Origin of enantioselection in chiral alcohol oxidation catalyzed by Pd[(-)-sparteine]Cl2

Article abstract of DOI:10.1021/ja053195p

A kinetic investigation into the origin of enantioselectivity for the Pd[(-)-sparteine]Cl₂-catalyzed aerobic oxidative kinetic resolution (OKR) is reported. A mechanism to account for a newly discovered chloride dissociation from Pd[(-)-sparteine]Cl₂ prior to alcohol binding is proposed. The mechanism includes (1) chloride dissociation from Pd[(-)-sparteine]Cl₂ to form cationic Pd(-)-sparteine]Cl, (2) alcohol binding, (3) deprotonation of Pd-bound alcohol to form a Pd-alkoxide, and (4) β-hydride elimination of Pd-alkoxide to form ketone product and a Pd-hydride. Utilizing the addition of (-)-sparteine HCl to control the [Cl ^-] and [H⁺] and the resulting derived rate law, the key microscopic kinetic and thermodynamic constants were extracted for each enantiomer of sec-phenethyl alcohol. These constants allow for the successful simulation of the oxidation rate in the presence of exogenous (-)-sparteine HCl. A rate law for oxidation of the racemic alcohol was derived that allows for the successful prediction of the experimentally measured k_rel values when using the extracted constants. Besides a factor of 10 difference between the relative rates of β-hydride elimination for the enantiomers, the main enhancement in enantiodetermination results from a concentration effect of (-)-sparteine HCl and the relative rates of reprotonation of the diastereomeric Pd-alkoxides.

Full text of DOI:10.1021/ja053195p

Published on Web 09/29/2005
Origin of Enantioselection in Chiral Alcohol Oxidation
Catalyzed by Pd[(-)-sparteine]Cl
2
Jaime A. Mueller, Anne Cowell, Bert D. Chandler,* and Matthew S. Sigman*,†
†
‡
,‡
Contribution from the Departments of Chemistry, UniVersity of Utah, 315 South 1400 East,
Salt Lake City, Utah 84112, and Trinity UniVersity, One Trinity Place,
San Antonio, Texas 78212-7200
Received May 16, 2005; E-mail: bert.chandler@trinity.edu; sigman@chem.utah.edu
Abstract: A kinetic investigation into the origin of enantioselectivity for the Pd[(-)-sparteine]Cl
aerobic oxidative kinetic resolution (OKR) is reported. A mechanism to account for a newly discovered
chloride dissociation from Pd[(-)-sparteine]Cl prior to alcohol binding is proposed. The mechanism includes
1) chloride dissociation from Pd[(-)-sparteine]Cl to form cationic Pd(-)-sparteine]Cl, (2) alcohol binding,
2
-catalyzed
2
(
(
2
3) deprotonation of Pd-bound alcohol to form a Pd-alkoxide, and (4) â-hydride elimination of Pd-alkoxide
-
to form ketone product and a Pd-hydride. Utilizing the addition of (-)-sparteine HCl to control the [Cl ]
and [H ] and the resulting derived rate law, the key microscopic kinetic and thermodynamic constants
+
were extracted for each enantiomer of sec-phenethyl alcohol. These constants allow for the successful
simulation of the oxidation rate in the presence of exogenous (-)-sparteine HCl. A rate law for oxidation
of the racemic alcohol was derived that allows for the successful prediction of the experimentally measured
k
rel values when using the extracted constants. Besides a factor of 10 difference between the relative rates
of â-hydride elimination for the enantiomers, the main enhancement in enantiodetermination results from
a concentration effect of (-)-sparteine HCl and the relative rates of reprotonation of the diastereomeric
Pd-alkoxides.
Introduction
difference in the relative reaction rates of the respective
enantiomers to provide reasonable yields of enantiomerically
The field of asymmetric catalysis has grown tremendously
as a means to access enantiomerically enriched building blocks.
Accessing nonracemic alcohols, which are important as starting
materials, by asymmetric catalysis has garnered significant
attention from the synthetic community. A key strategy to
2
1
enriched materials. Many kinetic resolutions do not achieve
this goal. In 2001, we reported the Pd-catalyzed aerobic
oxidative kinetic resolution (OKR) of secondary alcohols using
(
-)-sparteine which, for selected substrates, gives krel values
2
2,23
of >20.
We sought to investigate the origin of enantiose-
accessing highly enantiomerically enriched alcohols has been
kinetic resolution of racemic alcohols.1-6 The nonenzymatic
lectivity in this catalyst system in order to expand the system’s
2
4,25
utility,
potentially improve catalyst design, and provide a
kinetic resolution of secondary alcohols has been accomplished
7
-9
deeper understanding for asymmetric catalyst design in general.
using a variety of methods including epoxidation,
acylation
with nucleophilic catalysts,1
0-15
and oxidation.
16-20
(
12) For a recent review of kinetic resolutions using nucleophilic catalysts, see:
Jarvo, E. R.; Miller, S. J. In ComprehensiVe Asymmetric Catalysis,
Supplement; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer:
Berlin, Germany, 2004; pp 189-206.
Considering that the maximum theoretical yield of a kinetic
resolution is 50%, useful methods require at least a 20-fold
(13) Suzuki, Y.; Yamauchi, K.; Muramatsu, K.; Sato, M. Chem. Commun. 2004,
*
Corresponding author.
Department of Chemistry, University of Utah.
Department of Chemistry, Trinity University.
2770-2771.
†
(
14) Vedejs, E.; Daugulis, O. J. Am. Chem. Soc 2003, 125, 4166-4173.
‡
(15) Dalaigh, C. O.; Hynes, S. J.; Maher, D. J.; Connon, S. J. Org. Biomol.
