Evaluation of Catalyst Acidity and Substrate Electronic Effects in A Hydrogen Bond-Catalyzed Enantioselective Reaction
Evaluation of Catalyst Acidity and Substrate Electronic Effects in A Hydrogen Bond-Catalyzed Enantioselective Reaction
Evaluation of Catalyst Acidity and Substrate Electronic Effects in A Hydrogen Bond-Catalyzed Enantioselective Reaction
org/joc
sigman@chem.utah.edu
Received July 15, 2010
A modular catalyst structure was applied to evaluate the effects of catalyst acidity in a hydrogen
bond-catalyzed hetero Diels-Alder reaction. Linear free energy relationships between catalyst
acidity and both rate and enantioselectivity were observed, where greater catalyst acidity leads to
increased activity and enantioselectivity. A relationship between reactant electronic nature and rate
was also observed, although there is no such correlation to enantioselectivity, indicating the system is
under catalyst control.
7194 J. Org. Chem. 2010, 75, 7194–7201 Published on Web 10/04/2010 DOI: 10.1021/jo1013806
r 2010 American Chemical Society
Jensen and Sigman
JOC Article
SCHEME 1. Modular Hydrogen Bond Catalyst Design
Mechanistic Investigation
To gain a better understanding of how the observed trends
in yield and enantioselectivity based on catalyst acidity may
reflect the mechanism of the reaction, kinetic measurements
were acquired using in situ IR spectroscopy, and the reaction
rate dependence on each reaction component was deter-
mined (Figure 2, note: rate measured by monitoring the
change in IR absorbance of 1, which is correlated to [1], as
a function of time; rates reported in M/s). A first order
dependence on [catalyst] provides evidence against a dimer
or higher order species as the active catalyst (Figure 2A).
Initially, kinetic measurements were performed at room
temperature for practicality, but based on the observation
of a nonzero intercept, indicating a background reaction, all
further measurements were performed at reduced tempera-
ture (either 0 or -45 °C). Saturation in [aldehyde] was
observed (Figure 2B), as well as a first order dependence
on [diene] at high [aldehyde] (Figure 2 C). These results led to
the proposal of the mechanism shown in Scheme 2, where the
catalyst and aldehyde bind reversibly to form an activated
FIGURE 2. Reaction component dependencies with catalyst 3 complex C:A, which reacts with diene in the turnover limiting
measured by in situ IR spectroscopy. (A) First order dependence step to form the product.
on [catalyst 3], conditions: [benzaldehyde] = 0.37 M, [diene] = 0.18 Upon the basis of the proposed mechanism, rate law
M, rt. (B) Saturation in [aldehyde], conditions: [3] = 0.018 M, derivation was performed following the Michaelis-Menten
[diene] = 0.18 M, 0 °C. (C) First order dependence on [diene], kinetic model. Assuming the second step is the turnover
conditions: [3] = 0.018 M, [benzaldehyde] = 1.0 M, 0 °C.
determining step, the rate of product formation is propor-
catalyst leading to the highest enantiomeric excess. This result tional to the product of the concentration of the two inter-
was initially unexpected, as it was assumed that substituent size mediates involved in this step, where C:A is the activated
had the most significant effect on facial selectivity, and our catalyst-aldehyde complex (eq 1). As the catalyst exists in
assumption was that halogen substitution would have a minor two different states in this mechanism, we define [C]T as the
effect upon the steric nature of the catalyst. total catalyst concentration (eq 2). Using the steady state
Because 3 was discovered to catalyze the hetero-Diels- approximation (eq 3) allows us to solve for [C:A] (eq 4).
Alder reaction with high enantioselectivity, the scope of the Substitution into eq 1 results in the derived rate law (eq 5).
hetero-Diels-Alder reaction with this catalyst was exam-
ined (Table 2). In general, high enantioselectivity was obtained d½P
with aromatic aldehyde substrates (entries 1-5), highlighted ¼ k2 ½C : A½D ð1Þ
dt
by a 96% ee for 1-naphthaldehyde. An R,β-unsaturated alde-
hyde, cinnamaldehyde, also gives the product in high enantio-
½C ¼ ½CT - ½C : A ð2Þ
meric excess (entry 6). Enantioselectivity is diminished for
hydrocinnamaldehyde and hexanaldehyde (entries 7 and 8),
d½C : A
but this marks an improvement as prior experimentation ¼ k1 ð½CT - ½C : AÞ½A - k - 1 ½C : A
had shown catalyst 2 to be ineffective in the catalysis of the dt
hetero Diels-Alder reaction with aliphatic aldehydes. - k2 ½C : A½D ¼ 0 ð3Þ
FIGURE 3. Conversion of diene vs time for hydrogen bond cata- FIGURE 4. Linear free energy relationship between rate and pKa of
lyst 3-7, conditions: [benzaldehyde] = 0.78 M, [diene] = 0.18 M, the corresponding acetic acid derivative, conditions: [benzaldehyde] =
[catalyst] = 0.036 M, -45 °C. 0.78 M, [diene] = 0.18 M, [catalyst] = 0.036 M, -45 °C.
