} NMR spectrum obtained
at 233 K; (2b) simulation of the 13C{1H} NMR spectrum 2a; (3a) 13C{1H} NMR spectrum obtained at 223 K; (3b) simulation
of 13C{1H} NMR spectrum 3a at 223 K; (3c) simulation of 13C{1H} NMR spectrum 3c at 223 K; (3d) simulation of the
13C{1H} NMR spectrum obtained at 223 K of 3a :3d ) 1:0.4. The asterisk denotes an impurity.
The in situ IR experiments revealed that rhodium-
acyl complexes may be formed in the reaction of 2a ,
1-hexene, and CO. Rhodium-acyl carbon atoms have
characteristic chemical shifts in the 13C NMR spectrum
around 235 ppm. Therefore, the reaction was performed
with 13CO to enable the identification of the rhodium-
acyl resonance in 13C NMR. Figure 6 shows an overview
of the 13C{1H} NMR spectra obtained in a variable-
temperature experiment. All the resonances in the 13C
NMR spectrum were very broad at room temperature.
The resonances sharpened when the temperature was
decreased. This indicates that several intra- and inter-
molecular exchange processes are involved (vide infra).
The 13C NMR spectrum at 223 K (spectrum 3a, Figure
6) showed two rhodium-acyl resonances (δ 230.0 ppm
and δ 227.2 ppm) and three different rhodium-carbonyl
resonances (δ 194.7 ppm, δ 193.7 ppm and δ 190.6 ppm),
indicating that we are dealing with two different rho-
dium-acyl complexes. The structures of these two rho-
dium-acyl complexes will be discussed in the following
sections.
F igu r e 7. 103Rh-13C HMQC spectrum recorded at 183 K.
The delay time τ ) 1/[2J RhC] calculated from J RhC ) 70 Hz
was used.
Ch a r a cter iza tion of Rh od iu m -Acyl Com p lexes
3a ,b. The 13C{1H} NMR spectrum at 253 K (spectrum
3a, Figure 6) showed two rhodium-acyl resonances and
three different rhodium-carbonyl resonances. The 13C
COSY90 spectrum showed strong cross-peaks of the
carbonyl ligand at δ 193.7 and 190.6 ppm and weak
cross-peaks for the carbonyl ligand at δ 190.6 ppm and
the rhodium-acyl resonance at δ 230.0 ppm. Therefore,
we conclude that these three carbonyl resonances belong
to the same rhodium-acyl complex. We were not able
to obtain information about the rhodium-acyl reso-
nance at δ 227.2 ppm and the terminal CO resonance
at δ 194.7 ppm using two-dimensional NMR techniques,
because of the low concentration of this complex. Selec-
tive decoupling of the phosphorus resonance at δ 109
ppm at 233 K led to simplification of the carbon NMR
spectrum to doublets of doublets for the acyl resonance
at 230.0 ppm and the carbonyl resonance at 190.6 ppm.
From this spectrum the rhodium-carbon coupling
constants were calculated to be 18 Hz for the acyl carbon
and 42 Hz for the carbonyl ligand. This spectrum also
shows the two-bond carbonyl-acyl coupling constant of
2
28 Hz. The large J CC coupling constant of 28 Hz is
indicative of a trans orientation of the carbonyl ligand
and the acyl ligand.
Several 103Rh-13C HMQC spectra were recorded
using different delay times (τ ) 1/[2J RhC]). Figure 7
shows the 103Rh-13C HMQC spectrum obtained using
a delay time τ calculated from J RhC ) 70 Hz. The
carbonyl resonance at 193.7 ppm showed cross-peaks
with a rhodium resonance at approximately -70 ppm.
The multiplicity of the rhodium resonance (triplet)
shows that two phosphorus atoms with approximately
equal J RhP coupling constants are coordinated to the

438 Organometallics, Vol. 20, No. 3, 2001
van der Slot et al.
F igu r e 8. NMR data for the major rhodium-acyl complex formed in the reaction of 2a , 1-hexene, and CO.