Chem. 2005, 3, 981-984
(
1) For a review on kinetic resolutions using nonenzymatic catalysts, see:
Robinson, D. E. J. E.; Bull, S. D. Tetrahedron: Asymmetry 2003, 14, 1407-
(16) Kinetic resolution via alcohol oxidation was covered in a review on selective
alcohol oxidations, see: Arterburn, J. B. Tetrahedron 2001, 57, 9765-
9788.
1
446.
(
2) Turner Nicholas, J. Curr. Opin. Chem. Biol 2004, 8, 114-119.
3) For a review on kinetic resolutions of secondary alcohols, see: Somfai, P.
In Organic Synthesis Highlights IV; Schmalz, H.-G., Ed.; Wiley-VCH:
Weinheim, Germany, 2000; pp 175-181.
(
(17) For a recent example of aerobic oxidative kinetic resolution of alcohols,
see: Radosevich, A. T.; Musich, C.; Toste, F. D. J. Am. Chem. Soc. 2005,
127, 1090-1091.
(
4) Somfai, P. Angew. Chem., Int. Ed. 1998, 36, 22731-2733.
5) For a review on metal-catalyzed kinetic resolutions, see: Hoveyda, A. H.;
Didiuk, M. T. Curr. Org. Chem. 1998, 2, 489-526.
(18) Irie, R.; Katsuki, T. Chem. Rec. 2004, 4, 96-109.
(19) Sun, W.; Wang, H.; Xia, C.; Li, J.; Zhao, P. Angew. Chem., Int. Ed. 2003,
42, 1042-1044.
(
(
6) Cook, G. R. Curr. Org. Chem. 2000, 4, 869-885.
(20) Nishibayashi, Y.; Yamauchi, A.; Onodera, G.; Uemura, S. J. Org. Chem.
2003, 68, 5875-5880.
(
7) Martin, V. S.; Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda, M.;
Sharpless, K. B. J. Am. Chem. Soc. 1981, 103, 6237-6240.
8) Adam, W.; Humpf, H.-U.; Roschmann, K. J.; Saha-Moeller, C. R. J. Org.
Chem. 2001, 66, 5796-5800.
(21) For an excellent discussion of the practical aspects of kinetic resolutions,
see: Keith, J. M.; Larrow, J. F.; Jacobsen, E. N. AdV. Synth. Catal. 2001,
343, 5-26.
(
(
9) Yang, D.; Jiao, G.-S.; Yip, Y.-C.; Lai, T.-H.; Wong, M.-K. J. Org. Chem.
(22) Jensen, D. R.; Pugsley, J. S.; Sigman, M. S. J. Am. Chem. Soc. 2001, 123,
7475-7476.
2
001, 66, 4619-4624.
(
(
10) Yamada, S.; Misono, T.; Iwai, Y. Tetrahedron Lett. 2005, 46, 2239-2242.
11) Santi, C.; Tiecco, M.; Testaferri, L.; Tomassini, C.; Marini, F.; Bagnoli,
L.; Temperini, A. Phosphorus, Sulfur, Silicon Relat. Elem. 2005, 180,
(23) (a) A closely related Pd/(-)-sparteine system for aerobic oxidative kinetic
resolution was reported simultaneously and independently, see: Ferreira,
E. M.; Stoltz, B. M. J. Am. Chem. Soc. 2001, 123, 7725-7726. (b) Also
see: Stoltz, B. M. Chem. Lett. 2004, 33, 362-367.
1
071-1075.
10.1021/ja053195p CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 14817-14824
9
14817

A R T I C L E S
Mueller et al.
Scheme 1. Original Proposed Mechanism for Pd-Catalyzed
Oxidation of Alcohols
Table 1. Summary of Kinetic Studies of Pd-Catalyzed Oxidative
Kinetic Resolution of Secondary Alcohols
50 mol %
4 mol %
parameter
[(
−)-sparteine]
[(
−)-sparteine]
[
[
[
[
Pd(-)-sparteineCl2]
sec-phenethyl alcohol]
(-)-sparteine]
first order
first order
saturated
inhibition
1.31 ( 0.04
F ) 1.41 ( 0.15
11
first order
first order
first order
(-)-sparteine HCl]
KIE
1.04 ( 0.06
no correlation
6.1
Hammett correlation
intrinsic krel value
racemate krel value
25 ( 5.0
7.6 ( 2.0
elimination to liberate the ketone product and form a Pd-
hydride, followed by fast regeneration of 1 using O2. Kinetic
experiments revealed an intriguing dependence on exogenous
Our previous kinetic investigations into this system deter-
mined several of the mechanistic details and offered our first
insights into the origin of enantioselection.26,27 Further efforts
have focused on developing a more precise kinetic and
enantioselective model for chiral alcohol oxidation. Of note, a
structural model of enantioselection based on isolated Pd-
[(-)-sparteine]: (1) It acts as a Brønsted base to deprotonate
the Pd-bound alcohol, forming (-)-sparteine HCl as a transient
intermediate. (2) It induces a saturation event as it increases
from 4 to 50 mol %. This implies a change in rate-determining
steps at different [(-)-sparteine]. At low [(-)-sparteine], no
observed Hammett correlation and a unity KIE value for
C-H(D) in sec-phenethyl alcohol suggests that alcohol depro-
tonation is rate determining. At high [(-)-sparteine], a KIE value
for C-H(D) in sec-phenethyl alcohol of 1.3 ( 0.04 and a
F-value of -1.4 ( 0.5 support a shift from deprotonation to
â-hydride elimination as the rate-limiting step.
alkoxides that cannot undergo â-hydride elimination has been
proposed by Trend and Stoltz.28 A key component of this
structural model invokes diastereoselective chloride ion substitu-
tion as an important influence on asymmetric induction.
