Origin of pressure effects on regioselectivity and enantioselectivity in the rhodium-catalyzed hydroformylation of styrene with (S, S, S)-bisdiazaphos

Article abstract of DOI:10.1021/ja909619a

Gas pressure influences the regioselectivity and enantioselectivity of aryl alkene hydroformylation as catalyzed by rhodium complexes of the BisDiazaphos ligand. Deuterioformylation of styrene at 80 °C results in extensive deuterium incorporation into the terminal position of the recovered styrene. This result establishes that rhodium hydride addition to form a branched alkyl rhodium occurs reversibly. The independent effect of carbon monoxide and hydrogen partial pressures on regioselectivity and enantioselectivity were measured. From 40 to 120 psi, both regioisomer (b:l) and enantiomer (R:S) ratios are proportional to the carbon monoxide partial pressure but approximately independent of the hydrogen pressure. The absolute rate for linear aldehyde formation was found to be inhibited by carbon monoxide pressure, whereas the rate for branched aldehyde formation is independent of CO pressure up to 80 psi; above 80 psi one observes the onset of inhibition. The carbon monoxide dependence of the rate and enantioselectivity for branched aldehyde indicates that the rate of production of (S)-2-phenyl propanal is inhibited by CO pressure, while the formation rate of the major enantiomer, (R)-2-phenyl propanal, is approximately independent of CO pressure. Hydroformylation of α-deuteriostyrene at 80 °C followed by conversion to (S)-2-benzyl-4-nitrobutanal reveals that 83% of the 2-phenylpropanal resulted from rhodium hydride addition to the re face of styrene, and 83% of the 3-phenylpropanal resulted from rhodium hydride addition to the si face of styrene. On the basis of these results, kinetic and steric/electronic models for the determination of regioselectivity and enantioselectivity are proposed.

Full text of DOI:10.1021/ja909619a

Published on Web 07/12/2010
Origin of Pressure Effects on Regioselectivity and
Enantioselectivity in the Rhodium-Catalyzed Hydroformylation
of Styrene with (S,S,S)-BisDiazaphos
Avery L. Watkins and Clark R. Landis*
Department of Chemistry, UniVersity of Wisconsin-Madison, 1101 UniVersity AVenue,
Madison, Wisconsin 53706
Received November 19, 2009; E-mail: landis@wisc.edu
Abstract: Gas pressure influences the regioselectivity and enantioselectivity of aryl alkene hydroformylation
as catalyzed by rhodium complexes of the BisDiazaphos ligand. Deuterioformylation of styrene at 80 °C
results in extensive deuterium incorporation into the terminal position of the recovered styrene. This result
establishes that rhodium hydride addition to form a branched alkyl rhodium occurs reversibly. The
independent effect of carbon monoxide and hydrogen partial pressures on regioselectivity and enantiose-
lectivity were measured. From 40 to 120 psi, both regioisomer (b:l) and enantiomer (R:S) ratios are
proportional to the carbon monoxide partial pressure but approximately independent of the hydrogen
pressure. The absolute rate for linear aldehyde formation was found to be inhibited by carbon monoxide
pressure, whereas the rate for branched aldehyde formation is independent of CO pressure up to 80 psi;
above 80 psi one observes the onset of inhibition. The carbon monoxide dependence of the rate and
enantioselectivity for branched aldehyde indicates that the rate of production of (S)-2-phenyl propanal is
inhibited by CO pressure, while the formation rate of the major enantiomer, (R)-2-phenyl propanal, is
approximately independent of CO pressure. Hydroformylation of R-deuteriostyrene at 80 °C followed by
conversion to (S)-2-benzyl-4-nitrobutanal reveals that 83% of the 2-phenylpropanal resulted from rhodium
hydride addition to the re face of styrene, and 83% of the 3-phenylpropanal resulted from rhodium hydride
addition to the si face of styrene. On the basis of these results, kinetic and steric/electronic models for the
determination of regioselectivity and enantioselectivity are proposed.
Introduction
catalyst rates and large turnover numbers, particularly when
expensive metals and chiral ligands are used. In view of these
Asymmetric hydroformylation (AHF) is an efficient, low-
cost route to chiral aldehydes from prochiral alkenes.1 Because
aldehydes are such versatile synthetic intermediates, AHF could
be a key transformation in the production of chiral pharmaceu-
ticals and agrochemicals.2 Aryl alkenes are attractive substrates
for AHF primarily because oxidation of the branched aldehyde
to the corresponding 2-arylpropionic acid yields pharmacologi-
cally active, anti-inflammatory analgesics such as ibuprofen,
ketoprofen, and naproxen.
challenges, only a small number of catalyst systems capable of
achieving useful productivity of a single enantiomer with >90%
ee have been reported for rhodium-catalyzed AHF.3 Among the
successful ligands reported, BINAPHOS and (S,S,S)-BisDiaza-
phos (1) ligands have been demonstrated to exhibit selectivity
for structurally diverse alkenes.4,5 Using 1 and rhodium catalyst
precursors, we have found that AHF of terminal aryl alkenes
exhibits complex behavior as a function of the gas partial
pressures.6 Increased CO pressure yields higher regioselectivity
and enantioselectivity but suppresses the rate. In contrast,
changes in H2 pressure have little influence. We seek a better
understanding of the origin of regioselectivity and enantiose-
lectivity in AHF.
However, practical application of AHF faces several chal-
lenges. Production of a single branched aldehyde with high
enantiomeric purity requires precise control of chemo-, enantio-,
and regioselectivity. Large-scale application requires high
(3) For successful ligands for AHF, see: (a) Babin, J. E.; Whiteker, G. T.
World Patent WO 9393839, 1993. (b) Sakai, N.; Mano, S.; Nozaki,
K.; Takaya, H. J. Am. Chem. Soc. 1993, 115, 7033–7034. (c) Dieguez,
M.; Pamies, O.; Ruiz, A.; Claver, C. New J. Chem. 2002, 26, 827–
833. (d) Cobley, C. J.; Gardner, K.; Klosin, J.; Praquin, C.; Hill, C.;
Whiteker, G. T.; Zanotti-Gerosa, A.; Petersen, J. L.; Abboud, K. A.
J. Org. Chem. 2004, 69, 4031–4040. (e) Cobley, C. J.; Klosin, J.;
Qin, C.; Whiteker, G. T. Org. Lett. 2004, 6, 3277–3280. (f) Clark,
T. P.; Landis, C. R.; Freed, S. L.; Klosin, J.; Abboud, K. A. J. Am.
Chem. Soc. 2005, 127, 5040–5042. (g) Yan, Y. J.; Zhang, X. M. J. Am.
Chem. Soc. 2006, 128, 7198–7202. (h) Yan, Y. J.; Zhang, X. M. J. Am.
Chem. Soc. 2006, 128, 7198–7202. (i) Axtell, A. T.; Klosin, J.;
Abboud, K. A. Organometallics 2006, 25, 5003–5009.
(1) For recent reviews of AHF, see: (a) Agbossou, F.; Carpentier, J.-F.;
Mortreux, A. Chem. ReV. 1995, 95, 2485–2506. (b) Claver, C.; van
Leeuwen, P. W. N. M. Rhodium Catalyzed Hydroformylation; Kluwer
Academic Publishers: Dordrecht, 2000. (c) Breit, B.; Wolfgang, S.
Synthesis 2001, 1, 1–36. (d) Dieguez, M.; Pamies, O.; Claver, C.