rhodium nucleus. Cross-peaks were obtained for the
carbon resonances at 190.6 and 230.0 ppm and the
triplet at ∼-70 ppm in the rhodium dimension in the
103Rh-13C HMQC spectrum when a delay time τ calcu-
lated from J RhC ) 40 Hz was used. A rhodium resonance
at the same chemical shift was observed in the 103Rh-
31P HMQC spectrum. From the results of the various
NMR experiments we conclude that the major complex
obtained at 223 K after reaction of 2a , 1-hexene, and
CO is a rhodium-acyl complex having two equatorially
coordinated phosphorus ligands and two magnetically
inequivalent carbonyl ligands. The precise NMR data
obtained after simulation are presented in Figure 8; the
simulated 13C{1H} NMR spectrum corresponding to
these data is presented in Figure 6 (spectrum 3b). The
two phosphorus atoms are coordinated in the equatorial
plane of the trigonal-bipyramidal structure. The carbon
resonance at 193.7 ppm shows a large coupling constant
to the phosphorus nuclei (37 Hz), as also reported for
equatorial carbonyl ligands in rhodium-hydride com-
plexes (HRhLkL(CO)2, LkL ) bidentate phosphite
ligand).33 The large coupling constant of the equatorial
carbonyl to the rhodium nucleus (1J CRh ) 70 Hz) has
also been reported by Brown and Kent.2 The cis coupling
of the carbonyl with the acyl carbon resonance is
unresolved. The magnitude of the carbon-phosphorus
coupling constant for the apical carbonyl (20 Hz) is
the phosphorus atoms lost their time-averaged equiva-
lence. At 223 K, the slow-exchange region was reached
and a doublet of doublets appeared at δ 227.2 ppm (J RhC
) 17 Hz, J CP ) 83 Hz, J CP′ e 3 Hz). Simulation using
these data results in spectrum 3c (Figure 6). This acyl
resonance belongs to a complex that contains one
equatorial and one apical phosphorus atom (3c,d , Figure
4). A similar complex for triphenylphosphine has been
reported by Brown and Kent.2 The two carbonyl ligands
in the equatorial plane show a sharp doublet in the
terminal carbonyl region (δ 194.7 ppm, J RhC ) 79 Hz).
The terminal carbonyls exhibit equal coupling constants
to both phosphorus nuclei due to intramolecular ex-
change of the phosphorus atoms. When the temperature
is decreased, the terminal carbonyl resonance broad-
ened, and at 193 K a multiplet was observed. The J CP
coupling constants could not be resolved because of the
overlap of this resonance and the terminal carbonyl
resonance of the other acyl complex.
Summarizing, we conclude that in the reaction of 2a ,
1-hexene, and CO, two rhodium-acyl complexes are
formed. Spectrum 3d (Figure 6) shows the complete
simulation of the 13C{1H} NMR spectrum obtained at
223 K. The 13C{1H} NMR spectrum contains both
rhodium-acyl complexes 3a ,b and 3c,d in a ratio of 1
to 0.4. Both rhodium-acyl complexes have a trigonal-
bipyramidal structure with the acyl moiety coordinated
at an apical position. Both complexes contain two
phosphorus ligands and two carbonyl ligands. The major
complex formed contains two phosphorus ligands coor-
dinated in the equatorial plane of the trigonal bipyra-
mid; the minor complex has one phosphorus atom
coordinated at an equatorial position and one at an
apical position of the trigonal bipyramid. In Figure 4,
we proposed four different structures for the acyl
complexes present (3a -d ). On the basis of the NMR
data obtained, we are not able to distinguish between
the linear- and branched-acyl structures. However, since
the branched-acyl ligand introduces more steric hin-
drance close to the rhodium center than the linear-acyl
ligand, it is tempting to suggest that the major rho-
dium-acyl complex contains a linear acyl ligand (com-
plex 3a ), whereas the minor rhodium-acyl complex
contains a more bulky branched acyl ligand (complex
3d ). The linear-to-branched ratio found in the catalytic
reaction cannot be explained by the ratio of 3a to 3d
found in the NMR experiments, because these two
2
larger than normally observed for J CP in a pure cis
2
orientation. This relatively large J CP might be ex-
plained by distortion of the trigonal-bipyramidal struc-
ture. This is not unlikely, since distortion of the trigonal-
bipyramidal structure is also observed for hydride
complexes containing bulky monodentate and bidentate
phosphite and phosphorus diamide ligands.25,33
Ch a r a cter iza tion of Rh od iu m -Acyl Com p lexes
3c,d . The second (minor) acyl resonance in the carbon
NMR spectrum at δ 227.2 ppm shows a double triplet
at 253 K (Figure 6, spectrum 1a). Selective decoupling
of the phosphorus resonance at 109 ppm led to simpli-
fication of the acyl resonance to a doublet (J RhC ) 17
Hz). This indicates that the acyl resonance belongs to a
complex that contains two (time-averaged) equivalent
phosphorus atoms. The J CC coupling constant of the acyl
resonance and the carbonyl ligands is small, indicative
of a cis relation. When the temperature was decreased,
(33) van Rooy, A.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.;
Goubitz, K.; Fraanje, J .; Veldman, N.; Spek, A. L. Organometallics
1996, 15, 835.