Theoretical approaches have also been implemented to elucidate
the origin of enantioselectivity.29 After defining the most
energetically favored pathway, Nielsen and co-workers com-
pared krel values generated from transition-state calculations of
â-hydride elimination to experimental krel values. They con-
cluded that the enantioselectivity of Pd-catalyzed alcohol OKR
is primarily due to kinetic â-hydride elimination and is greatly
influenced by the nature of the Pd-catalyst counterion. A
limitation of their study is that deprotonation/reprotonation could
not be addressed due to the large number of atoms in the relevant
structures.
As a complement to the structural and theoretical studies,
we report herein a kinetic model of the enantiodifferentiation
between chiral alcohols during Pd-catalyzed aerobic oxidation
using (-)-sparteine. This model allows for the extraction of
thermodynamic and kinetic constants for each enantiomer of
sec-phenethyl alcohol and successfully predicts the experimental
krel values under synthetically relevant reaction conditions.
The OKR enantioselectivity for secondary alcohols is sub-
stantially influenced by [(-)-sparteine]. At 4 mol % (-)-
sparteine, the intrinsic k , calculated by comparing the measured
rel
single enantiomer oxidation rates, is similar to the racemate krel
measured during OKR experiments (Table 1). However, at 50
mol % (-)-sparteine, the racemate krel is more than double the
intrinsic krel of 11. Several mechanistic factors may influence
enantiodifferentiation, including the role of chloride ion and the
presence of (-)-sparteine HCl, a large, chiral Brønsted acid.
Enantiodifferentiation may also be affected by thermodynamic
differences between the diastereomeric alkoxide stabilities,
relative rates of â-hydride elimination, and competition between
alcohol enantiomers for a single chiral catalyst. While the change
in rate-limiting step at low [(-)-sparteine] to deprotonation is
interesting, the highest selectivities were garnered at high [(-)-
sparteine] conditions. Therefore, the addition of several key
experiments and the reconsideration of previous results have
been utilized to elucidate the thermodynamic and/or kinetic
factors that control asymmetric induction at high [(-)-sparteine].
Background
Previous kinetic studies support a mechanism involving
alcohol binding by Pd[(-)-sparteine]Cl2 (1), followed by
deprotonation of the Pd-bound alcohol to form a Pd alkoxide
Results
(
Scheme 1). The Pd-alkoxide species can undergo â-hydride
Single Enantiomer Experiments. Earlier studies demon-
strated that exogenous (-)-sparteine HCl inhibits the rate of
alcohol oxidation, providing additional evidence of rate-limiting
â-hydride elimination at high [(-)-sparteine]; however, these
results were initially interpreted in terms of increasing the rate
of alkoxide reprotonation. It is also possible that the additive
(-)-sparteine HCl inhibits the alcohol binding equilibrium, with
(24) Caspi, D. D.; Ebner, D. C.; Bagdanoff, J. T.; Stoltz, B. M. AdV. Synth.
Catal. 2004, 346, 185-189.
(
25) Ali, I. S.; Sudalai, A. Tetrahedron Lett. 2002, 43, 5435-5436.
26) Mueller, J. A.; Jensen, D. R.; Sigman, M. S. J. Am. Chem. Soc. 2002, 124,
(
8
202-8203.
(27) Mueller, J. A.; Sigman, M. S. J. Am. Chem. Soc. 2003, 125, 7005-7013.
(28) Trend, R. M.; Stoltz, B. M. J. Am. Chem. Soc. 2004, 126, 4482-4483.
(29) Nielsen, R. J.; Keith, J. M.; Stoltz, B. M.; Goddard, W. A., III. J. Am.
Chem. Soc. 2004, 126, 7967-7974.
14818 J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005

Enantioselection in Chiral Alcohol Oxidation
A R T I C L E S
Scheme 2. Revised Proposed Mechanism for Oxidation of
Alcohols with Pd[(-)-sparteine]Cl2
The Pd mass balance,
Pd] ) [1] + [3] + [4] + [5]
[
t
can be solved for [5] using the equilibrium constant expressions
for the chloride dissociation and alcohol binding equilibria (KD
and KR, respectively) and applying the steady-state approxima-
tion
+
0
) k [B][4] - k [5] - k [HB ][5]
2
R
3R
-2R
using [5] as the steady-state intermediate. The resulting rate law
(full derivation available in Supporting Information) is
-
ν ) K K k k [AH ][B][Pd] /{(K + [X ] + K K
R
D
R
2R 3R
R
t
D
D
R
the additional chloride ion driving the back reaction. To test
this, we attempted comparable inhibition experiments using
several other sources of exogeneous chloride ion, including
tetraalkylammonium chlorides and PPh3(CH2Ph)Cl. The rapid
formation of Pd-black or poor solubility of exogenous chloride
ion sources made the extraction of kinetic and thermodynamic
data from these experiments untenable.
+
[AH
])(k-2R[HB ] +k3R) + K K k2R[AH ][B]} (2)
D R R
R
where [AHR] is (R)-sec-phenethyl alcohol, [B] is (-)-sparteine,
-
+
[Pd]t is total Pd catalyst, [X ] is chloride ion, and [HB ] is
+
(-)-sparteine H . Additionally, the alcohol intermediate has
been unobservable, even in the presence of large excess of
Attempts to monitor alcohol binding and chloride dissociation
using conductivity have been more successful. Preparation of
the Pd[(-)-sparteine)]Cl2 complex in DCE resulted in a solution
with substantial conductivity. Individual solutions of (-)-spar-
teine and Pd(NCCH3)2Cl2 in DCE showed essentially no con-
ductivity prior to mixing, indicating that Pd[(-)-sparteine]Cl2
exhibits a dissociative pre-equilibrium in the absence of alcohol.
Comparison against a calibration curve prepared from PPh3(CH2-
Ph)Cl solutions allowed for the determination of the chloride
dissociation equilibrium constant (KD ) 1.9 × 10 M, see
Supporting Information). Based on the determined value for KD,
approximately 35% of the complex dissociates a chloride under
typical OKR reaction conditions (1 mM Pd[(-)-sparteine]Cl2).