Tetrahedron: Asymmetry 2004, 15, 2113–2122. (e) Dieguez, M.;
Pamies, O.; Claver, C. Top. Organomet. Chem. 2006, 35–64.
(2) Botteghi, C.; Paganelli, S.; Marchetti, M. Chirality 1991, 3, 355–369.
9
10306 J. AM. CHEM. SOC. 2010, 132, 10306–10317
10.1021/ja909619a  2010 American Chemical Society

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
Scheme 1. General Mechanism for Rhodium-Catalyzed Hydroformylation
In this contribution, we present the methods, results, and
interpretations of mechanistic investigations of styrene hydro-
formylation using Rh(BisDiazaphos) catalysts. First, we report
investigations of AHF using D2 and CO; these studies probe
the reversibility of Rh-D addition to styrene. This is followed
by kinetic studies of the effect of CO and H2 pressures on rates
of formation of the achiral linear isomer and both branched
enantiomers. Data presentation concludes with isotopic labeling
experiments that probe the extent to which the linear product
results from reaction at the re and si faces of styrene. These
data are interpreted according to a simple kinetic model that
accommodates all the known data and constitutes a robust model
for the origins of selectivity and gas pressure effects in AHF of
styrene as catalyzed by rhodium complexes of 1.
the product regio- and stereoselectivity may be set at one of
four stages: (1) irreversible alkene coordination to rhodium, (2)
reversible alkene coordination followed by irreversible rhodium
hydride addition, (3) reversible rhodium alkyl formation fol-
lowed by irreversible alkyl migration to CO to form a rhodium
acyl complex, or (4) reversible rhodium acyl formation followed
by irreversible rhodium acyl hydrogenolysis.
For rhodium-based catalyst systems, the regiochemistry of
the aldehydes commonly is thought to be set by largely
irreversible hydride addition to coordinated alkene.8,9 The
structure of the five-coordinate π-olefin complex is proposed
(4) (a) Sakai, N.; Nozaki, K.; Takaya, H. J. Chem. Soc. Chem. Comm.
1994, 395–396. (b) Nanno, T.; Sakai, N.; Nozaki, K.; Takaya, H.
Tetrahedron: Asymmetry 1995, 6, 57. (c) Nozaki, K.; Li, W.; Horiuchi,
T.; Takaya, H. Tetrahedron Lett. 1997, 38, 4611–4614. (d) Horiuchi,
T.; Ohta, T.; Shirakawa, E.; Nozaki, K.; Takaya, H. Tetrahedron 1997,
53, 7795–7804. (e) Nozaki, K.; Sakai, N.; Nanno, T.; Higashijima,
T.; Mano, S.; Horiuchi, T.; Takaya, H. J. Am. Chem. Soc. 1997, 119,
4413–4423. (f) Nozaki, K.; Takaya, H.; Hiyama, T. Top. Catal. 1997,
4, 175–185. (g) Nozaki, K.; Matsuo, T.; Shibahara, F.; Hiyama, T.
AdV.Synth. Catal. 2001, 343, 61–63. (h) Aghmiz, A.; Masdeu-Bulto,
A. M.; Claver, C.; Sinou, D. J. Mol. Catal. 2002, 184, 111–119.
(5) Clark, T. P.; Landis, C. R.; Freed, S. L.; Klosin, J.; Abboud, K. A.
J. Am. Chem. Soc. 2005, 127, 5040–5042.
(6) Watkins, A. L.; Hashiguchi, B. G.; Landis, C. R. Org. Lett. 2008, 10,
4553–4556.
(7) Heck, R. F.; Breslow, D. S. J. Am. Chem. Soc. 1961, 83, 4023–4027.
(8) (a) Lazzaroni, R.; Uccello-Barretta, G.; Benetti, M. Organometallics
1989, 8, 2323–2327. (b) Lazzaroni, R.; Settambolo, R.; Uccello-
Barretta, G. Organometallics 1995, 14, 4644–4650. (c) Casey, C. P.;
Petrovich, L. M. J. Am. Chem. Soc. 1995, 117, 6007–6014. (d)
Horiuchi, T.; Shirakawa, E.; Nozaki, K.; Takaya, H. Organometallics
1997, 16, 2981–2986.
Background
(9) van Leeuwen has reported an example of reversible hydride addition
in the hydroformylation of 1-hexene. See: van der Slot, S. C.; Duran,
J.; Luten, J.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Organo-
metallics 2002, 21, 3873–3833.
According to the generally accepted mechanism for hydro-
formylation, first proposed by Heck and Breslow7 (Scheme 1),
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10307

A R T I C L E S
Watkins and Landis
Scheme 2. Asymmetric Hydroformylation of Styrene under
to play a crucial role in controlling the regioselectivity of the
reaction.10 However, this π-olefin complex has not been
observed directly. Brown and Kent11 have shown that the PPh3-
modified L2Rh(CO)2H complex exists as a mixture of two
rapidly equilibration trigonal bipyramidal isomers in a diequa-
torial (ee) to equatorial-apical (ea) isomer ratio of 85:15. Studies
on electronically modified bisphosphine ligands have demon-
strated similar dynamic equilibria between ee and ea.12 Casey
and Whiteker13 developed the concept of the natural bite angle
as means of characterizing diphosphine ligands on the basis of
molecular mechanics calculations. The Casey group reported
that the regioselectivity of the rhodium-catalyzed hydroformy-
lation of 1-alkenes was dramatically affected by the bite angle
of bidentate bisphosphine ligands.14 The correlation between
regioselectivity and ligand natural bite angle was rationalized
on the basis of a change in the ratio of ee and ae isomers as a
function of ligand bite angle. However, subsequent work by
Van Leeuween suggests that, despite the correlation between
regioselectivity and natural bite angle, the chelation mode of
the diphosphine ligand is not the key parameter controlling the
regioselectivity.15
Standard Conditions
alkyl rhodium formation is reversible, the regio- and enanti-
oselectivity must also be controlled, at least in part, by a step
(or steps) occurring later in the catalytic cycle. Casey et al.19
found that the nature of the selectivity-determining steps for
Pt-catalyzed hydroformylation of styrene changed with tem-
perature and gas pressures. On the basis of deuterioformylation
studies, they showed that enantioselectivity was determined in
the hydride addition step at 40 °C but at 100 °C was controlled
in latter stages of the catalytic cycle. Understanding the effect
of CO and H2 pressures on the selectivity is critical to
determining the origin of selectivity in AHF.
Results
Hydroformylation of Styrene at 80 °C. Standard conditions
(Scheme 2) for these hydroformylation studies comprise styrene
at 2.9 M and Rh-BisDiazaphos catalyst (1) at 6.7 × 10-4 M in
toluene (1.5 mL total solution volume) at 80 °C with 80 psi
syngas (1:1 CO/H2) performed in a pressure bottle (total volume
ca. 50 mL) placed in an oil bath and fitted with a Teflon-coated
stir bar. At 90% conversion of styrene, chiral GC analysis
showed 3-phenylpropanal (2) and 2-phenyl propanal (3)
[branched:linear (b:l) ) 4.5:1, 71% ee 2-phenylpropanal (3(R))].
In separate experiments, the regioselectivity and enantioselec-
tivity were observed to remain constant throughout the reactions.
These results were in good agreement with earlier studies that
were performed on a larger scale.