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 439
F igu r e 9. Different exchange processes of rhodium-acyl complexes.
rhodium-acyl complexes can have different reactivities
toward hydrogen.
Exch an ge P r ocesses Obser ved for th e Rh odiu m -
Acyl Com p lexes. The rhodium-acyl complexes 3a -d
are involved in several exchange processes, depending
on the temperature. An overview of the exchange
processes is presented in Figure 9. The complexes
containing the branched aldehyde have been omitted
for the sake of brevity. At room temperature one broad
resonance for the free ligand is observed in the 31P{1H}
NMR spectrum, indicating that there is an exchange
F igu r e 10. 31P{1H} NMR spectrum obtained after reaction
of 2a , 1-hexene, and CO at 193 K: (1a) experimental data;
(1b) simulated 31P{1H} NMR spectrum of rhodium-acyl 3a
at 193 K, The resonance denoted by an asterisk belongs
probably to the (minor) rhodium acyl 3d .
with the coordinated ligand. A broad doublet at 109.5
ppm and a sharp resonance for the free ligand are
observed when the temperature is decreased to 253 K.
The exchange process of the coordinated and free
ligands is slow on the NMR time scale at this temper-
ature. One average resonance is observed for the two
rhodium-acyl complexes with ee and ea coordinated
phosphorus ligands. The isomerization between 3a ,b
and 3c,d is still fast on the NMR time scale. The
phosphorus NMR broadens again upon cooling to 193
K, indicating that the isomerization becomes slow on
the NMR time scale. In both complexes the phosphorus
ligands lose their equivalence. Because of the low
concentration of complex 3d we were not able to resolve
phosphorus coupling constants and chemical shifts of
this complex. The 31P NMR spectrum obtained at 193
K is shown in Figure 10. It should be noted that the
free ligand shows fluxional behavior between two con-
formers. A variable-temperature 31P NMR experiment
of the free ligand at 173 K showed two different
conformers in a 1:1 ratio (δ 65.7 and 64.7 ppm). The
31P NMR spectrum 1a displayed in Figure 10 is a result
of two different processes in which complex 3a is
involved. First, the two coordinated phosphorus atoms
become magnetically inequivalent upon cooling (δ(P1)
Ta ble 7. 31P {1H} NMR Da ta of 3a a fter Sim u la tion
of th e Sp ectr u m Obta in ed a t 193 K
compd
δ(31P) (ppm)
J PP (Hz)
J RhP (Hz)
isomer 1
P(1) 112.3
P(2) 105.8
P(1) 112.4
P(2) 106.2
117
P(1) 202
P(2) 235
P(1) 203
P(2) 222
isomer 2
115
∼112 ppm, δ(P2) ∼105 ppm). Second, interconversion
between two ligand conformations becomes important;
one of the phosphorus ligands is still in a fast exchange
and the second phosphorus atom reaches the slow-
exchange region. Simulation of these two processes
results in the spectrum presented for spectrum 1b in
Figure 10. Chemical shifts and coupling constants are
presented in Table 7.
The carbonyl ligands of complex 3a are involved in
several exchange processes at temperatures above 223
K. One averaged signal at δ 193 ppm for both carbonyl
ligands and dissolved 13CO was observed in the 13C

440 Organometallics, Vol. 20, No. 3, 2001
van der Slot et al.
NMR spectrum at room temperature, and one very
broad resonance in the acyl region was observed. The
averaged signal at 193 ppm is due to fast exchange of
the carbonyl ligands with the dissolved 13CO. The
resonance of the equatorial CO is still very broad at 253
we conclude that the regioselectivity for either the linear
or the branched aldehyde is determined in these reac-
tion steps.
In previous studies both we and others have discussed
whether a bis-equatorial coordination mode of the
phosphorus ligands should lead to high linear-to-
branched ratios. The phosphorus diamide ligand used
forms exclusively a hydride complex in which the
phosphorus ligands are coordinated in the equatorial
plane. The rhodium-acyl complexes observed contain
phosphorus ligands coordinated in both the ee and ea
coordination modes. As already argued by van der Veen
and co-workers,31 formation of a hydride complex having
bis-equatorially coordinated phosphorus ligands does
not guarantee a high linear-to-branched ratio in the
product. As shown above, monodentate ligands may well
have the flexibility to form ea complexes for the more
hindered branched-alkyl and acyl complexes.
