L-proline functionalized polymers prepared by RAFT polymerization and their assemblies as supported organocatalysts

Article abstract of DOI:10.1021/ma201256m

We have prepared a range of well-defined copolymers of styrene and l-proline functionalized styrene (5-11 kDa) using reversible addition-fragmentation chain transfer (RAFT) polymerization techniques and explored their use in supported catalysis. Upon depr

Full text of DOI:10.1021/ma201256m

ARTICLE
pubs.acs.org/Macromolecules
L-Proline Functionalized Polymers Prepared by RAFT Polymerization
and Their Assemblies as Supported Organocatalysts
Annhelen Lu,† Thomas P. Smart,‡ Thomas H. Epps, III,‡ Deborah A. Longbottom,*,§ and
Rachel K. O’Reilly*,†
†Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, United Kingdom
‡Department of Chemical Engineering, University of Delaware, 150 Academy Street, Newark, Delaware 19716, United States
§Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, United Kingdom
S Supporting Information
b
ABSTRACT: We have prepared a range of well-defined co-
polymers of styrene and L-proline functionalized styrene
(5À11 kDa) using reversible additionÀfragmentation chain
transfer (RAFT) polymerization techniques and explored their
use in supported catalysis. Upon deprotection of the L-proline
functionalities, the solution self-assembly of these copolymers
was investigated in mixed solvent systems. The resulting
assemblies were characterized by dynamic light scattering,
transmission electron microscopy (on graphene oxide sub-
strates, along with cryo-TEM and tomography), and scanning
electron microscopy. The application of these functional as-
semblies as supported catalysts for the aldol condensation reac-
tion was explored using cyclohexanone and 4-nitrobenzaldehyde. The rate and selectivity of solution catalysis in our self-assembled
system were comparable to those of L-proline, and a significant advantage of our system was that the polymer support could be
utilized at lower catalyst loadings with comparable activity and also could be recycled a number of times while maintaining activity
and selectivity.
’ INTRODUCTION
The use of organic molecules as catalysts in asymmetric
reactions has garnered great attention in recent years as these
molecules serve as attractive alternatives to transition metal and
enzyme based catalysts.1À3 This class of catalysts, named orga-
nocatalysts, are not only often less expensive but also sometimes
Figure 1. Structures of the bifunctional organocatalysts L-proline (left)
and its derivative 4-hydroxy-L-proline (right).
readily available from the chiral pool and experimentally tractable,
as the majority of them are stable to both water and oxygen.1,4,5
Nature’s aldolase enzymes carry out the asymmetric aldol
reaction via an enamine based mechanism,6À8 and the amino
acid L-proline has been named the “simple enzyme”, as it is
proposed to mimic the aldolase enzymes in this mechanistic
aspect.9 The organocatalytic properties of L-proline were first
reported in the 1970s by Hajos and Parrish10 and Eder et al.,5 and
L-proline has since become one of the most popular organoca-
talysts in asymmetric synthesis. L-Proline and its derivatives
(Figure 1) are highly effective and selective catalysts used in a
number of asymmetric reactions including Mannich reactions,4,11
Michael additions,12,13 Robinson annulations,14 and aldol con-
densation reactions.4,6 These reactions are of interest as a new
asymmetric carbonÀcarbon bond is formed, and thus they are
often key steps in the synthesis of natural products and non-
natural drug molecules.8,15
and in particular the acid functionality. Barbas,16,17 Hayashi,18
and others19,20 showed high selectivity with some of their deriv-
atives; however, a number of derivatives have also resulted in
reduced performance.21À23 These studies highlight the key
role of sterics and electronics around the active centers of the
L-proline in mediating highly selective and efficient catalysis.
Additionally, it has been suggested that the presence of a polymer
scaffold may have a beneficial influence on the catalyst activity, as
it can to an extent be considered as a peptide-backbone mimic.