Deuterioformylation of Styrene. The reversibility of rhodium
hydride addition to coordinated styrene was probed with
deuterioformylation experiments. If hydride addition were
irreversible, deuterium would be incorporated in the product
with one deuterium ꢀ to the carbonyl group and one in the
formyl group (Scheme 3), only. However, reversible hydride
addition could yield incorporation of deuterium in multiple sites
of both styrene and the aldehyde products. Quenching the
reaction at low substrate conversion (<10% conversion) mini-
mizes the possibility of multiple exchanges and the opportunity
for deuterium-exchanged styrene to undergo deuterioformyla-
tion. Under such conditions one expects the major deuterium-
containing products to be the dideuterio-aldehyde products
(linear and branched) and deuterium-exchanged styrene; ap-
pearance of deuterium in the R-position of styrene indicates
reversible formation of the linear Rh-alkyl, whereas reversible
formation of the branched alkyl is implicated by observation
of deuterium in the ꢀ (or terminal) position.
Reaction conditions dictate which step fixes the regioselec-
tivity of rhodium-catalyzed aryl alkene hydroformylation.16 For
styrene hydroformylation with unmodified rhodium catalysts,
rhodium hydride addition strongly favors production of the
branched isomer and is irreversible at low temperature reaction
conditions (25-80 °C, 10-20 atm syngas). However, at higher
temperatures (90-120 °C), rhodium hydride addition can
become reversible. Under these conditions regioselectivity is
controlled as combination of the alkene insertion step and
subsequent steps such as CO binding and acyl formation.
Compared with achiral catalyst systems, the mechanistic
details of rhodium-catalyzed asymmetric hydroformylation
reactions are poorly understood. By analogy to achiral systems,
the regiochemistry and stereochemistry of AHF should be set
by the hydride addition step. On this basis, the origin of
enantioselectivity in asymmetric hydroformylation has been
discussed with the simple model proposed by Pino and
Consiglo.17 However, Nozaki and co-workers found that in the
AHF of styrene with BINAPHOS, the hydride addition becomes
reversible when the syngas pressure is reduced from 10 to 1
atm.18 Concomitant with the onset of reversibility was a modest
decrease in both regioselectivity and enantioselectivity (from
9:1 b:l to 5:1 and from 94% to 89% ee). We also observe strong
gas pressure effects on the regio- and stereochemistry of aryl
alkene AHF using Rh(BisDiazaphos) catalysts (vide supra). If
(10) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.;
Gavney, J. A.; Powell, D. R. J. Am. Chem. Soc. 1992, 114, 5535–
5543.
(11) Brown, J. M.; Kent, A. G. J. Chem. Soc., Perkin Trans. 2 1987, 1597–
1607.
Deuterioformylation of styrene catalyzed by 1 was conducted
at 80 psi 1:1 CO/D2 at 80 °C. After 10 min of reaction time, 1H
NMR analysis showed that 5% of the styrene had been converted
to aldehydes with b:l ) 4.2 and 69% ee 3(R). Integration of 1H
and 2H spectra of the products revealed 100% incorporation of
deuterium in the formyl group of the aldehyde products (Scheme
4, A-D) but only 20% deuterium incorporation ꢀ to the
carbonyl groups (B and C). No deuterium was observed at the
position R to the carbonyl groups of either aldehyde regioisomer
(12) Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.;
Petrovich, L. M.; Matter, B. A.; Powell, D. R. J. Am. Chem. Soc.
1997, 119, 11817–11825.
(13) Casey, C. P.; Whiteker, G. T. Isr. J. Chem. 1990, 30, 299–304.
(14) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.;
Gavney, J. A.; Powell, D. R. J. Am. Chem. Soc. 1992, 114, 5535–
5543.
(15) 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, 11616–11626.
(16) Lazzaroni, R.; Raffaelli, A.; Settambolo, R.; Bertozzi, S.; Vitulli, G.
J. Mol. Catal. 1989, 50, 1–9.
(17) Consiglio, G.; Pino, P. Top. Curr. Chem. 1982, 105, 77–123.
(18) Nozaki, K.; Matsuo, T.; Shibahara, F.; Hiyama, T. Organometallics
2003, 22, 594–600.
(19) Casey, C. P.; Martins, S. C.; Fagan, M. A. J. Am. Chem. Soc. 2004,
126, 5585–5592.
9
10308 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
Scheme 3. Products Expected from Deuterioformylation of Styrene
Scheme 5. Pathways by Which d1-Aldehydes and ꢀ-d1-Styrene Are
Produced
faster than it converts to aldehdye. Much more deuterium shows
up in the recovered styrene than in the ꢀ-position of the
aldehydes (51% F compared to 9% (B + C)), suggesting that
when a branched rhodium alkyl reverts to a rhodium alkene
hydride, the styrene most frequently dissociates from rhodium.
Styrene complexation to rhodium is apparently rapid and
reversible as compared to rhodium hydride addition. Because
deuterium is not found in the internal R position of the unreacted
styrene, linear alkyl rhodium appears much more likely than
branched alkyl rhodium to progress to aldehyde without
competitive reversion to styrene. However, some caution is
required in interpreting this result because the methylene
hydrogens of the linear alkyl are diastereotopic and could
undergo ꢀ-hydride elimination stereospecifically such that
deuterium exchange does not result.
Scheme 4. Product Distribution Observed in the
Deuterioformylation of Styrene Performed under Standard
Conditionsa
Most of the aldehydes formed during deuterioformylation
under standard conditions have just one deuterium in the
product. Although all linear and branched formyl groups contain
deuterium, the ꢀ-positions of the aldehyde products are primarily
protio (80% (G + H)). The high level of protium in the
ꢀ-position of the aldehyde suggests a 4-fold higher steady-state
concentration of Rh-H species than Rh-D species. Although acyl
hydrogenolysis regenerates rhodium deuterides exclusively, the
deuterium is rapidly incorporated into the unreacted styrene by
a rapid hydride addition/ꢀ-hydride elimination sequence (see
Scheme 5).
Hydroformylation Selectivity Varies with CO Pressure. The
independent effects of CO and H2 partial pressure on the
regioselectivity and enantioselectivity of styrene hydroformy-
lation as catalyzed by 1 at 80 °C was investigated. One set of
experiments used 40 psi H2 pressure while varying CO pressure
systematically from 15 to 200 psi.
a Products outlined by boxes require reversible insertion of styrene into
the Rh-D bond. Percentages outside of parentheses indicate product yield
based on initial styrene; percentages within indicate fractional yields among
transformed styrene molecules.
(A and D). Of the recovered styrene, 5.9% contained deuterium
in the ꢀ vinyl positions (F). No deuterium was found at the
R-vinyl position of the recovered styrene (E). Thus, of the
approximately 11% of the initial styrene that was transformed,
48% yielded aldehydes and 52% gave ꢀ-deuteriostyrene. The
gas headspace above the reactor is sufficiently large (50 mL)
that the accumulated HD is less than 5% of the H2 present.
The extensive incorporation of deuterium in the recovered
styrene requires reversible formation of the branched rhodium
alkyl. Once formed, this deuterated alkyl rhodium complex
ꢀ-hydride eliminates to give ꢀ-deuteriostyrene about 7 times
faster than it is converted to aldehyde (51.2% F vs 7.2% C).