Addition of up to 0.6 M racemic alcohol caused no increase
in conductivity, indicating that alcohol binding does not
substantially perturb the dissociative equilibrium. Attempts to
study the alcohol binding equilibrium with NMR spectroscopy
have also been unsuccessful. Similarly, alcohol dependence
studies with up to 0.4 M alcohol show first-order kinetics;
saturation in alcohol is not observed under any conditions. On
the basis of these results, we conclude that the equilibrium
constant for alcohol binding must be significantly smaller than
alcohol; hence
K [AH ] , 1
R
R
Using this conclusion, the rate law simplifies to
K k k [AH ][B][Pd]
R
2R 3R
R
t
νR )
-
[
X ]
+
1
+
(k-2R[HB ] + k ) + K k [AH ][B]
3R R 2R R
(
)
KD
-
4
(
3)
The rate law includes all of the key information supported
by the mechanistic studies. The second term in the denominator
allows for the observed base saturation behavior. Finally, the
revised rate law predicts a complicated inhibition by exogenous
(-)-sparteine HCl, which impacts both the [X ] and [HB ]
terms.
-
+
Dependence on Added [(-)-Sparteine HCl]. The derived
-
+
+
rate law includes chloride ion [X ] and sparteine H [HB ] in
the denominator, which corresponds to the role of (-)-sparteine
HCl as a reaction inhibitor. (-)-Sparteine HCl can both inhibit
the initial dissociative equilibrium and promote reprotonation
of the Pd-alkoxide, rendering a complex rate dependence on
[(-)-sparteine HCl]. This dependence can be elucidated by
examining the rate law with a 1/ν plot and assuming that, when
1.
Based on these findings, a slightly modified mechanism is
proposed for the Pd(II)-catalyzed oxidation of alcohols (Scheme
). The revised mechanism contains all of the essential features
+
2
excess (-)-sparteine HCl is added, the concentrations of HB
-
of the original mechanism, with the addition of a rapid pre-
equilibrium for the chloride ion dissociation from the Pd[(-)-
sparteine]Cl2 complex. The addition of a distinct chloride
dissociation step is a relatively minor modification to the original
mechanism; further, the modified mechanism is entirely con-
sistent with all of the original kinetic data.
and X are controlled by this amount. In other words, the
+
-
additional production of HB and X from the Pd starting
material’s dissociative equilibrium and the deprotonation of the
alcohol are negligible relative to the amount intentionally added.
This assumption introduces less than a 10% error in the assumed
-
[X ] when only 1 equiv of (-)-sparteine HCl is added (relative
Using the mechanism in Scheme 2 and an enantiomerically
pure alcohol substrate, a revised rate law can be derived.
Assuming that all steps after â-hydride elimination are fast, the
reaction rate of the (R)-enantiomer can be expressed in terms
of the â-hydride elimination step:
to [Pd[(-)-sparteine]Cl ]) at 1 mol % in these experiments.
Errors in [HB ] are smaller, and these systematic errors become
2
+
negligible at higher [(-)-sparteine HCl]. This allows for the
+
-
replacement of both the [HB ] and [X ] terms with the
concentration of added (-)-sparteine HCl ([SpHCl]) in eq 3:
+
-
ν ) k [5]
R
3R
[SpHCl] ) [HB ] ) [X ]
J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005 14819

A R T I C L E S
Mueller et al.
yielding the rate expression
νR )
K k k [AH ][B][Pd]
R
2R 3R
R
t
k-2R
k3R
2
[
SpHCl] + k
+
[SpHCl]+k +K k [AH ][B]
3R R 2R R
(
-2R
)
KD
KD
(4)
The concentration of [(-)-sparteine HCl] is then related to 1/νR
by the following quadratic equation:
k3R
k-2R + KD [SpHCl]
2
(
)
k-2R[SpHCl]
1
)
+
νR K K k k [AH ][B][Pd] K k k [AH ][B][Pd]
D
R
2R 3R
R
t
R
2R 3R
R
t
k3R + K k [AH ][B]
R
2R
R
+
(5)
K k k [AH ][B][Pd]
R
2R 3R
R
t
This further simplifies to
k3R
k-2R
+
[SpHCl]
2
(
)
[
B][Pd]t k-2R[SpHCl]
KD
)
+
kR-obs
K K k k
D
R
2R 3R
K k k
R
2R 3R
[AH ][B]
1
R
+
+
(6)
K k
k3R
R
2R
Figure 1. Quadratic fits of alcohol oxidation rate dependence on [SpHCl]:
(A) (R)-sec-phenethyl alcohol; (B) (S)-sec-phenethyl alcohol.
Oxidation rates for each enantiomer of sec-phenethyl alcohol
were measured under high (-)-sparteine conditions (where
â-hydride elimination is rate-limiting) in the presence of added
(
Table 2. Constants from Quadratic Nonlinear Least-Squares
Analyses
-)-sparteine HCl (eq 7).30 The data were plotted in the form
constant
(R)
(S)
equation
of quadratic eq 6. Nonlinear least-squares analysis of each
enantiomer yielded excellent fits of the measured data to the
k-2Y
2
4
AY )
AY
(6.0 ( 1.0) × 10
(1.2 ( 0.25) × 10
K K k2Yk
3Y
D
Y
quadratic model (Figure 1). The coefficients AY, BY, and CY
2
k3Y + k-2YKD
K K k2Yk
(
Y ) R or S) of f(x) ) AYx + BYx + CY were extracted from
B
Y
3.6 ( 1.2
52 ( 29
BY )
the nonlinear least-squares analysis for each alcohol enantiomer
D
Y
3Y
(Table 2).
k3Y + KYk2Y[AH ][B]
Y
C
Y
0.042 ( 0.003
0.51 ( 0.07
CY
)
K k2Yk
Y
3Y
Table 3. Extracted Rate Constants for Pd-Catalyzed
sec-Phenethyl Alcohol Oxidation
parameters
(R)
(S)
(R)/(S)
1
1
M
500 ( 500
106 ( 60
14
K k
)
Y
2Y
-2
-1
-2
-1
Extraction of Kinetic and Thermodynamic Constants.