Therefore, great efforts have been dedicated to the immobiliza-
tion of L-proline on solid supports to enhance catalytic activity as
well as to facilitate the recycle and reuse of the catalyst.24
Received: June 3, 2011
Thus, much research has been concentrated on improving the
performance of L-proline via modifications to the active centers
Revised:
Published: September 06, 2011
August 13, 2011
r
2011 American Chemical Society
7233
dx.doi.org/10.1021/ma201256m Macromolecules 2011, 44, 7233–7241
|

Macromolecules
ARTICLE
Scheme 1. Coupling Reaction of 1 and VBA To Give Monomer 2
Benaglia and co-workers25À27 have anchored L-proline onto a
poly(ethylene glycol) (PEG) scaffold, which was successfully
recovered and reused several times in the aldol condensation,
with close to unchanged yields and only small losses in selectivity.
Polystyrene (beads),28À31 Merrifield resin,32 ionic liquids,33,34
and dendrimers35 are examples of other reported supports for
L-proline. While these studies highlight the potential to selec-
tively incorporate L-proline into a polymer scaffold without a loss
in activity, the selective incorporation of L-proline functionalized
monomers into polymers is a less well-explored area. This
is somewhat surprising given the pioneering work of Endo
and Sanda36 and more recently others on the preparation,
polymerization, and properties of a range of vinyl amino acid
monomers including L-phenylalanine, L-leucine, L-alanine, and
L-histidine.37À40
reaction. Various solvents were screened, and the mixed solvent
systems with DMSO/water and DMF/water gave the most
promising results. Conversions of up to 95%, anti/syn ratios
up to 95/5, and enantiomeric excesses up to 94% were achieved
when the aldol reaction was carried out in the mixed solvent
systems. These findings were not entirely unexpected, as water
has been reported to enhance the rate of reaction, as it helps
promote the formation of the enamine transition state.60À64
Given these initial promising results, we are further exploring this
chemistry with our RAFT derived polymers to allow for access
to highly active recyclable supports. Furthermore, we propose
that the selective incorporation of the hydrophilic amino acid
L-proline into a hydrophobic polystyrene backbone will yield
polymers with tunable solution properties to allow for self-as-
sembly of the resultant copolymers. We hypothesize that the
formation of higher order assemblies will affect the catalytic
activity of the L-proline in the copolymers and allow for improved
postreaction catalyst recovery. These self-assembly effects on
activity and recovery are explored in this work.
In recent years the group of Hansen and co-workers24,41,42 has
reported the synthesis of several L-proline functionalized metha-
crylate monomers and a range of L-proline functionalized styrenic
and methacrylic copolymer beads. The L-proline functionalized
beads showed both high activity and selectivity in water and were
successfully recycled five times without a significant loss in
catalytic activity. Although this technique may be used on a large
industrial scale, with simple work-up and purification and good
control of catalyst loading, this type of (co)polymerization
technique does not allow for the controlled synthesis of more
complex or well-defined polymer architectures such as block
copolymers/amphiphiles, which may undergo self-assembly to
form higher-order structures and perhaps find utility as organo-
catalytic nanoreactors.43
To obtain improved copolymerization polydispersity control,
controlled radical polymerization (CRP) techniques are often
employed instead of conventional free radical methods. In
particular, the application of reversible additionÀfragmentation
chain transfer (RAFT) polymerization for the preparation of
well-defined functional copolymers has received much attention
in the past decade.44À53 RAFT polymerization is a versatile
polymerization technique as it proceeds under relatively mild
reaction conditions, is highly tolerant to a variety of monomers
including monomers with unprotected functional groups (such
as acids, alcohols, and amines), and is amenable to a number of
postpolymerization modifications including end-group removal
and chain extension to afford well-defined block copolymers.