K (Figure 6, spectrum 1a), but the resonance of the
13
apical CO sharpens. No resonance for the dissolved
-
CO gas was observed at this temperature. Simulation
of the 13C{1H} NMR data (Figure 6, spectrum 1b)
showed that the carbonyl ligands exchange with dis-
solved 13CO gas with different rates (Figure 9). The
pseudo-first-order rate constants obtained are k1 ) 450
s-1 and k2 ) 25 s-1. The acyl group is involved in an
insertion-deinsertion equilibrium (k3 ) 25 s-1). The 13C-
{1H} NMR shows the immediate appearance of the acyl
resonance of complex 3a after bubbling 13CO through a
solution of the 12CO rhodium-acyl complex at 253 K.
This proves the presence of an insertion-deinsertion
equilibrium. The rate of this insertion-deinsertion
reaction (Figure 9, k3) is much lower than the exchange
rate of the equatorial carbonyl ligand with carbon
monoxide in solution. Both the apical carbonyl and the
rhodium-acyl resonances are well resolved at 233 K,
indicating that exchange of both of these ligands with
dissolved 13CO is slow on the NMR time scale at this
temperature (Figure 6, spectrum 2a, simulated spec-
trum 2b). The resonance of the equatorial carbonyl
ligand is still very broad (k1 ) 175 s-1). At 223 K (Figure
6, spectrum 3a), the exchange of the equatorial carbonyl
ligand with dissolved 13CO is also slow on the NMR time
scale (k3 ) 20 s-1) and a sharp resonance was observed
for the equatorial carbonyl ligand together with a sharp
resonance for dissolved 13CO. The 13C{1H} NMR spec-
trum broadens again when the temperature was de-
creased further. This broadening is due to loss of
equivalence of the phosphorus atoms. We were not able
to resolve the different J CP coupling constants of the
inequivalent phosphorus atoms because the spectra
remained broad upon cooling to 193 K.
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 using standard Schlenk tech-
niques. All solvents were distilled from sodium. Ligand 1 was
prepared according to literature procedures.25 One-dimensional
(high-pressure) NMR spectra (1H, 31P, and 13C) were recorded
on a Bruker DRX-300 spectrometer. The two-dimensional
(high-pressure) NMR spectra were recorded on a Bruker AMX-
200 spectrometer. The NMR flow cell experiments were
recorded on a Bruker AM 200 WB spectrometer using a probe
developed in Liverpool.12 Chemical shifts are given in ppm
referenced to TMS or H3PO4 (external). The in situ IR
experiments were recorded on a Bio-Rad FTS-60A spectro-
photometer. Every 1 s, seven IR spectra were recorded that
were averaged to one spectrum. The IR spectra for the
stoichiometric reactions were recorded on a Nicolet 510 FT-
IR spectrometer.
Hyd r ofor m yla tion Exp er im en ts. These were performed
in a stainless steel (SS 316) autoclave (196 mL). The autoclave
is stirred mechanically and equipped with a reservoir, a
pressure transducer, a thermocouple, and a sampling device.
The autoclave is kept under constant pressure by a second
pressurized stainless steel autoclave equipped with a reducing
valve. In a typical experiment Rh(acac)(CO)2 and ligand 1 were
dissolved in 15 mL of toluene and introduced to the autoclave.
After the autoclave was flushed with CO/H2 (1/1), the autoclave
was put under a pressure of 20 bar. The autoclave was heated
to 41 °C, and after 2 h the substrate solution was charged into
the reservoir and added to the reaction mixture by overpres-
sure. The alkene was filtered over neutral alumina to remove
peroxides. During the reaction, several samples were taken
and immediately quenched by adding an excess of P(O-n-Bu)3,
to deactivate hydroformylation or isomerization active rhodium
species. The samples were analyzed by GC using decane as
internal standard. In a typical deuterioformylation experi-
ments, 10 mg (3.8 µmol) of Rh(acac)(CO)2 and 10 equiv of
ligand 1 were dissolved in 15 mL of toluene and introduced to
the autoclave. The autoclave was pressurized with 15 bar of
CO/D2 (1/2). The autoclave was heated to 41 °C, and after 2 h
the substrate solution was charged into the reservoir and
added to the reaction mixture by overpressure of CO. The
alkene was filtered over neutral alumina to remove peroxides.
The autoclave was pressurized up to 20 bar using CO. During
the reaction several samples were taken and quenched im-
mediately by adding an excess of P(O-n-Bu)3, to deactivate
hydroformylation- or isomerization-active rhodium species.