K
D
(B
Y
- K
D
A
Y
)
min
M
min
Thermodynamic and kinetic rate constants (KYk2Y, k3Y, k-2Y)
for each enantiomer can be extracted from the quadratic fits to
the [SpHCl] dependence inhibition studies (Table 3). The alcohol
intermediate has been unobservable and therefore must be
present in minute concentrations. Consequently, the alcohol
binding and deprotonation steps can only be treated as a single
thermodynamic/kinetic event that is described by the combined
microscopic constants KYk2Y. Values for k3Y can also be
determined from previously published [(-)-sparteine] saturation
kinetics data using a Lineweaver-Burk plot (see Supporting
[
AH ][B]
Y
0
min
.125 ( 0.042
0.010 ( 0.006
12
k3Y
)
1
-1
-1
min
C
-
(
Y
K k
)
Y
2Y
2
M
1 ( 10
2.5 ( 2
8.4
k-2Y ) AYKDKYk2Yk3Y
-
1
-1
-1
-1
min
M
min
k-2Y
170 ( 98
M
240 ( 250
0.69
-
1
-1
k3Y
M
-
2
-1
in reasonable agreement with values determined from the
[SpHCl] dependence inhibition studies. Additionally, the Gibbs’
free energies of activation were calculated using the values for
k3Y determined from the [SpHCl] dependence inhibition studies
(∆GR ) 23.7 ( 0.2 kcal/mol and ∆GS ) 25.3 ( 0.3 kcal/mol).
These free energy values are in reasonable agreement with those
previously measured when no added (-)-sparteine HCl is used
Information). The values (k3R ) 8.7 × 10 min , k3S ) 8.5
10 min , k3R/ k3S ) 10) determined using this method are
-
3
-1
×
(
30) The oxidation rate of (R)-sec-phenethyl alcohol was measured by monitoring
the appearance of acetophenone by in situ IR spectroscopy. Oxidation of
(S)-sec-phenethyl alcohol was significantly slower, challenging the limita-
tions of that method of monitoring. Instead, the oxidation rate of the (S)-
enantiomer was monitored by GC. See Supporting Information for further
experimental details.
14820 J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005

Enantioselection in Chiral Alcohol Oxidation
A R T I C L E S
Figure 3. Simulated OKR krel values using enantioselective model in eq
11.
Scheme 3. Mechanism of Pd-Catalyzed OKR of sec-Phenethyl
Alcohol
Figure 2. Simulation of (A) (R)-sec-phenethyl alcohol oxidation rates and
(B) (S)-sec-phenethyl alcohol oxidation rates with additive [SpHCl].
(∆GR ) 22.5 ( 0.6 kcal/mol and ∆GS ) 23.8 ( 2.6 kcal/mol),
further supporting that the extracted k3Y values reflect the
experimental events.
Both the (R) and (S) inhibition studies were successfully
simulated using the extracted rate constants, the experimental
KD, and the rate law associated with the modified mechanism
(Figure 2).
Implications of the Intrinsic krel. The intrinsic krel value (the
relative observed rate constants for experiments with enantio-
merically pure alcohols) can be described by
K k k
R
2R 3R
krel
)
×
(
K k k
)
S
2S 3S
-
(
K + [X ])(k [SpHCl] + k ) + K K k [B][AH ]
D
-2S
3S
D
S
2S
S
(8)
(
-
)
(
K + [X ])(k [SpHCl] + k ) + K K k [B][AH ]
D
-2R
3R
D
R
2R
R
The data clearly show that oxidation of both alcohol enantiomers
is inhibited by exogenous (-)-sparteine HCl, but changes in
the intrinsic krel values are of greater practical importance. An
intrinsic krel model was generated using eq 8 and the extracted
parameters in Table 3. The model indicates that [(-)-sparteine
HCl] has a significant effect on the intrinsic krel. Values for the
intrinsic krel increase from approximately 12 to 20 over the
concentration range studied, corresponding to a relative rate
amplification of 1.7 (Figure 3).
to be the same as with the single enantiomers; however, an
important distinction must be included. The racemate rate law
must also describe the competition between alcohol enantiomers
for a single chiral Pd catalyst. Even though the individual
mechanistic steps remain unchanged, this competition substan-
tially complicates the reaction network and the rate law (Scheme
3).
Oxidative Kinetic Resolution in the Presence of Exogenous
(
-)-Sparteine HCl. Under high [(-)-sparteine] conditions, a
significant difference between measured intrinsic krel values (ca.
1) and racemate krel values (ca. 25) is observed. The individual
steps in the reaction mechanism for the racemate are assumed
1
The derivation of the racemate rate law is qualitatively similar
to the single enantiomer rate law derivation, with the exception
J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005 14821

A R T I C L E S
Mueller et al.
of a more complicated Pd mass balance. Assuming that all steps
after â-hydride elimination are fast, the reaction rates can be
expressed in terms of the â-hydride elimination step:
To test this enantioselective model, racemate krel values were
31
measured with additive (-)-sparteine HCl (eq 12). The model
successfully predicts the krel values of OKR of racemic sec-
phenethyl alcohol (Figure 3).