Indeed, Endo and co-workers54À57 have reported the successful
use of RAFT techniques to incorporate L-proline and 4-hydroxy-
L-proline units into a polyacrylamide backbone, though not
leaving the key amine functionality available to play its essential
role in the aldol reaction. More recently, our group has reported
the successful synthesis of L-proline functionalized polystyrene
copolymers58,59 via both nitroxide-mediated polymerization
(NMP)58 and RAFT polymerization techniques.59 The organo-
catalytic activity and selectivity of the copolymers synthesized via
the NMP technique were carefully studied in a typical aldol
’ RESULTS AND DISCUSSION
Monomer Synthesis. To afford the means to attach the
organocatalytically active L-proline to the polymer backbone,
while leaving both the amino and carboxylic acid functionalities
of the L-proline available for catalysis, a bis-protected 4-hydroxy-
L-proline functionalized monomer, protected at the amino and
acid functionalities, was synthesized. First, commercially avail-
able Boc-O-benzyl-L-hydroxyproline was modified via two pro-
tection/deprotection reactions to yield 1.65 Then, precursor 1
was coupled with 4-vinylbenzoic acid (VBA) via the Steglich
esterification66 reaction to provide the desired bis-protected
L-proline functionalized styrenic monomer 2 (Scheme 1). Our
group has previously shown that this coupling technique does
not result in the racemization of the key chiral C2 carbon.59
The amino and acid functionalities on L-proline were pro-
tected with a tert-butyloxycarbonyl (Boc) group and a tert-butyl
ester (tBu) group, respectively. These protecting groups were
selected due to the simple deprotection strategy that may be
employed to simultaneously reveal both the L-proline carboxylic
acid and amino functionalities.67 This deprotection strategy was
more straightforward than the benzyl carbamate (Cbz) and
benzyl ester (Bn) protection methodology previously reported
by our group.59 The resultant monomer 2 was characterized by
1H and 13C NMR spectroscopy, high-resolution mass spectros-
copy (HR-MS), IR, and elemental analysis; the IR spectrum is
1
shown in Figure S1. A number of signals in the H NMR
spectrum of 2 showed unexpected splitting patterns, and thus
the assignment of the protons present in the L-proline five-
membered ring (labeled fÀi in Figure 2) was initially explored
using 2D COSY NMR spectroscopy. The observed splitting
patterns of the protons labeled g and i arise because these protons
7234
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Figure 2. 1H NMR (400 MHz) spectrum of monomer 2, in CDCl3 at 25 °C and 60 °C.
Scheme 2. Scheme Representing Bulk Copolymerization of 2 with Styrene Using 3 as CTA To Give Copolymers 4
are situated between two stereocenters in the ring and experience
different environments, causing the splitting pattern shown in
Figure 2. The unusual splitting noted at δ 4.30 (proton h) was
due to a rotameric effect of the neighboring Boc N-protecting
group, which was resolved by high-temperature 1H NMR
spectroscopy at 60 °C in CDCl3, verifying that no racemization
had taken place during the coupling process.
Copolymerization of Functionalized Monomer with Styr-
ene Using RAFT. RAFT polymerization was utilized to incorpo-
rate the L-proline functionalized monomer into a polystyrene
based copolymer (Scheme 2). Dodecyl 1-phenylethyl trithio-
carbonate 3 (Figure S2) was used as the RAFT agent, or chain
transfer agent (CTA), as trithiocarbonates are suitable for the
polymerization of styrene and styrenic monomers.68À70
The polymerization resulted in the desired catalyst loading on
the polymer (ca. 5%), with a narrow polydispersity index
(ca. 1.1), showing the polymerization proceeded with good control.
Percentage conversion of the two monomers was determined by
comparing distinct monomer signals (δ 5.30 ppm for L-proline and
δ 5.13 ppm for styrene) to a distinct polymer signal at δ
5.45 ppm for L-proline and δ 6.20À7.30 ppm for polystyrene
(Figure S3), and the 1H NMR signals were utilized to estimate the
molecular weight of the resultant copolymer. Percentage incor-
poration of 2 was determined by comparing a polymer signalin the
1H NMR spectrum for each monomer to a distinct end-group
signal from the chain transfer agent, CTA (at δ 3.20 ppm). This
information was further used to confirm the molecular weight of
the copolymer; however, it should be noted this method assumes
7235
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Table 1. Copolymerization Data for 2 and Styrene, with
95:5:1 Ratio of Styrene:2:3, Polymerization in Bulk at 110 °C
(Reaction Quenched Depending on the Desired DP of
Styrene)
Table 2. Data Showing Conversion, Final Diastereomeric
Ratio (dr), and Enantiomeric Excess (ee), Using Copolymer
5c at 20 mol % and Unsupported L-Proline at 20 mol % after
24 h
c
d
Mn
Mn
% incorporation
polymer DPa DPb (kDa) (kDa) Mw/Mn
of 2c
d
4a
4b
4c
4d
4e
4f
93
34
65
64
93
83
74
4
3
3
3
3
4
4
11.7
5.2
9.5
5.0
1.06
1.07
1.06
1.13
1.07
1.07
1.07
4.0
8.0
4.5
4.5
3.0
4.6
5.0
% water
in DMF
% conv
HPLC
% conv
NMR
8.4
8.7
entry
cat.