Taking labeling statistics (1/3 of reversions of branched rhodium
alkyl to styrene leave an isotopic trace) into account and
assuming no significant isotope effect, the branched alkyl
rhodium complex reverts to styrene approximately 21 times
As shown in Figure 1, both the branched:linear aldehyde ratio
(b:l) and the enantiomeric ratio (R:S) show a strong dependence
on the CO pressure. The experimental data for b:l and R:S ratios
roughly conform to a phenomenological rate expression of the
type shown in eq 1.
A[CO]
B + C[CO]
rate ratio )
(1)
Competitive trapping by CO results in a transition from
apparent first order dependency in the low CO pressure regime
to independence in the high CO pressure regime. Many scenarios
could give rise to his kinetic behavior. However, the conclusion
(from analysis of deuterioformylation results) that some revers-
ibility accompanies branched alkyl formation immediately
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10309

A R T I C L E S
Watkins and Landis
Scheme 6. Kinetic Scheme Used To Extract Phenomenological
Rate Constants (klinear, kbranched, kexchange) by Simulation of Product
Concentrations Observed for Deuterioformylation of Styrene As
Catalyzed by Rh(acac)(CO)2 in the Presence of 1 at Small
Conversions
Figure 1. Graphs depicting the influence of CO and H2 pressure on the
enantiomeric ratio (R:S ) [3(R)]/[3(S)]) and branched:linear aldehyde ratio
(b:l ) [3]/[2]) for styrene hydroformylation catalyzed by 1 at 80 °C in
toluene. Smoothed trend lines are visual aids only.
suggests a role of CO in the competition between ꢀ-hydride
elimination from a branched alkyl rhodium species and conver-
sion to aldehyde product. For example, coordination of CO to
a branched alkyl rhodium species may promote acyl formation
and, hence, production of branched aldehyde. Because increased
CO pressure influences the enantiomeric ratio also, it appears
that the reaction manifold that leads to 3(R) is most affected.
Thus, raising the CO pressure could promote formation of 3(R)
over both 2 and 3(S), which would simultaneously increase
reaction regioselectivity and enantioselectivity.
Rh-D changes during the initial rate period as H is scrambled
from styrene to the catalyst under D2.
Hydroformylation Selectivity Is Independent of H2 Pres-
sure. To independently measure the effect of H2 partial pressure
on regio- and enantioselectivity, the catalytic hydroformylation
of styrene in the presence of Rh-BisDiazaphos (1) at 80 °C was
performed under different H2 pressures. The CO pressure was
held at constant 40 psi, and the H2 partial pressure was varied
from 15 to 160 psi. As shown in Figure 1, neither the
regioselectivity (b:l) nor enantioselectivity (R:S) varied ap-
preciably as the H2 pressure was changed. Therefore, both the
regiochemistry and stereochemistry of the reaction are likely
set prior to acyl hydrogenolysis.
Rates of Product Formation and Deuterium Exchange
Vary with CO Pressure. The data provided so far address the
relative but not absolute rates of 2, 3(R), and 3(S) formation.
Absolute rates were probed by deuterioformylation experiments
in which the D2 pressure is held constant at 40 psi, and the CO
pressure was varied systematically from 40 to 160 psi. Com-
parison with experiments performed under H2 revealed no
measurable deuterium isotope effect on either the b:l or R:S
ratios. Simulation of the initial deuterioformylation rates at <10%
conversion according to a simple kinetic model (Scheme 6)
using the kinetic modeling software package COPASI20 leads
to an excellent fit to experimental data for the appearance of
dideuterio and monodeuterio isotopologues of linear and
branched aldehydes and for monodeuterio styrene using just
three unique, phenomenological rate constants (all kinetic
isotope effects are assumed to be small). Because the CO and
D2 pressures are constant during each experiment, each step in
the kinetic model has a rate given by the product of the styrene
concentration, catalyst concentration, and rate constant. For
example, the following rate equation describes deuterium
incorporation into styrene, the last step in the kinetic model:
δ[d1-styrene]/δt ) kexchange[Rh-D][styrene]. Kinetic modeling is
required because the distribution of catalyst between Rh-H and
Data for the effect of CO pressure on the observed rate
constants for formation of 2 and 3 and d1-styrene are shown in
Figure 2. The rate constant for H/D exchange is greater than
the respective rates for both 2 and 3, even at high CO pressure.
The rates of linear aldehyde 2 formation and H/D exchange
into unlabeled styrene are inhibited by increasing CO pressure.
Inhibition by CO plausibly results from the need to dissociate
CO from the catalyst resting state, HRh(CO)2L, in order to
generate HRh(CO)L, which is capable of reacting with styrene.
The effect of PCO on the rate of branched aldehyde formation
is more complex. In the low pressure region, the rate is
unaffected by increasing PCO. However, at higher pressures, the
rate appears to be slightly inhibited by increasing PCO
.
Separation of the rates for branched aldehyde formation into
rates for each enantiomer (see Figure 2) provides further insight
into the influence of pressure on regioselectivity and enanti-
oselectivity. The phenomenological rate constant kbranched com-
prises the sum of the rate constants, kbranched(R) + kbranched(S), for
the two enantiomeric products. As shown in Figure 2, plots of
apparent rate constants as a function of CO pressure indicate
that the rate for minor enantiomer 3(S) is inhibited by CO
pressure, whereas formation of 3(R) is independent of CO,
perhaps exhibiting the onset of inhibition at higher pressure.
Hydroformylation Rates Exhibit Linear Dependence on
Catalyst and Alkene Concentration. The effect of catalyst and
alkene concentrations on hydroformylation rates is depicted in
Figure 3. A plot of the initial rate of aldehyde formation versus
catalyst concentrations (0.2 to 2 mM) indicates a first-order
dependence on catalyst concentration. The reaction order with
respect to alkene was determined by monitoring the entire
reaction time course using in situ infrared spectroscopy. As
shown in Figure 3, the formation of aldehyde exhibits a well-
behaved first-order dependence on alkene concentration.
Hydroformylation of r-Deuteriostyrene Reveals the Regi-
oselectivity for Different Styrene Enantiofaces. Under the
hydroformylation conditions examined thus far, rhodium hydride
adds preferentially to the re face of complexed styrene to form
(20) Hoops, S.; Sahle, S.; Gauges, R.; Lee, C.; Pahle, J.; Simus, N.; Singhal,
M.; Xu, L.; Mendes, P.; Kummer, U. Bioinformatics 2006, 22, 3067–
3074.
9
10310 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
Figure 2. Graphs depicting the influence of CO pressure on the observed rate constants for formation of d1-styrene (kexchange, upper left plot), 3-phenylpropanal
and 2-phenylpropanal (klinear and kbranched, upper right), and the partitioning of the branched rate constant into (R)-2-phenylpropanal and (S)-2-phenylpropanal
components (kbranched(R) and kbranched(S), lower left). All reactions performed at 80 °C, [styrene]0 ) 2.9 M, [1] ) 6.7 × 10-4 M, PH ) 40 psi. Smoothed lines
2
serve as guides only.
Figure 3. Effects of catalyst (left) and styrene concentrations (right) on reaction rate for reactions performed at 80 °C, 80 psi syngas, [styrene]0 ) 2.9 M.