ν ) k [5-(R)]
R
3R
The Pd mass balance,
Pd] ) [5-(R)]+ [4-(R)] + [1] + [3] + [4-(S)] + [5-(S)]
[
t
Origin of Enantioselectivity. Of fundamental importance is
the origin of enantioselectivity for the system, particularly the
enantioselectivity enhancement observed using the racemate.
The original proposal for the Pd-catalyzed aerobic oxidation of
chiral alcohols was that enantioselectivity at high [(-)-sparteine]
was due to a combination of kinetic differences in â-hydride
elimination and a thermodynamic difference in Pd-alkoxide
stabilities favoring the (R)-enantiomer of alcohol. This proposal
does not explain our present observations. The current investiga-
tions indicate that relative thermodynamics play an unclear role
and that enantiodifferentiation in this system arises largely from
reactivity differences and the details of the reaction kinetics.
Both thermodynamic and kinetic factors can affect selectivity
in a catalytic system, although they do not necessarily work in
concert. In a seminal report on enantioselective hydrogenation
by Rh(diphos) complexes, Halpern described kinetic and
thermodynamic limiting scenarios in determining enantioselec-
tivity for asymmetric catalytic systems: (1) the prevailing
reactive species is determined by the initial substrate binding
event, and (2) the prevailing reactive species arises from the
minor catalyst-substrate diastereomer by virtue of its much
can be solved for [5-(R)] in the same fashion as for the single
enantiomer experiments, assuming that the chloride dissociation
and alcohol binding equilibria are fast (established) pre-
equilibria. Because both reactions occur simultaneously, both
[5-(R)] and [5-(S)] must be treated as steady-state intermediates,
described by the equations
+
0
) k [4-(R)][B] - k [5-(R)] - k [HB ][5-(R)]
2R 3R -2R
or
0
) k [4-(R)][B] - k3R*[5-(R)]
2R
+
where k3R* ) (k3R + k-2R[HB ] and
0
) k [4-(S)][B] - k3S*[5-(S)]
2S
+
where k3S* ) (k3S + k-2S[HB ]). Using these assumptions and
that KY is significantly less than 1, the following rate laws were
derived (see Supporting Information):
32
higher reactivity. In this study, Halpern and co-workers found
that the thermodynamic stability of intermediates is of secondary
importance to their intrinsic reactivity. Specifically, the minor
diastereomeric species undergoes a far more facile addition of
hydrogen than the major diastereomer, illustrating a selective
process that is purely kinetically controlled. This is an elegant
proof that the Curtin-Hammett principle must be considered
when dealing with selectivity in catalytic systems. Simply stated,
the Curtin-Hammett principle asserts that one cannot draw a
causal relationship between product ratios and the relative
νR )
K K k k [AH ][B][Pd]
t
D
R
2R 3R
R
k [B]
-
2S
K K k [AH ][B] + k [X ]+ K + K K [AH ]
D
R
2R
R
3R*
D
D
S
S
( k
)
[
]
3
S*
(
9)
νS )
K K k k [AH ][B][Pd]
t
D
S
2S 3S
S
3
3
proportions of intermediates in a fast equilibrium.
k [B]
-
2R
(a) Thermodynamic Influences. The small alcohol binding
constants make evaluation of thermodynamic influences in this
system difficult, at best. We have been unable to observe alcohol
binding to the Pd center and have treated this binding as a fast,
established pre-equilibrium. Consequently, it is not possible to
differentiate between thermodynamic and kinetic factors in the
relative binding rates of (R)- and (S)-sec-phenethyl alcohols.
Using the presently available data, it is only possible to point
out that the combination of binding constants (KY) and depro-
tonation rate constants (k2Y) results in an approximately 10-
fold greater net reactivity for (R)-sec-phenethyl alcohol.
For the OKR reaction, the magnitude of the k-2/k3 ratios ((R)
K K k [AH ][B] + k [X ]+ K + K K [AH ]
D
S
2S
S
3S*
D
D
R
R
( k
)
[
]
3
R*
(10)
An important consequence of this rate law is that the rate of
alcohol oxidation for each enantiomer now depends on the
oxidation rate for the other, reflecting the competition for the
Pd catalyst.3
4,35
The overall krel value can be expressed by
K k k
R
2R 3R
krel
)
×
(
)
K k k
S
2S 3S
-
1
-1
)
170 M , (S) ) 240 M , Table 3) suggests that the
K K k [AH ][B]
-
D
R
2R
R
reprotonation reactions are facile and the system may approach
K K k [AH ][B]+k [X ]+K +
D
S
2S
S
3S*
(
D
)
[
]
k3R*
(
31) The krel was calculated using krel ) ln[(1 - C)(1 - ee)]/ln[(1 - C)(1 +
ee)], where C is the conversion and ee is the enantiomeric excess, see:
Kagan, H. B.; Fiaud, J. C. Kinetic Resolution. Top. Stereochem. 1988, 18,
249-329.
K K k [AH ][B]
-
D
S
2S
S
(
)
K K k [AH ][B]+k [X ]+K +
D
R
2R
R
3R*
(
D
)
[
]
k3S*
(
32) Halpern, J. Science 1982, 217, 401-407.
(
11)
(33) Curtin, D. Y. Rec. Chem. Prog. 1954, 15, 111.
1
4822 J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005

Enantioselection in Chiral Alcohol Oxidation
A R T I C L E S
equilibrium. However, since the magnitude of k2Y (the depro-
tonation rate constant) cannot readily be determined, it is also
possible that the two parallel intermediates never establish a
true thermodynamic equilibrium. For clarification, we define
this thermodynamic equilibrium specifically as the ratio of Pd-
alkoxide intermediates that is governed by their free energies
of formation. The kinetic steady state, on the other hand, requires
only that the concentrations of intermediates remain constant
over time; it does not imply that those concentrations are
equivalent to expected concentrations based on thermodynamic
equilibria or the relative stabilities of the two Pd-alkoxides. It
is then possible that relative stabilities may play little or no
role in determining the product ratios in the catalytic system.
