Control of aldol reaction pathways of enolizable aldehydes in an aqueous environment with a hyperbranched polymeric catalyst

Full text of DOI:10.1021/ja806584q

Published on Web 11/25/2008
Control of Aldol Reaction Pathways of Enolizable Aldehydes in an Aqueous
Environment with a Hyperbranched Polymeric Catalyst
Yonggui Chi, Steven T. Scroggins, Emine Boz, and Jean M. J. Fre´chet*
DiVision of Materials Sciences, Lawrence Berkeley National Laboratory and College of Chemistry, UniVersity of
California, Berkeley, California 94720-1460
Received August 19, 2008; E-mail: frechet@berkeley.edu
A central challenge of organic chemistry resides in the control
of reaction pathways to avoid undesired reactions and develop new
or more efficient syntheses. Nature uses precise control to amplify
kinetically or thermodynamically unfavorable transformations with
the assistance of enzymes, catalytic biopolymers that fold into
sophisticated tertiary structures in water.1 A number of important
advances have been made in approximating enzymes with synthetic
materials. These include for example Miller’s synthetic peptide
catalysts,2a Breslow’s polymer/pyridoxamine enzyme mimics,2b
Reymond’s peptide dendrimer catalysts,2c Moore’s catalytic phe-
nylene ethynylene foldamer,2d the metal complex self-assemblies
of Bergman and Raymond,2e and the unimolecular free energy pump
reactor of Fre´chet and Hawker.2f However, it is still extremely
difficult to mimic the complex and precise functional makeup of
an enzyme. In this study we attempted to replicate an enzyme’s
ability to control competing reaction pathways by using synthetic
polymers that promote reactions in an aqueous environment. We
anticipated that such macromolecule catalysts might be used to
direct the reaction pathways in a way not readily achievable with
typical small molecule catalysts.
A class of reactions that attracted our attention are aldehyde
transformations, such as cross ketone/aldehyde aldol condensations.
These reactions are of fundamental importance in organic chemistry
and have broad applications from the manufacture of basic
chemicals to the preparation of fine pharmaceuticals.3 Indeed,
aldehydes are the most common substrates in the recent explosive
development of enamine,4 iminium,5 and SOMO6 catalysis. How-
ever, since enolizable aldehydes are often very reactive as both
nucleophiles and electrophiles, controlling the competing pathways
to avoid self-aldol reactions is an intrinsically unsolved chemose-
lectivity challenge. Usually, a large excess of one reagent is used
to ensure high yielding reactions. For example, cross ketone/
aldehyde reactions are typically carried out using the ketone
substrates as solvent in the presence of either inorganic bases such
as NaOH or amines such as proline as the catalyst.7 Here we report
a soluble polymer catalyst that can eliminate the self-aldol reactions
in an aqueous environment by suppressing an irreversible aldol
pathway, thereby allowing for the amplification of otherwise
unfavorable reactions without the need for excess reactants.
Our investigation began with a careful re-examination of the
widely studied proline-catalyzed aldol reaction4,8 with enolizable
aldehydes as the substrates. A typical model self-aldol of butanal
(1) shows two major competing pathways at room temperature (rt)
(Scheme 1). One is the irreVersible formation of R,ꢀ-unsaturated
aldehyde 2, presumably through a Mannich-type pathway involving
an iminium intermediate formed between aldehyde 1 and proline;
the other is the reVersible formation of ꢀ-hydroxy-aldehyde 3 via
an enamine intermediate. These results (see Supporting Information
[SI]) were consistent with observations and postulations in the
literature.4,8 Our aim was to eliminate the irreversible pathway
leading to the formation of 2. Thus the reversible formation of 3
can be turned into dynamic catalysis in developing new or more
efficient syntheses. We speculated that the undesired pathway might
be disrupted by controlling the charged iminium species involved
in the Mannich-type reaction leading to 2. Initial studies indicated
that the use of typical small molecule amine catalysts (e.g., proline,
pyrrolidine), changes in solvent polarity, or the use of water9 as
solvent or cosolvent failed to suppress the formation of 2. Here we
are interested in the development of soluble synthetic polymers to
address this chemoselectivity problem.