k1 ½CT ½A use the pKa values of acetic acid derivatives measured in
½C : A ¼ ð4Þ water as the reference, with the key assumption being that the
k1 ½A þ k - 1 þ k2 ½D
electron-withdrawing ability of the R substituent affects the
d½P k1 k2 ½CT ½A½D acidity of our catalyst and the acidity of the corresponding
¼ ð5Þ acetic acid derivative similarly.
dt k1 ½A þ k - 1 þ k2 ½D
Importantly, the derived rate law is consistent with the logðkÞ ¼ - RðpKa Þ þ C ð8Þ
empirical observations. Specifically, when [aldehyde] is low
the rate law can be simplified to eq 6 which is consistent with To evaluate the effect of catalyst acidity upon substrate
a first order dependence of rate on [aldehyde], and when binding, the saturation curves for the most and least acidic
[aldehyde] is high the rate law can be simplified to eq 7 catalysts, 3 and 7, respectively, were examined (Figure 5,
which is consistent with zero order dependence of rate on note y-axis scales are different). Comparing the curvature of
[aldehyde], and first order dependence on [catalyst] and these two plots, saturation is reached at much lower [alde-
[diene]. hyde] for the more acidic catalyst (Figure 5A) than for the
least acidic catalyst (Figure 5B), indicating that binding
d½P k1 k2 ½CT ½A½D equilibrium is also perturbed by catalyst acidity.
¼ ð6Þ
dt k - 1 þ k2 ½D With the understanding that catalyst acidity has a direct
effect on both substrate binding and rate of reaction with
d½P diene, we were interested in whether the observed trends in
¼ k2 ½CT ½D ð7Þ
dt enantioselectivity could also be directly correlated to catalyst
acidity. Considering that enantiomeric ratio is a measure-
ment of the relative rate of formation of each enantiomer
Catalyst Acidity Effects (er = kS/kR), a Brønsted-type plot was constructed by plotting
In light of the proposed mechanism, the question re- log(er) vs pKa (Figure 6). A linear free energy relationship was
mained whether catalyst acidity is affecting the binding equili- observed between enantioselectivity and acidity, a unique
bria of catalyst with substrate (k1/k-1), the rate of reaction of example of such a direct correlation in a hydrogen bond
the activated aldehyde C:A with diene (k2), or both. To catalyzed reaction. This indicates that enantioselectivity can
determine the influence upon the second step, the reaction be directly perturbed by the hydrogen bond donating ability.
rate was measured for the complete series of catalysts 3-7
under saturation conditions. A clear trend was observed,
Substrate Electronic Effects
with the most acidic catalyst leading to the highest rate
(Figure 3). In an attempt to further understand the subtleties of the
To determine whether this trend in reaction could be system, we chose to evaluate the influence of substrate electronic
directly correlated to the electronic nature of the catalyst, a perturbations. We hypothesized that the electronic nature of the
Brønsted-type plot was constructed (Figure 4). The Brønsted substrate could potentially affect both the rate at which the
equation (eq 8)25 relates the rate of an acid-catalyzed reac- aldehyde reacts with the diene, as well as the Lewis basicity and
tion to the pKa of the corresponding acid. The resulting thus the binding of substrate to catalyst. Therefore, a series of
linear free energy relationship is similar to the more common benzaldehyde derivatives with either electron-donating or
Hammett relationship, where pKa’s of benzoic acids are used electron-withdrawing substituents in the para position were
as the reference measurement for the effect of electronic examined. In situ IR was used to monitor the rate of the reaction.