The left plot shows initial rates measured at varying concentrations of catalyst. The right-hand plot represents the IR absorbance due to aldehyde products
as a function of time as monitored at 1725 cm-1 and the line computed for a simple first-order dependence on styrene.
the branched rhodium alkyl that ultimately produces the major
aldehyde product 3(R). Although the other enantiomer, 3(S),
arises from binding the si face of styrene to the catalyst, it is
not clear if the linear isomer originates primarily from the si or
re face of the styrene-Rh adduct. To examine the stereochemistry
of 3-phenylpropanal formation, the hydroformylation of R-deu-
teriostyrene was carried out and the stereochemistry of the
deuterated 3-phenylpropanal was determined (Scheme 7). We
note that an alternative strategy, the hydroformylation of
ꢀ-deuteriostyrene, was precluded by the rapid exchange of H
and D in the ꢀ-position during the hydroformylation (vide
supra).
the styrene had been converted to 2 and 3 with b:l ) 4.3 and
63% ee 3(R). The stereochemistry of 2 was determined by
conversion to the diastereomeric (2)-benzyl-4-nitrobutanal 6,
using the enantioselective organocatalytic Michael addition
protocol developed by Gellman and co-workers.21 The ratio of
2
diastereomers 6a and 6b was determined by H NMR integra-
tions of the diastereotopic benzylic methylene protons (PhHDC)
resonances, which appear at 3.17 and 2.8 ppm, respectively.
The 5:1 ratio of these resonances gives 63% de, which is similar
to 63% ee. Molecular mechanics modeling of conformational
1
distributions and experimental H NMR NOEs establish that
Hydroformylation of R-deuteriostyrene in the presence of 1
at 80 °C under 80 psi 1:1 CO/H2 for 6 h showed that 99% of
(21) Chi, Y.; Guo, L.; Kopf, N. A.; Gellman, S. H. J. Am. Chem. Soc.
2008, 130, 5608–5609.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10311

A R T I C L E S
Watkins and Landis
Scheme 7. Determination of Linear Product (3-Phenylpropanal) Stereochemistry Obtained by Hydroformylation of R-d1-Styrene
Scheme 8. Stereochemical Outcome of the Rhodium-Catalyzed Hydroformylation of R-d1-Styrene in the Presence of 1
Scheme 9. Proposed Kinetic Model for Determination of Regioselectivity in AHF of Styrene
(2S,3S)-6b forms in excess. Thus, 2 is generated primarily from
hydride addition to the si face of coordinated styrene.22
The stereochemistry of aldehyde formation under the conditions
used for the hydroformylation of R-deuteriostyrene is summarized
in Scheme 8. The branched isomer, 2-phenylpropanal, constitutes
81% of the aldehydes; 83% of the branched aldehyde is formed
by rhodium hydride addition to the re face of styrene. In contrast,
3-phenylpropanal constitutes 19% of the aldehydes, and 84% of it
is formed by rhodium hydride addition to the si face of styrene.
Overall, 70% of the aldehyde products result from rhodium hydride
addition to the re face of styrene. While rhodium hydride addition
to the re face of styrene exhibits excellent selectivity toward
branched aldehydes (22:1 b:l), rhodium hydride addition to the si
face of styrene is unselective, resulting in neapear aldehydes. Casey
and co-workers reported a similar selectivity trend for the platinum
catalyzed AHF of styrene.19
shown in Scheme 9. This model differs from the mechanism
shown in Scheme 1 by the addition of diastereomeric pathways
that produce enantiomers 3(R) and 3(S). Starting from 18 e-
complex A, loss of CO generates catalytically active B.
Coordination of styrene to B yields the diasteremeric alkene
complexes C(S) and C(R). From alkene complex C(S), in which
the alkene is coordinated at its si face, rhodium hydride can
then add to the coordinated styrene to give a linear alkyl rhodium
complex Dl or a branched alkyl rhodium complex D(S). These
rhodium alkyls are formed irreversibly and, hence, are com-
mitted to forming 2 and 3(S). Rhodium hydride addition from
C(R), in which styrene is coordinated at the re face, can lead
to either the linear rhodium alkyl Dl or the branched rhodium
alkyl D(R). Whereas the linear alkyl Dl is formed irreversibly,
branched D(R) undergoes reversion to styrene via ꢀ-hydride
elimination at a rate that is competitive, depending on the CO
pressure, with conversion to aldehyde. Thus, for the reaction
manifolds leading to 2 and 3(S), rhodium hydride addition fixes
the selectivity and limits the turnover frequency. Qualitatively,
inhibition of the flux by CO along these manifolds results from
shifting of the equilibrium A·B + CO toward A. For 3(R), the
turnover frequency is determined by the product of the frequency
of hydride addition to give Dl(R) multiplied by the fraction of
Discussion
Proposed Kinetic Model for Hydroformylation. On the basis
of the experimental kinetic and isotopic labeling data, we
propose the kinetic model for enantioselective hydroformylation
(22) The absolute stereochemistry at C-2 has been previously assigned as
the S-enantiomer (see ref 21).
9
10312 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
additions that are trapped to give 3(R); this fraction is
proportional to CO pressure. Qualitatively, the rate of formation
of 3(R) is independent of CO pressure due to a cancellation of
the inhibitory term (due to shifting of the equilibrium A·B +
CO) by the trapping efficiency term (which increases with CO
pressure until it reaches near unity).
A full rate expression for the proposed kinetic model can be
derived straightforwardly by application of the steady state
approximation on all rhodium-containing species. Because of
the many steps and intermediates, it is useful to simplify the
treatment with some reasonable approximations. The rate
expressions of eqs 2-5 result from the assumptions that (1)
the equilibrium between A and B is rapid as compared to
hydride addition, (2) equilibration of B with C(R) and C(S) is
rapid, (3) trapping of the alkyl intermediates Dl and Db(S), but
not Db(R), is rapid at all CO pressures, (4) the concentration
of Db(R) is described well by the steady-state approximation,
and (5) nearly all of the catalyst pools in the form of A, i.e.,
[RhTOT] ≈ [A]. With these assumptions, the rate expressions
for 2, 3(R), and 3(S) adopt the forms shown in eqs 2-4.
KCO(KSty(R)k3l(R) + KSty(S)k3l(S))[RhTOT][styrene]
d[2]
dt
rate2 )
)
PCO
(2)
(3)
K
COKSty(S)k3b(S)[RhTOT][styrene]
d[3(S)]
rate3(S)
)
)
dt
PCO
K
COKSty(R)k3b(R)k4b(R)[RhTOT][styrene]
k-3b(R) + k4b(R)PCO
d[3(R)]
dt
rate3(R)
)
)
(4)
Figure 4. Plots illustrating phenomenological rate constants klinear(upper
left), kbranched(R)(upper middle), kbranched(S) (middle left), kexchange ) kd -styrene
(middle right), excess selectivites (enantiomeric excess, ee, and “bran1ched”
excess, be; lower left), and the fitted values (eqs 8-11, solid and dashed
lines) versus CO partial pressure at 80 °C, 40 psi H2 partial pressure, [Rh]TOT
) 6.7 × 10-4 M, [styrene]0 ) 2.9 M. All rates measured over the first 10%
consumption of styrene.
d[D-styrene]
rateD-styrene
)
)
dt
1/3KCOKSty(R)k3b(R)k-3b(R)[RhTOT][styrene]
PCO(k-3b(R) + k4b(R)PCO
)
(5)
constant (kexchange) the one parameter fit was equivalent to
the two parameter fit: the data for kexchange can be described
most simply as inverse first-order in PCO. The empirical value
of kbranched(R) is modeled marginally better with a two
parameter fit. Thus, a good fit to all of the data can be
obtained with four or five parameters obtained from a sensible
kinetic model.