Scheme 1. Competing Pathways of Enolizable Aldehyde Self-Aldol
Reactions
We chose a commercial hyperbranched polyethyleneimine (PEI)
as the scaffold to develop our polymer catalyst. In contrast to the
use of PEI and related amine polymers and dendrimers10 as catalyst
supports11 (e.g., for metal nanoparticles), we aimed to use this type
of water-soluble highly branched polymer to facilitate catalytic
reactions in an aqueous environment.12 Pristine commercial PEI
contains primary and secondary amine groups and is thus not
suitable for our purpose because these amino groups catalyze the
aldol reaction without any control over the competing pathways
shown in Scheme 1. However, simple chemical modification of
PEI readily affords polymers of tunable structures and properties;
for example, reaction of PEI with propylene oxide gave a water-
soluble polymer (a possible structure is illustrated by 4) that met
our requirements (Scheme 2). Instead of introducing catalytically
active sites via covalent linkages, we took advantage of the tertiary
amino groups in 4 as noncovalent handles to attract proline
catalysts13 via electrostatic interactions,14 affording the desired
polymer catalyst (5). This polymer catalyst 5 provided nearly
complete control over the two reaction pathways when the aldol
reaction was carried out in water (Scheme 2).
The formation of ꢀ-hydroxy aldehyde 3 is facile and reversible.
Very little (typically less than 1%) unsaturated aldehyde 2 was
detected even at long reaction times (2 days or longer). Very
interestingly, 5 did not provide any control over the two competing
pathways when the reaction is carried out in organic solVents such
as DMF, dimethyl sulfoxide, or CH2Cl2. For example, a typical
self-aldol reaction of 1 with 10 mol%13 catalyst 5 in water reached
equilibrium with reversible formation of 50-70% ꢀ-hydroxy
aldehyde 3. The remainder was starting aldehyde 1 while essentially
no unsaturated aldehyde 2 was observed after 24 h (chemoselec-
tivity, the ratio between 3 and 2 is greater than 60; Table 1, entry
3). The same reaction (with 5 as the catalyst) carried out in organic
9
10.1021/ja806584q CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 17287–17289 17287

C O M M U N I C A T I O N S
Scheme 2. Control of Aldol Reaction Pathways with a
Hyperbranched Polymer Catalyst in Aqueous Environment
polymer catalyst 5 performs best in controlling the aldol reaction
pathways (Scheme 2).
Scheme 3. Dynamic Catalytic Cross Ketone/Aldehyde Reaction
In view of the numerous chemical transformations involving
enolizable aldehydes as substrates, the catalyst system with polymer
complexes such as 5 might prove useful in the development of more
efficient syntheses. In particular, a dynamic catalytic process might
be developed for the amplification of desired products by reactions
involving either the aldehyde substrate or the reversibly formed
ꢀ-hydroxy aldehyde adduct.16 Initially, we chose to study the cross-
aldol condensation between butanal and acetone to demonstrate this
concept and highlight the potential of this catalytic system. The
product of this condensation, an R,ꢀ-unsaturated ketone, is a key
intermediate in the commercial production of methyl amyl ketone,
an FDA listed food additive.17 In general, R,ꢀ-unsaturated ketones
are both important commercial chemicals and common functional
groups found in complex molecules such as natural products.
Previous methods used to prepare these compounds typically require
the use of a large excess of ketone to compete with the generally
more rapid aldehyde self-condensation.
In a model cross aldol reaction between acetone and butanal using
catalytic complex 5 in water (Scheme 3), the self-aldol reaction of
6 to form ꢀ-hydroxy aldehyde 7 proceeded much faster than the
cross-aldol reaction, reaching maximum conversion in ∼30 min.
However, the facile reversibility of the self-aldol reaction in our
catalytic system led to the eventual formation of the kinetically
disfavored cross-aldol product 8 in more than 90% yield in 22 h.
In this instance, a slight excess of ketone was used to facilitate
monitoring of reaction progress, but it is not necessary in preparative
scale syntheses. Surprisingly, little ꢀ-hydroxy ketone 9 was detected
in the reaction. A sample of 9 prepared using a literature procedure
did not yield any dehydration product (the unsaturated ketone) when
subjected to the same catalytic condition,7b suggesting that the cross
condensation proceeds exclusively via a Mannich-type mechanism.
The Mannich-type pathway is proposed to involve the reaction of
an enamine intermediate formed between acetone and proline as
well as an iminium ion derived from aldehyde and a second
molecule of proline. It remains unclear exactly why the same
iminium intermediate does not react with the enamine derived from
aldehyde and proline to yield the aldehyde self-condensation product
under the aqueous catalytic conditions with polymer complex 5
(see the SI).
Table 1. Self-Aldol Reaction of Enolizable Aldehydea
conv.