anti/syna
% eeb
8.3
8.5
11.6
10.3
10.2
9.5
1
2
3
4
5
6
7
8
5c
5c
5c
5c
0
50
90
82
87
49
98
90
90
21
99
99
98
89/11
95/5
95/5
91/9
91/9
92/9
92/8
87/13
91
89
85
94
83
86
88
82
12.3
11.6
5
4g
7.5
12
0
a Of styrene, determined by 1H NMR spectroscopy, in CDCl3. b Of 2,
determined by 1H NMR spectroscopy, in CDCl3. c Determined by 1H
NMR spectroscopy, in CDCl3. d Determined by GPC, in THF, PS
calibration.
L-proline
L-proline
L-proline
L-proline
5
99
99
97
7.5
12
100% end-group fidelity. The presence of the CTA end group in
the copolymers was confirmed by examining the UV trace signal
collected from GPC analysis at 309 nm, the wavelength char-
acteristic for the trithiocarbonate end group.
a Determined by H NMR spectroscopy in CDCl3. b Determined by
1
chiral HPLC, Chiralpak IA column (80:20 hexane:IPA).
r1 Â r2 ∼ 1, confirming an extreme ideal copolymerization given
the very different reactivity ratios for the pair of monomers. This
difference in reactivity ratios implies that M1 (monomer 2) is
preferentially added to the propagating species, supporting
results obtained via the linear model. A plot showing this extreme
ideal behavior generated in Contour is shown in Figure S5. This
behavior is not entirely unexpected as the presence of the
electron-withdrawing ester linkage present at the para-position
on monomer 2 is expected to increase the rate of polymerization
of 2 compared to that of styrene. However, we propose that this
nonrandom distribution of monomer 2 in the polymer backbone
will lead to an amphiphilic copolymer (once monomer 2 is
deprotected), potentially allowing for interesting self-assembly
behavior.73,74
Additionally, the active functionalities of monomer 2 were
deprotected prior to polymerization (see Supporting In-
formation), and this new monomer could be successfully copo-
lymerized using RAFT with styrene (under the same conditions
highlighted in Table 1) to afford a copolymer of similar composi-
tion to the protected analogue (Mn,NMR = 14.3 kDa, Figure S5);
however, its SEC characterization was difficult due to the
amphiphilic nature of the copolymers, and hence the protected
monomer 2 was used in this study.
Reactivity Ratio of Copolymerization. Previous initial in-
vestigations into the copolymerization kinetics carried out in our
group suggested a faster polymerization rate for monomer 2
compared to styrene.59 This led us to hypothesize a nonrandom
polymerization and instead propose a gradient-type structure for
these L-prolineÀstyrene copolymers 4. To confirm this hypoth-
esis, the reactivity ratios of the two monomers in this copolym-
erization were investigated; thus, different ratios of the two
monomers (styrene/2 with ratios 90/10, 80/20, 70/30, 60/40,
and 50:50) were polymerized with 3 under the same conditions
(in bulk at 110 °C). An aliquot was taken after 30 min to
determine the conversion by 1H NMR spectroscopy. The mole
ratio of each monomer in the resultant copolymer and the initial
monomer feed were used to determine the reactivity ratios of the
monomers. The linearized FinemannÀRoss (FR) model was
used for this analysis,71 and the model concluded that r1 > r2 (r1 =
9.3; r2 = 0.11), in the case where monomer 2 is M1 and styrene is
M2, suggesting blocklike polymerization and thus validating our
hypothesis as to the nature of the copolymer. The parameters
used to calculate r1 and r2 using the FR model are given in
Table S1.