These rate expressions agree with the kinetic data. For
example, the net rates of formation of 2 and 3(S) exhibit
inhibition by CO due to the CO-dependent initial pre-
equilibrium between A, CO, and B. For 3(R), trapping of
Db(R) is inefficient at low CO pressure (k-3b(R) > k4b(R)[CO]);
this causes the rate of its formation to be approximately
independent of CO. Upon raising the CO pressure, trapping
of Db(R) by CO becomes competitive with ꢀ-hydride
elimination (k4b(R)[CO] ≈ k-3b(R)) and the rate of 3(R)
formation should be inhibited by CO. In the limit of high
CO pressure, the rates of formation of all products are
inhibited by CO and the regioselectivity and enantioselectivity
reach limiting values. The data are insufficient to uniquely
determine all nine constants in eqs 2-5. Therefore, the data
were fit to the simpler expressions of eqs 6-9
KCO(KSty(R)k3l(R) + KSty(S)k3l(S)
)
2.59
klinear
)
≈
(6)
(7)
PCO
PCO
KCOKSty(S)k3b(S)
1.70
PCO
kbranched(S)
)
≈
PCO
Global fitting of the four phenomenological rate constants
(klinear, kbranched(R), kbranched(S), and kexchange), which were measured
over the CO pressure range of 40-160 psi, and the observed
3:2 and 3(R):3(S) ratios, which were measured over the CO
pressure range of 40-400 psi, yield the results shown in eqs
6-9 and Figure 4. For kbranched(R) and kexchange, both one and
two parameter fits were performed. Effectively the two
parameter fit tests the sensitivity to a two-term denominator
(a + bPCO) versus a single term. For the exchange rate
K
COKSty(R)k3b(R)k4b(R)
kbranched(R)
)
k-3b(R) + k4b(R)PCO
0.27
1 + 0.0034PCO
≈
(two parameter fit)
(8)
K
COKSty(R)k3b(R)k4b(R)
kbranched(R)
)
k-3b(R) + k4b(R)PCO
≈ 0.27 (one parameter fit)
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10313

A R T I C L E S
Watkins and Landis
1/3KCOKSty(R)k3b(R)k-3b(R)
efficiency to yield a net rate of 3(R) formation that is
independent of PCO. Because the rate of formation of 3(S) and
2 is inhibited by CO pressure under all observed conditions,
the overall enantioselectivity and regioselectivity increase with
increasing CO in the low pressure regime.
kexchange
)
PCO(k-3b(R) + k4b(R)PCO
)
66
≈
(two parameter fit)
PCO(1 + 0.0PCO
)
(9)
1/3KCOKSty(R)k3b(R)k-3b(R)
It is common to identify individual steps in the catalytic cycle
as “enantiodetermining”, “regiodetermining”, or “turnover-
limiting”. However, such language is too simplistic to encom-
pass the kinetic behavior observed for enantioselective styrene
hydroformylation over a broad range of CO pressures. In the
limit of high pressure, ratios of product enantiomers and
regioisomers are controlled by the relative transition state free
energies for insertion of styrene into Rh-H. In this limit, which
corresponds to Curtin-Hammett conditions with respect to
alkene coordination, one can reasonably speak of insertion as
selectivity-determining and turnover-limiting. However, at lower
pressures the situation is more complex: intermediates Dl and
Db(S) are committed to form 2 and 3(S) but Db(R) is not. The
fate of Db(R) is controlled by a kinetic competition between
transition states (for deinsertion to give C(R) versus CO
association and insertion to give aldehyde product 3(R)) that
are similar in free energy.
Enantiofacial Discrimination Largely Controls Regioselec-
tivity. Hydride addition to the si enantioface of styrene is
nonregioselective, providing a 1:l mixture of branched and linear
aldehydes. In contrast, hydride addition to the re enantioface
of styrene displays excellent regioselectivity (22:1 b:l). Thus,
the enantiofacial selectivity of styrene binding strongly impacts
both enantioselection and regioselection. This observation
suggests that changes in catalyst design that promote discrimi-
nation in binding the re versus the si enantioface of styrene to
the catalyst could result in greater regioselectivity and enanti-
oselectivity.
Qualitative Free Energy Surface Depicting the Influence
of CO Pressure on Enantioselectivity. The CO pressure strongly
influences the hydroformylation selectivities of aryl alkenes such
as styrene but not other alkenes. Similarly, deuterioformylation
is more likely to result in D exchange into the alkene for aryl
alkenes than for other substrates.18 When styrene does undergo
D exchange, it does so primarily at the ꢀ-position via the
benzylic Rh-alkyls (Db(R) and Db(S)). These data suggest the
possible involvement of η3-coordination modes. For example,
van Leeuwen23 has proposed for nonenantioselective hydro-
formylation of styrene that the initial unsaturation at Rh resulting
from alkene insertion into the Rh-H bond is “released by
coordination in an η3-fashion.” This coordination mode inhibits
CO association, slows conversion to an acyl intermediate, and
the “Rh-η3-1-phenethyl remains in a state that may easily
undergo deinsertion.” In this model, η3-coordination decelerates
formation of the acyl intermediate. However, our data indicate
that alkyl Db(R) is formed fastest and quasi-reversibly while
yielding most of the aldehyde product; stabilization of the
intermediate Db(R) by η3-coordination should slow aldehyde
production along this path relative to other paths.
kexchange
)
PCO(k-3b(R) + k4b(R)PCO
)
66
PCO
≈
(one parameter fit)
Although the kinetic data are insufficient to uniquely deter-
mine the nine independent rate constants of the model depicted
in Scheme 9 and eqs 2-5, with reasonable assumptions these
equations do accommodate all the observed rate, selectivity, and
exchange data under all conditions. For example, assuming
equilibrium constant values for KCO, KSty(R), and KSty(S) that are
compatible with the observation of rates that are inhibited by
CO and first order in styrene, an adequate fit to all of the
observed data may be obtained by varying k3l(R), k3l(S), k3b(R)
,
k3b(S), k4b(R), and k-3b(R) (see the Supporting Information for
details). Although these rate constants are not uniquely deter-
mined, the foregoing demonstrates that the proposed kinetic
model accommodates all of the observed data. It must be noted,
however, that other, so far unimagined, kinetic scenarios might
equally well account for these data.
CO Pressure Effects Originate from Kinetic Competition
along the Manifold That Produces 3(R). The proposed kinetic
model features reversible formation of the fastest formed
rhodium alkyl, Db(R), and irreversible formation of rhodium
alkyls Db(S) and Dl. The branched:linear and enantiomer ratios
are given by eqs 10 and 11. In the limit of very high CO
pressure, trapping of Db(R) becomes faster than the ꢀ-hydride
elimination (k4b(R)[CO] . k-3b(R)). In this pressure regime, the
favored enantiomer and regioisomer are determined by the
equilibrium concentration of the diastereomeric alkene bound
complexes (KSty(R)/KSty(S)) and the relative rates at which they
undergo hydride addition (k3b(R)/k3b(S)). The enantiomeric ratio
and b:l ratio (see eqs 10 and 11) therefore approach indepen-
dence with respect to the CO pressure as the pressure is raised.