(2+3)%b
selectivity
(3:2)
entry
cocatalyst
solvent
pHc
1
2
3
4
5
7
8
9
-
-
DMF
H2O
H2O
DMF
H2O
H2O
H2O
H2O
82
<1
56
86
80
82
77
86
1.4
-
-
-
9.2
-
polymer 4
polymer 4
Et3N
CH3N(EtOH)2
CH3N(EtOH)2
N(EtOH)3
63
0.6
1.0
5.3
3.3
15
11.1
10.1
-
d
e
9.5
a See SI for experimental details. b Measured by 1H NMR of the
crude reaction mixture; diastereoselectivity of aldol product 3 is ∼1.2:1;
enantioselectivity was not determined. c pH of the catalyst solution in
water (before the addition of aldehyde substrate). d N-Methyl-
diethanolamine. e Sodium dodecyl sulfate (a surfactant) was added (see
SI).
solvents such as DMF under otherwise identical conditions led to
poor chemoselectivity with the irreversible formation of 2 as the
major product (Table 1, entry 4). In initial studies, triethylamine
(TEA) tested as the small molecule analogue of polymer 4 under
otherwise identical conditions showed poor reaction control (Table
1, entry 5). Further experiments using tertiary amines containing
alcohol groups (N-methyldiethanolamine, triethanolamine) in an
aqueous medium gave results better than that with TEA (Table 1,
entry 6-9). The pH values of the reaction medium were measured.
It appears that pH has an effect on the reaction selectivity, and
reactions in aqueous media with pH ∼9.0 gave optimal results.15
The catalytic conditions provided by the polymer can be ap-
proximated but are difficult to duplicate; therefore significant
formation of the undesired self-aldol product 2 was still observed
with the small molecule tertiary amines tested under various
conditions, including different pH’s (see SI). The use of alcohols,
such as methanol, 2-propanol, and 2,2,2-trifluoroethanol as additives
or cosolvents, does not improve reaction selectivity. In all cases,
To probe the substrate scope, we first examined cross acetone/
aldehyde condensations with unhindered straight-chain aliphatic
aldehydes of different size and hydrophobicity. Increasing aldehyde
hydrophobicity leads to a large decrease in reaction efficiency due
to the polar nature of the aqueous polymer catalytic phase (see SI).
This selectivity determined by substrate hydrophobicity may be used
to design polymer-assisted polarity gradient-directed chemoselective
9
17288 J. AM. CHEM. SOC. VOL. 130, NO. 51, 2008

C O M M U N I C A T I O N S
Scheme 4. Catalytic Cross Ketone/Aldehyde Condensation
Supporting Information Available: Experimental details and
additional discussions. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) (a) Garcia-Viloca, M; Gao, J.; Karplus, M.; Truhlar, D. G. Science 2004,
303, 186. (b) Benkovic, S. J.; Hammes-Schiffer, S. Science 2003, 301,
1196.
(2) (a) Davie, E. A. C.; Mennen, S. M.; Xu, Y.; Miller, S. J. Chem. ReV. 2007,
107, 5759. (b) Liu, L.; Breslow, R. J. Am. Chem. Soc. 2002, 124, 4978.
(c) Delort, E.; Darbre, T.; Reymond, J. L. J. Am. Chem. Soc. 2004, 125,
15642. (d) Heemstra, J. M.; Moore, J. S. J. Am. Chem. Soc. 2004, 126,
1648. (e) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Science 2007,
316, 85. (f) Piotti, M. E.; Rivera, F., Jr.; Bond, R.; Hawker, C. J.; Fre´chet,
J. M. J. J. Am. Chem. Soc. 1999, 121, 9471.
(3) (a) Mahrwald, R. Modern Aldol Reactions, Vols. 1-2; Wiley-VCH:
Germany, 2004. (b) Evans, D. A. Aldrichimica Acta 1982, 15, 23.
(4) (a) List, B.; Lerner, R. A.; Barbas, C. F., III J. Am. Chem. Soc. 2000, 122,
2395. (b) Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Chem. ReV
2007, 107, 5471.
(5) (a) Ahrendt, K. A.; Borths, C. J.; MacMillan, D. W. C. J. Am. Chem. Soc.
2000, 122, 4243. (b) Erkkila, A.; Majander, I.; Pihko, P. M. Chem. ReV.
2007, 107, 5416.
(6) Beeson, T. D.; Mastracchio, A.; Hong, J.; Ashton, K.; MacMillan, D. W. C.
Science 2007, 316, 582.
(7) (a) Barnicki, S. D.; McCusker-Orth, J. E.; Miller, J. L. U.S. Pat Appl. Publ.
US2005004401 and reference therein. (b) List, B.; Pojarliev, P.; Castello,
C. Org. Lett. 2001, 3, 573.
(8) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. Engl. 1971, 10,
496.