perturbations on rate or equilibrium. In our case, we chose to We first examined the two most electronically disparate cases
within the series, p-anisaldehyde 10 and p-trifluoromethylben-
(25) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic Chem- zaldehyde 11. Initial reaction rates were measured over a range
istry; University Science Books: Sausalito, CA, 2006; p 466. of [aldehyde]. Similar rates were observed at low [aldehyde],
J. Org. Chem. Vol. 75, No. 21, 2010 7197
JOC Article Jensen and Sigman
FIGURE 5. Comparison of saturation in [aldehyde] for catalysts 3 and 7. Conditions: [diene] = 0.075 M, [catalyst] = 0.018 M, -45 °C. Fit to
y = ax/(b + x). (A) a = 5.47 10-5 ( 3.1 10-6, b = 0.074 ( 0.017, R2 = 0.97; (B) a = 1.45 10-5 ( 6.4 10-7, b = 0.209 ( 0.029, R2 = 0.99.
Conclusion
We have presented a detailed examination of the effect
of catalyst acidity in an enantioselective hydrogen bond-
catalyzed reaction. This systematic study investigates the effects
of electronic perturbations to both the catalyst and substrate.
A process where catalyst acidity controls enantioselectivity is
revealed. Since our initial report of a direct correlation
between catalyst acidity and enantioselectivity,10 examples
FIGURE 9. Hammett plots show no linear free energy relationship of similar trends in other hydrogen bond-catalyzed reactions
for enantioselectivity, conditions: [benzaldehyde] = 0.36 M, [diene] = have emerged,11,12 implying this effect is not unique to our
0.18 M, [catalyst] = 0.036 M, -45 °C. system. However, ideal catalyst acidity is likely to differ
depending on the specific reaction of interest, as the role of
the catalyst is to stabilize the transition state more than the
bound substrate intermediate, therefore it should not be
assumed that a more acidic catalyst will always lead to
improved enantioselectivity.40 Optimal matching of the pKa’s
of catalyst and transition state structure should lead to
improved catalyst performance in other systems as well,
and the understanding gained in this study may aid in future
catalyst design.
Experimental Section
Preparation of Catalyst 3 (2,2,2-Trifluoro-N-((R)-1-((S)-
4-(hydroxydiphenylmethyl)-4,5-dihydrooxazol-2-yl)-2-phenylethyl)-
acetamide). The synthesis of the Cbz oxazoline intermediate (benzyl
FIGURE 10. Effect of substituents on acidity of hydrogen bond (R)-1-((S)-4-(hydroxydiphenylmethyl)-4,5-dihydrooxazol-2-yl)-2-
donor and acceptor in proposed the transition state. phenylethylcarbamate) 12 has been previously reported.13 A 100
mL round-bottom flask was charged with 44 mg of 10% Pd/C and
inconsistent with the lack of a correlation between substrate the flask was flushed with nitrogen. In a separate 100 mL flask,
electronic nature and enantioselectivity. 442.3 mg of oxazoline 12 (0.873 mmol) was dissolved in 6 mL of dry
A reasonable explanation is that a stronger hydrogen MeOH, and this solution was transferred via cannula into the flask
bond between the catalyst and the substrate at the transition containing Pd/C (on occasions when this is difficult, gentle warm-
ing or addition of excess MeOH is helpful and has no discernible
state is formed with the more acidic catalyst, leading to
effect on overall yield). A further 4 mL of MeOH (2 2 mL) was
increased rigidity in the transition state. It has been reported used for rinsing. The flask was then evacuated under water
that if the pKa of the hydrogen bond donor and that of the aspirator pressure and filled with H2 from a balloon. The cycle
protonated hydrogen bond acceptor are closely matched, a was repeated thrice more, and the reaction mixture was stirred
shorter and stronger hydrogen bond is formed.36-38 While under a H2 balloon atmosphere. On completion of reaction (7-8 h,
the pKa difference between catalyst and protonated substrate by disappearance of oxazoline 12 on TLC, Rf = 0.6 with 1:1
is substantial, they are more closely matched with a more EtOAc/hexanes), the reaction mixture was filtered through a pad
(36) Perrin, C. L. Science 1994, 266, 1665. (39) Based on calculated pKa values reported on Scifinder (Calculated
(37) Cleland, W. W.; Kreevoy, M. M. Science 1994, 264, 1887. using Advanced Chemistry Development Software V8.14).
(38) Schwartz, B.; Drueckhammer, D. G.; Usher, K. C.; Remington, S. J. (40) Davis, T. A.; Wilt, J. C.; Johnston, J. N. J. Am. Chem. Soc. 2010, 132,
Biochemistry 1995, 34, 15459. 2880.