PCOKSty(R)k3b(R)k4b(R)
KSty(S)k3b(S)
+
k-3b(R) + k4b(R)PCO
KSty(R)k3l(R) + KSty(S)k3l(S)
b
l
rate3(S) + rate3(R)
)
)
rate2
(10)
PCOKSty(R)k3b(R)k4b(R)
k-3b(R) + k4b(R)PCO
KSty(S)k3b(S)
3(R)
3(S)
rate3(R)
rate3(S)
)
)
)
k3b(R) KSty(R)
PCO
k4b(R)
·
·
·
(11)
k3b(S) KSty(S) k-3b(R) + k4b(R)PCO
In the low CO pressure limit, the branched rhodium alkyl,
Db(R), undergoes ꢀ-hydride elimination faster than it is
converted to the corresponding acyl (k4b(R)[CO] , k-3b(R)). The
rate of formation of 3(R), is determined by the rate of formation
of Db(R), which is inversely dependent on PCO, multiplied by
An alternative, simpler explanation for the high susceptibility
of the alkyl Db(R) to revert to Rh-H and alkene and for the
influence of CO pressure on regio- and enantioselectivitiy is
that the barrier for styrene insertion to giVe Db(R) is uniquely
low. This is depicted in the hypothetical free energy surfaces
for low and high CO pressure conditions shown in Figure 5.
a trapping efficiency factor equal to k4b(R)PCO/(k4b(R)PCO
+
k-3b(R)). Under conditions of low pressure the trapping efficiency
of Db(R) increases linearly with increasing CO pressure. In this
pressure regime, the inhibitory effect of CO on the rate of
formation of Db(R) cancels the promoting effect on trapping
(23) van Rooy, A.; Orij, E. N.; Kramer, P. C. J.; van Leeuwen, P. W. N. M.
Organometallics 1995, 13, 34–43.
9
10314 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
Figure 6. Empirical steric-electronic map that rationalizes observed
regioselectivity and enantioselectivity patterns for the hydroformylation of
monosubstituted alkenes using Rh(BisDiazaphos) catalysts.
Although the empirical data from this kinetic study are
insufficient to determine all of the intermediate and transition
state free energies, the surfaces graphically summarize the
primary features of the kinetic behavior. Salient features include
(1) dissociation of CO from A to give B is uphill and reversible,
(2) styrene binding to B at either enantioface is energetically
unfavorable with respect to the resting state A and rapidly
reversible, (3) at low CO pressure, formation of Db(S) is
irreversible which leads to an enantiomeric ratio that is
determined by the free energy difference between the transition
state for acyl formation on the S-branched pathway and the
insertion transition state for alkyl formation along the R-
branched, and (4) at high pressure of CO, conversion of C(S)
to give Db(S) is irreversible and the enantiomeric ratio is
determined by the relative free energies of the styrene insertion
transition states, leading to an increase in enantioselectivity.
Figure 5. Hypothetical free energy surfaces for low CO pressure (a) and
high CO pressure (b) conditions that illustrate the relative increase in rate
of S-branched aldehyde (red, solid) versus the R-branched aldehyde (blue,
dashed) with increasing CO pressure. Free energy surfaces for the formation
of linear products are not shown explicitly but closely track the S-branched
(red, solid) surface. This figure provides a graphical depiction of the kinetic
behavior only; no additional information such as computed energies are
represented.
Free energy surfaces for the formation of linear products are
not shown explicitly but closely track the S-branched (red, solid)
surface. Note that these surfaces do not represent the full
catalytic cycle and end at the last kinetically relevant intermedi-
ates, acyl complexes. At low CO pressure, the branched alkyl
Db(R) reverts to the styrene adduct because the reversion barrier
is lower than the barrier for association of CO and migratory
insertion to form an Rh-acyl. Increasing the CO pressure
increases the free energy of states having two free CO’s to a
greater extent than those with one free CO. Thus, at high CO
pressure, the barrier of deinsertion becomes higher than as-
sociation of CO and migratory insertion to form an acyl
complex. Under these conditions, insertion of styrene to give
the kinetically preferred, branched alkyl Db(R) becomes ir-
reversible and the observed regioselectivity reached its maxi-
mum value. Because the two diastereomeric pathways for
making linear aldehydes approximately track the S-branched
surface, then the increase in enantioselectivity with increasing
CO pressure has the same origin as the enantioselectivity
increase. Accordingly, if there is a role of η3-coordination in
this model, it is to lower the energy of the transition state that
connects C(R) and Db(R).
Qualitative Steric-Electronic Map for Understanding Rela-
tive Rates of Insertion of Styrene into Rh-H Bonds. The
dominant source of enantioselectivity and regioselectivity in
styrene hydroformylation as catalyzed by Rh(BisDiazaphos)
complexes is the kinetic preference for styrene insertion to give
the branched alkyl Db(R). Previous studies by our group have
demonstrated a strong Hammett-like correlation for which more
electron-withdrawing para subsituents significantly increase the
branched:linear ratio but have no systematic effect on the
enantiomeric ratio.6 Substrate electronic factors control, to a
first approximation, regioselectivity but not enantioselectivity.
Consistent with the trend observed for para-substituted styrenes,
we generally find that monosubstituted alkenes undergo hydro-
formylation with higher branched:linear ratios as the inductive
effect of the substituent becomes more electron withdrawing
(for example, allyl cyanide vs 1-decene). However, the general
sense of chiral induction favors addition of the formyl to the re
face of the alkene for all 1-alkenes examined by us so far.
Enantioselectivity is primarily controlled by steric interactions
in the diastereomeric transition states for styrene insertion into
the Rh-H bond.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10315

A R T I C L E S
Watkins and Landis
Figure 6 depicts a qualitative steric-electronic map that
summarizes these findings. We presume that the pseudo-five
coordinate transition state structure for alkene insertion favors
axial-equatorial coordination of the diphosphine ligand. A two-
dimensional quadrant steric map is constructed such that the
plane of the map contains the Rh-H vector and lies parallel to
the plane of alkene CdC. Because the axial P atom lies in the
plane of the map and the equatorial P atom lies behind the plane,
the steric influence of the axial P substituents dominates. The
equatorial CO ligand is small and exerts minimal steric effect,
leading to a quadrant coloring with two sterically small
quadrants (uncolored, right-hand side), one quadrant that is
sterically large (bottom left, dark blue), and one that lies between
the two extremes (top left, light blue). Electronic effects can be
superimposed by coloring the top two quadrants (red dots),
indicating that orientation of an alkene substituent in the
quadrants that give rise to linear products is disfavored. The strength
of the electronic effect varies with the inductive effect of the X
substituent. As shown in Figure 6, this map rationalizes that the
most favored transition state corresponds to the branched-Re
orientation while the least-favored transition state corresponds to
the linear re orientation. The free energies of the linear and
branched si face orientations lie between these extremes.
regio- and enantioselection can be elucidated for a useful
enantioselective hydroformylation catalyst.
Experimental Methods
Tetrahydrofuran (THF) and toluene were distilled over Na/
benzophenone and further deoxygenated by at least three freeze-thaw
cycles prior to use. Rh(acac)(CO)2 was recrystallized from toluene/
hexanes (green needles) prior to use. Dodecane and styrene were
purchased from Aldrich, distilled under vacuum, and deoxygenated
prior to use. (R)-Diphenylprolinol silyl ether was purchased from
Aldrich and used as received. R-Deuteriostyrene and nitroethylene
were prepared according to literature procedures.19,24 1H NMR and
2H NMR (76.77 MHz) NMR data were recorded using a Varian
Unity 500 MHz spectrometer. Gas chromatography (GC) was
performed on a Varian Chrompack instrument using a ꢀ-DEX 120
column from Supelco, (30 m × 0.25 mm i.d.). All reaction pressures
are reported as absolute pressures in pounds per square inch (psi;
not gauge pressures, psig).