(9) Zotova, N.; Franzke, A.; Armstrong, A.; Blackmond, D. G. J. Am. Chem.
Soc. 2007, 129, 15100.
(10) (a) Fre´chet, J. M. J.; Tomalia, D. A. Dendrimers and Other Dendritic
Polymers; Wiley: New York, 2001. (b) de Brabander-van den Berg,
E. M. M.; Meijer, E. W. Angew. Chem., Int. Ed. Engl. 1993, 32, 1308.
(11) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V. L.; Yeung, K. Acc. Chem.
Res. 2001, 34, 181.
(12) For an example of using PEI to enhance reaction rate, see: Royer, G. P.;
Klotz, I. M. J. Am. Chem. Soc. 1969, 91, 5885–5886. Also see ref 2b for
related examples.
(13) Proline and 4 used in a 1:2 molar ratio (based on amino groups) to prepare
5. Preliminary studies indicated that increasing the amount of 4 leads to
faster aldol reaction. The catalyst loading discussed in this manuscript refers
to proline, unless otherwise specified.
reactions, such as aldehyde/aldehyde cross-aldol reactions.18 The
mild catalytic conditions that prevail with 5 should enable the use
of this catalyst with substrates containing functional groups such
as esters, which are not compatible with the use of strong bases
(e.g., NaOH) as catalysts. A small set of cross-aldol condensation
products, including variations in ketone substrates, is illustrated in
Scheme 4.19 Our method provides access to R,ꢀ-unsaturated ketones
in an efficient manner without the need for excess reagents. While
our polymer catalyst was solely designed to control reaction
pathways, a side benefit of such polymer catalyst systems is the
relative ease of catalyst recycling without further proline addition.20
In summary, we have developed an aqueous polymer catalyst
system that controls the challenging aldol reaction pathways of
enolizable aldehydes, a problem that had remained intrinsically
unsolved previously. Such control of reaction pathways allows
dynamic catalytic processes for the amplification of otherwise
unfavorable reactions. Although we have primarily focused on
addressing the chemoselectivity problems at this point, studies in
progress indicate that stereoselective reactions are achievable using
this catalyst system. Ongoing work also suggest that very chal-
lenging reactions, such as those involving hindered and unreactive
substrates with the generation of quaternary carbon centers, can be
developed. Given the large number of transformations that involve
enolizable aldehydes as substrates, we anticipate that this catalyst
and its polymer or small molecule analogues may be useful in a
broad range of new or more efficient syntheses. Overall, the design
of macromolecule catalysts that can address fundamental challenges
is of both conceptual and practical importance in chemistry.
(14) (a) Bahulekar, R.; Ayyangar, N. R.; Ponrathnam, S. Enzyme Microb.
Technol. 1991, 13, 858. (b) Haimov, A.; Cohen, H.; Neumann, R. J. Am.
Chem. Soc. 2004, 126, 11762.
(15) A conclusive relation between pH and reaction selectivity cannot be drawn
at this moment (for example, at similar pH, reaction in a mixture of water
and 2,2,2-trifluoroethanol gave poor selectivity; see SI). For relevant
discussions, see:Janda, K. D.; Dickerson, T. J. J. Am. Chem. Soc. 2002,
124, 3220. Brogan, A. P.; Dickerson, T. J.; Janda, K. D. Angew. Chem.,
Int. Ed. 2006, 45, 8100. Reymond, J. L.; Chen, Y. J. Org. Chem. 1995,
60, 6970. Other parameters (such as solvent properties) can have profound
effects on the reaction outcome too.
(16) The aldol product 3 may undergo further aldol reactions, especially at
elongated reaction time (for an example, see:Cordova, A.; Notz, W.; Barbas,
C. F., III J. Org. Chem. 2002, 67, 301. ) suggesting that useful reaction
may be developed by reacting with this aldol product.
(17) Letts, J. B. U.S. Patent 4739122, 1988.
(18) Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 6798.
1
(19) Reaction conditions were not optimized. Conversion is determined by H
NMR analysis of the crude reaction mixture (see the SI). Isolated yields
after column chromatography are in the range 50% to 80% for 11c and
11h.
(20) For example, in a scalable preparation of 11c and 11h (Scheme 4),
distillation yielded a mixture with two separate layers: an organic layer
containing the desired product in high purity and a water layer. The white
solid residue remaining after distillation contains the catalyst, which was
directly reused without purification, showing only a slight loss in catalytic
activity.
Acknowledgment. Financial support from DOE-BES (DE-
AC02-05CH11231) is acknowledged with thanks.
JA806584Q
9
J. AM. CHEM. SOC. VOL. 130, NO. 51, 2008 17289