Consequently, for the OKR reaction, the influence of relative
Pd-alkoxide stabilities is not at all clear, not readily determin-
able, and may even have little consequence for the origin of
the enantioselectivity.
Enantioselectivity Enhancement in Racemate Experi-
ments. The increased enantiodifferentiation (relative to the
single enantiomer experiments) observed upon oxidation of the
racemate can only be reasonably attributed to kinetic effects.
In examining this assertion, the possibility that the mechanism
changes for the racemate must be considered. Along those lines,
it is possible that, for the racemate, the diastereomers interact
in some manner to change the intrinsic stabilities and/or
reactivities of the intermediates. Given the extremely low
concentrations of the active species and the quality of the fit to
the experimental data in Figure 3, these possibilities are unlikely
to be of consequence.
Rather, two key properties unique to the oxidative kinetic
resolution appear to be at work. First, the racemate experiments
contain an inherent competition between the alcohols for the
catalytically active Pd species. This competition is not present
in single enantiomer experiments, and this explanation relies
only on the clear experimental evidence that the Pd-(R)-
alkoxide is produced at a faster rate than the Pd-(S)-alkoxide.
When the two alcohols are in competition for the catalyst, this
kinetic enhancement can be thought of as “pulling” the
equilibrium farther toward the (R) reaction network, partially
(b) Kinetic Influences. The kinetic evidence indicates that
each of the mechanistic steps favors (R)-alcohol oxidation. The
primary adjustable parameters in this reaction are [(-)-sparteine]
+
and [HB ]; increasing the concentration of either reagent
enhances enantioselectivity in this system. For comparison
purposes, the “baseline” selectivity for this system is the intrinsic
selectivity under low [(-)-sparteine] conditions, where alcohol
deprotonation is rate determining and enantioselectivity is set
by the relative rates of deprotonation. The intrinsic krel measured
under these conditions shows deprotonation of the (R)-enanti-
omer to be favored by a factor of 6 (Table 1). The determined
value of KRk2R/KSk2S (14, Table 3) is consistent with this
observation (the ratio of the microscopic constants is larger than
the measured krel because kobs values also include several terms
in the denominators).
3
4,35
starving the (S) network.
It is difficult to quantify the
importance of this competition; however, its direct impact on
the increased enantioselectivity in racemate experiments is likely
to be small. The relatively large value for chloride dissociation
and small values for alcohol binding imply that a large excess
of the monochloro-Pd species 3 is available; hence, the
individual alcohol binding equilibria (and subsequent deproto-
nation and oxidation) are unlikely to substantially impact one
another.
The second property of this competition is more subtle, but
appears to drive the higher selectivities observed in racemate
over single enantiomer experiments. The assumption that rate
constants remain the same from the single enantiomer to
racemate conditions (vida supra) does not imply that intermedi-
ate concentrations remain constant. Hence, the absolute rates
of individual steps, which depend on both intermediate con-
centrations and rate constants, are not necessarily identical in
Subsequent mechanistic steps serve to enhance this baseline
selectivity. At higher [(-)-sparteine], the intrinsic krel (11, Table
1
) nearly doubles as the rate-determining step moves from Pd-
bound alcohol deprotonation to Pd-alkoxide â-hydride elimina-
tion. With this change in rate-limiting step, determined â-hydride
elimination rate constants (k3, Table 3) also show an approximate
factor of 10 difference between the (R)- and (S)-alkoxides. This
enhancement must be independent of the relative alkoxide
thermodynamic stabilities, as those terms can appear only in
steps leading up to alcohol deprotonation. Consequently, the
â-hydride elimination enhancement is purely kinetic in nature,
reflecting a greater reactivity of the (R)-bound alkoxide.
+
both experiments. In the OKR reaction, [HB ] appears to be
the key intermediate in explaining the observed enantiodiffer-
entiation enhancement in racemate experiments, which result
from a suppression of the (S)-enantiomer pathway and an
effective increase in krel.
In single enantiomer experiments, the Pd-(R)-alkoxide is
produced roughly an order of magnitude faster than the Pd-
Although the reprotonation reaction was proposed in the
original mechanism, this step appears to be far more important
than we initially realized, as experimentally determined intrinsic
krel values increase substantially with added (-)-sparteine HCl,
approaching 20 (Figure 3). For both bound alkoxides, the
determined values for k-2 are 2 orders of magnitude greater
than k3. Further, the k-2/k3 values ((R)/(S) ) 0.69, Table 3) are
also the only kinetic terms that are larger for the (S)- than for
the (R)-enantiomer. The net result is that, for this system, even
the reprotonation reaction works in concert with the other kinetic
steps to favor oxidation of the (R)-alkoxide. Additionally, it is
clear that reprotonation plays a critical role in determining the
concentration of Pd-alkoxide at high [(-)-sparteine], and likely
has an important role even when the added [(-)-sparteine] is
low.
(
S)-alkoxide (KRk2R/KSk2S ) 14, Table 3). Because these steps
+
in the reaction produce one equivalent of HB , it is reasonable
to conclude that, once steady state is established, the [HB ] in
+
the (R) single enantiomer experiments is approximately an order
+
of magnitude greater than the [HB ] in the (S) single enantiomer
experiments. However, a single steady state is established in
experiments with the racemate. Since the oxidation of the (R)-
+
alcohol is the dominant pathway, the [HB ] in racemate
experiments will be largely determined by (R)-alcohol oxidation
kinetics. For the (S)-enantiomer pathway in OKR, this results
+
in a roughly 10-fold increase in [HB ] relative to the single
+
enantiomer experiment. This increase in [HB ] drives the (S)-
(
34) Blackmond, D. G. J. Am. Chem. Soc. 2001, 123, 545-553.