General Procedure for Asymmetric Hydroformylation.
Ligand solutions (THF) and Rh(acac)(CO)2 (toluene) were
prepared at room temperature in a nitrogen-filled drybox at room
temperature. Hydroformylation solutions were prepared by
addition of ligand (1 µmol) and Rh(acac)(CO)2 (0.5 µmol) stock
solutions to an oven-dried 50 mL pressure bottle equipped with
a magnetic stir bar. Styrene (4.4 mmol) containing dodecane
(0.4 mmol) was added, and the total volume was adjusted to
1.5 mL by addition of toluene. The pressure bottle was then
connected to a gauged pressure reactor assembly and removed
from the drybox. The pressure reactor was then purged three
times with syngas (1:1 CO/H2) and then charged to 150 psi.
The reaction bottle was then placed in the oil bath with vigorous
stirring and allowed to react for 2 h. After 2 h, the reaction was
removed from the oil bath, cooled in an ice bath, and vented.
General Procedure for Deuterioformylation Kinetic Stud-
ies. Deuterioformylation solutions were prepared by addition of
ligand (1 µmol) and Rh(acac)(CO)2 (0.5 µmol) stock solutions to
an oven-dried 50 mL pressure bottle equipped with a magnetic stir
bar. The total volume of each reaction was adjusted to 0.9 mL by
addition of toluene. The pressure bottle was then connected onto a
gauged pressure reactor assembly, and the reactor was removed
from the drybox. The reactor was purged three times with CO,
charged with 40 psi CO followed by 40 psi D2, and allowed to stir
in an 80 °C oil bath for 30 min. After 30 min, styrene (4.4 mmol)
containing dodecane internal standard (0.4 mmol) was added via
gastight syringe. Samples for NMR analysis were removed after 5
and 11 min via gastight syringe.
High Pressure IR Spectroscopy. A solution of Rh(acac)(CO)2
(1 µmol), 1(1.2 µ mol), styrene (5 mmol), and toluene (3.5 mL)
was placed in a high-pressure reaction vessel fitted with an in situ
attenuated total reflectance IR detector of a ReactIR apparatus. The
pressure vessel was sealed under nitrogen and removed from the
glovebox. The autoclave was heated to 80 °C followed by
pressurization with syngas (1:1 CO/H2), with rapid stirring. IR
spectra were recorded every 2 min. Recording was started 30 s
before addition of syngas. The appearance of aldehydes was
monitored as the change in intensity of the absorption peak centered
at 1725 cm-1 corresponding to the overlapping carbonyl stretching
frequencies of the branched and linear aldehydes.
Organocatalytic Michael Addition. Organocatalytic Michael
reactions were performed in a manner similar to that reported
by Gellman and co-workers.21 To a 12 mL vial equipped with
a small magnetic stir bar were added 1.2 mL of dry toluene,
(R)-diphenylprolinol silyl ether (0.4 mL stock solution in toluene,
[catalyst] ) 0.05M), 0.2 mmol 3-nitrobenzioc acid B (33.4 mg),
and the crude hydroformylation products (approx 2.0 mmol neat).
Are the qualitative ideas presented by Figures 5 and 6 convenient
fantasy or an essential representation of the free energy landscape?
Insight requires further experimental data and/or computer model-
ing. On the experimental side, trapping of key intermediates and
direct measurements of their relative reactivities would provide
critical information. Full computational modeling of the reaction
free energy surface, although a daunting task, would enable
verification of these hypothetical models.
Summary
Rhodium catalysts bearing the chiral BisDiazaphos ligand
exhibit extraordinary rates and regio- and enantioselectivities
for the hydroformylation of alkenes. Aryl alkenes, such as
styrene, yield branched aldehydes with high enantiomeric purity
under optimized conditions, but the regioselectivity and enan-
tioselectivity erode as the syngas pressure is lowered. It is the
CO partial pressure, not that of dihydrogen, that influences b:l
and enantiomer ratios. Deuterioformylation studies demonstrate
that formation of the branched rhodium alkyl that ultimately
yields the major enantiomeric aldehyde is reversible, whereas
other diastereomeric branched and linear rhodium alkyls are
formed irreversibly. Thus, the pressure effect on regio- and
enantioselectivity arises from a kinetic competition between CO-
dependent conversion of one branched rhodium alkyl diastere-
omer to an acyl and its reversion to a rhodium hydride and
styrene. Studies with deuterium-labeled styrene establish that
coordination of re versus si enantiofaces of styrene leads to very
different branched:linear ratios. Models of the reaction free
energy surfaces under different CO pressures and steric-
electronic mapping of the transition state for insertion of bound
styrene into the Rh-H bond provide a graphical summary of
the key experimental findings. By exhibiting strong CO pressure
effects on regio- and enantioselectivity, aryl alkenes appear to
be the exception rather than the rule. It appears that the origin
of these effects is the unusually low insertion barrier for
formation of a single branched alkyl diastereomer vis-a`-vis the
free energy of its CO-dependent pathway for acyl formation.
These studies constitute a rare example in which the details of
(24) Wright, J. P.; Gaunt, M. J.; Spencer, J. P. Chem.-Eur. J. 2006, 12,
949–955. See the Supporting Information for detailed procedures.
9
10316 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Pressure Effects in the Hydroformylation of Styrene
A R T I C L E S
The mixture was stirred in an ice bath for 5 min, followed by
addition of 1.0 mmol nitroethylene (0.2 mL stock toluene
solution). The mixture was stirred in a cold room (3 °C) for 4
days. After 4 days, toluene was removed under reduced pressure,
and the product purified via SiO2 column chromatography eluting
with EtOAc (20%) in hexanes to give the (S)-2-benzyl-4-
NSF CHE-0342998, NIH 1 S10 RR13866-01, NSF CHE-9629688,
and NSF CHE-9208463.
Supporting Information Available: Details of NMR and GC
analyses of hydroformylation and deuterioformylation products,
tabulation of all kinetic and product distribution data (Table S1),
phenomenological rate constants and errors determined by fitting
observed product distributions to the simple kinetic model of
Scheme 6 (Table S2), details of the preparation of R-d1-styrene
and the analysis of linear hydroformylation products derived
therefrom, results of fitting rate constants to experimental data
according to eqs 6-9, and tabulation of observed data and values
calculated from a set of reasonable, assumed kinetic constants. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
nitrobutanal. H NMR (500 MHz, C6D6): δ 1.46 (m, 1H), 1.70
(m, 1H), 2.08 (X, 1H), 2.15 (dd, J ) 7.5, 14 Hz, 1H), 2.42 (dd,
J ) 7.5, 14 Hz, 1H), 3.64-3.68 (AB, 2H), 7.0-7.21 (m, 5H),9.0
(s, 1H).
Acknowledgment. This work was supported by the National
Science Foundation (CHE-0715491). We thank Professor Charles
P. Casey for helpful discussions concerning the mechanistic aspect
of this work. Dr. Tulay Atesin performed the molecular mechanics
simulation of rotamer populations used in analyzing the enantio-
facial origins of enantioselective hydroformylation. NMR instru-
mentation used in this work was supported by the following grants:
JA909619A
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10317