The samples were analyzed by GC using decane as internal
standard. The deuterium contents in the substrate and
products during the reaction were determined using gas
Con clu d in g Rem a r k s
While both kinetics and in situ studies seem rather
complicated at first sight, the system presents an ideal
case to illustrate the basics of hydroformylation. The
two most common types of kinetics are observed, and
the effects of raising or lowering the concentration of
any of the substrates leads to the result expected. In
most ligand-modified rhodium systems the oxidative
addition of H2 plays no role in the kinetics. As shown
previously,25 the present ligands have high ø valuess
higher than those of most phosphitessand a relatively
slow oxidative addition may indeed be expected. The
rate-determining step of the hydroformylation reaction
using this catalyst system cannot be attributed to one
single step and is strongly dependent on the reaction
conditions used. Therefore, the coordinatively saturated
rhodium-hydride complex 2a and the coordinatively
saturated rhodium-acyl complexes 3a -d are present
in solution during the hydroformylation reaction, which
was proven by in situ HP IR and HP NMR studies.
The deuterioformylation experiments showed that the
rhodium-alkyl complex 2c is irreversibly formed after
alkene coordination and hydride migration. Therefore,

Hydroformylation of 1-Alkenes
Organometallics, Vol. 20, No. 3, 2001 441
chromatography/mass spectrometry. After 2 h, complete con-
version was reached and the reaction mixture was quenched
with P(O-n-Bu)3. The aldehydes were distilled off and further
purified by column chromatography on silica gel. As eluent
pentane was used to remove toluene and decane from the
sample (Rf(aldehydes) ) 0), dichloromethane was used to
recover the mixture of linear and branched aldehydes (Rf-
reservoir. A 0.5 mL (3.2 mmol) amount of 1-octene was added
by overpressure. The IR spectra were recorded at 40 °C.
Recording was started several seconds before the alkene was
added.
High -P r essu r e NMR Exp er im en ts. In a typical NMR
tube experiment, 20 mg (0.079 mmol) of Rh(acac)(CO)2 and
114 mg (0.39 mmol) of ligand 1 were dissolved in 2.0 mL of
toluene-d8. The yellow solution was introduced to an argon-
flushed (10 mm) sapphire HP NMR tube31 and put under a
pressure of 15 bar of CO/H2 (1/1). Complete conversion to 2a
was reached after 1 h of heating at 80 °C, and the NMR tube
was depressurized. CO was bubbled through the solution, and
0.3 mL (2.4 mmol) of 1-hexene was added. The pale yellow
solution becomes bright yellow when 1-hexene is added. The
HP tube was pressurized with 3 bar of CO, and the NMR
spectra were recorded. A 10 mm thick-walled glass tube
instead of the sapphire HP tube31 was used for the two-
dimensional NMR experiments to obtain better resolution. In
a typical NMR flow cell12 experiment, 20 mg (0.079 mmol) of
Rh(acac)(CO)2 and 114 mg (0.39 mmol) of ligand 1 were
dissolved in 5 mL of toluene-d8. The cell was pressurized to
20 bar of CO/H2 (1/2). The sample was cooled to room
temperature after 1 h of heating at 80 °C. 1-Hexene was
injected at -20 °C, and the NMR spectra were recorded.
Simulations of the carbon and phosphorus NMR data were
performed using gNMR V4.1.0.
1
2
(aldehydes) ) 1). H and H NMR data of the aldehydes were
identical with the data obtained by Casey et al.29
High -P r essu r e F T IR Exp er im en ts. These were per-
formed in a stainless steel (SS 316) 50 mL autoclave equipped
with IRTRAN windows (ZnS, transparent up to 700 cm-1, 10
mm i.d., optical path length 0.4 mm), a mechanical stirrer, a
temperature controller, and a pressure transducer. In a typical
experiment 10 mg (0.039 mmol) of Rh(acac)(CO)2 and 114 mg
(0.39 mmol) of ligand 1 were dissolved in 15 mL of cyclohexane.
The yellow solution was introduced to the autoclave and put
under a pressure of 15 bar of CO/H2 (1/1). Complete conversion
to HRhL2(CO)2 (2a ) is reached after 1 h of heating at 80 °C or
2 h of heating at 40 °C. For the stoichiometric reactions, after
complete conversion to the hydride complex, the autoclave was
cooled to room temperature. After depressurization the auto-
clave was flushed with CO to remove the H2, and 1.0 mL (6.4
mmol) of 1-octene was added slowly. The autoclave is put
under a pressure of 5 bar of CO, and the IR spectra were
recorded at room temperature. Finally 10 bar of H2 was added,
and the reaction was followed using IR spectroscopy. In the
in situ IR experiments the autoclave is equipped with an extra
OM000785H