(35) Andraos, J. J. Phys. Chem. A 2003, 107, 2374-2387.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005 14823

A R T I C L E S
Mueller et al.
alkoxide reprotonation reaction faster in OKR than in the (S)
single enantiomer experiment, resulting in the net slowing of
the (S) pathway and an observed increase in krel for the racemate.
This explanation is strongly supported by the kinetic models
and the (-)-sparteine HCl inhibition studies. As greater amounts
of exogenous (-)-sparteine HCl are added to the reaction, the
eomeric alkoxides being formed, depending on which chloride
dissociates from the original Pd[(-)-sparteine]Cl2. A good
pictorial model is proposed of three-dimensional interactions
that induce enantiodifferentiation during â-hydride elimination.
However, to achieve this, the experimental conditions under
which OKR occurs were avoided and could not be simulated
computationally.
+
catalytic cycle’s contribution to the overall [HB ] diminishes.
Once the catalytic cycle’s contribution becomes negligible, the
exogenous (-)-sparteine HCl sets the [HB ] in both the single
On many levels our kinetic model is consistent with that put
forth by Nielsen and co-workers. Both models agree that the
+
29
enantiomer and racemate experiments, and there should be little
difference between the intrinsic and racemate krel values. As
shown in Figure 3, the differences between intrinsic and
racemate krel values rapidly diminish with added (-)-sparteine
HCl, and the two curves approach a common krel of ap-
proximately 20, in good agreement with the experimental data.
The close approach of these two curves also supports the
conclusion that competition for the catalyst plays only a minor
role in the increased enantiodifferentiation during racemate
experiments.
kinetic contribution of â-hydride elimination is significant.
However, our model differs from the computational model of
Nielsen and co-workers in that they arrive at the conclusion
that the enantioselectivity is exclusively limited to â-hydride
elimination. This could be due to the incomplete representation
of all the steps involved due to computational limitations. Our
kinetic studies complement their computational studies, as we
were able to gain insight into the deprotonation/reprotonation
steps that eluded calculation.
Conclusion
Relevance to Other Models and Systems. The kinetic
enantioselective model is a good complement to the current
A mathematical model for the enantioselective oxidation of
chiral alcohols using Pd[(-)-sparteine]Cl2 has been elucidated.
It successfully predicts experimental krel values where the
2
8
29
structural and computational models. These studies have
provided a detailed, albeit complex mathematical picture of how
each component of the reaction interacts with the others to define
the overall enantioselectivity between alcohols. An important
question arises when considering the role of reprotonation in
enhancement of enantioselectivity in Pd-catalyzed aerobic OKR
of sec-phenethyl alcohol. This is true whether the relative
difference in reprotonation is due to the diastereomeric interac-
tions of a bulky chiral acid and a bulky chiral Pd-alkoxide
species or due to the inherent ease of reprotonation of one
diastereomer of Pd-alkoxide over the other, regardless of the
nature of the acid. If it is the latter, this would imply that an
achiral base/acid system could be used with Pd[(-)sparteine]-
Cl2 catalyst for OKR of alcohols. In fact, several achiral bases
have been shown to give good selectivities in Pd-catalyzed
aerobic OKR, with carbonates in tBuOH being some of the
+
concentration of (-)-sparteine H can be approximated. Enan-
tioselection during Pd-catalyzed OKR is a complex combination
of thermodynamic and kinetic events. It can be summarized as
a competition between enantiomers of alcohol for an active Pd
catalyst and a balance of kinetic â-hydride elimination and
reprotonation by (-)-sparteine HCl. The model predicts changes
of experimental krel values compared to intrinsic krel values at
various [(-)-sparteine HCl]. Additionally, krel improvement is
possible through the addition of (-)-sparteine HCl. Using this
information, we plan to design and evaluate bifunctional
catalysts for the palladium-catalyzed oxidative kinetic resolution
of alcohols.
Acknowledgment. The genesis of this study arose from a
senior-level undergraduate interdisciplinary course co-taught by
B.D.C. and Dr. Nancy Mills at Trinity University. We very
gratefully thank Dr. Mills for her contributions to this study
and the students enrolled in the course for their discussions.
B.D.C. and A.C. gratefully acknowledge the Research Corpora-
tion (Grant No. W-1552) for financial support of this work.
M.S.S. thanks the National Institutes of Health (NIGMS No.
RO1 GM63540) and the Dreyfus Foundation (Teacher Scholar
Award) for financial support. J.A.M. thanks the American
Association of University Women Educational Foundation for
a 2004-2005 American Fellowship.
3
6,37
best.
There are several limitations to the insights our studies
provide. The chloride ions on Pd[(-)-sparteine]Cl2 are diaste-
reotopic; therefore, two separate alkoxides can be formed from
each enantiomer of alcohol. Thus, the extracted kinetic constants
are combinations of the two possible intermediates for each
enantiomer of alcohol at each step. These two possible diaster-
eomeric forms of alkoxides for each enantiomer of alcohol are
kinetically indistinguishable. The kinetic models provide no
further insight into the active species or the origin of the
“baseline” preference for oxidation of the (R)-enantiomer of the
alcohol; hence, we cannot comment on structural aspects of the
model. In contrast, the solid-state structural and computa-
tional29 models have addressed the possibility of four diaster-
Supporting Information Available: Experimental procedures,
graphs, data tables for the kinetic study, and rate law derivations.
This material is available free of charge via the Internet at
http://pubs.acs.org.
28
(
36) Mandal, S. K.; Sigman, M. S. J. Org. Chem. 2003, 68, 7535-7537.
37) Bagdanoff, J. T.; Ferreira, E. M.; Stoltz, B. M. Org. Lett. 2003, 5, 835-
(
8
37.
JA053195P
14824 J. AM. CHEM. SOC.
9
VOL. 127, NO. 42, 2005