Polymer Deprotection. The Boc and tBu protecting groups
for the amine and carboxylic acid, respectively, can be readily
deprotected under acidic conditions.67 Thus, copolymers 4aÀg
were deprotected in dry CH2Cl2 in the presence of trifluoroacetic
acid (TFA) to afford copolymers 5aÀg. Successful deprotection
was confirmed by comparing signals the 1H NMR data given in
t
Figures S3 and S4. The Bu signals disappeared following
deprotection, causing a simplification in the 1.15À2.60 ppm
region, and the characteristic L-proline signal from the proton
labeled “a” has now shifted from δ 5.45 to 5.63 ppm (Figure S4).
To confirm that the ester, which links the catalytically active
L-proline moiety to the polymer backbone, was not also affected
under these strong acidic conditions, monomer 2 was stirred in
TFA under the deprotection conditions. Under these conditions
the L-proline was successfully deprotected, and the ester linkage
was found to be stable.
Additionally, Contour, a computer program developed by van
Herk72 using a nonlinear least-squares fitting method, was used
to investigate our copolymer system. The F1 and f1 values of the
copolymerization were entered into the program, and a set of
calculations were performed to rapidly determine the reactivity
ratio of the two monomers that fits the input criterion, where F1 =
copolymer composition and f1 = feed composition. The gener-
ated values for r1 and r2 were 12.4 and 0.01, respectively, and
Organocatalytic Properties of L-Proline Functionalized
Polymer. The catalytic activity of the synthesized L-proline
functionalized copolymers (5aÀg) was tested using a represen-
tative aldol reaction between cyclohexanone and 4-nitrobenzal-
dehyde. This system has been previously studied with other
solid-supported L-proline catalysts, and thus the catalyst activity
can be directly compared.29,32,75 Previously, our group has shown
that our NMP copolymers with different catalyst incorporations
7236
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Figure 4. Kinetic plot comparing the rate of the aldol reaction in 100%
DMF (O, b) and in DMF/5 vol % water (0, 9), for polymer 5e as
determined by HPLC analysis.
Figure 3. Kinetic plot showing rate of reaction against time in the
L-proline catalyzed aldol reaction, at 20 mol % of copolymer 5c with
varying water concentrations and with unsupported L-proline at 5 vol %
water for rate comparison; all conversions were determined by HPLC
analysis.
Table 3. Comparison of the Activity and Selectivity of the
Aldol Reaction Carried Out at 10 and 20 mol % Catalyst
Loading after 24 h
successfully catalyzed the aldol reaction at 10 mol % catalyst
loading and at relatively high comonomer incorporation.59 To
further explore the potential of the L-proline-supported system,
the RAFT derived copolymers were synthesized with lower
catalyst incorporation, ca. 5%. It was proposed that lowering
the catalyst loading would change the copolymer’s solution
properties and aid in the recycling of the support. To confirm
that the RAFT derived copolymer 5c (as a representative
example) was able to catalyze the aldol reaction with high yield
and selectivity, the reaction was first carried out using 20 mol %
catalyst loading. It was interesting to note that the reaction did
indeed take place and that, with water as a cosolvent, reactions
uniformly gave a higher conversion after 24 h compared to those
reactions performed in the absence of water (Table 2).60À62,76
The reaction with 5 vol % water present gave the best overall
result (98% conversion, 95:5 anti:syn, 89% ee after 24 h), while
those with supported L-proline at 7.5 and 12 vol % water gave
lower conversions. The addition of more than 12 vol % water was
not possible, as this caused the polymer-supported catalyst to
precipitate out of solution, presumably due to the hydrophobic
nature of the polystyrene.
To provide a direct comparison, the aldol reaction was carried
out under the same conditions with unsupported L-proline at
20 mol % with 5, 7.5, and 12 vol % water (Table 2), all of which
gave high conversion (98À99%) but slightly reduced enantios-
electivity (82À88%) compared with the polymeric system
(85À94%). Similarly to the supported system, when the un-
supported L-proline reaction was carried out in pure DMF, a
lower conversion after 24 h was observed, which was consider-
ably lower than the polymer supported system. The kinetic plots
for the unsupported L-proline catalyzed aldol reactions at differ-
ent DMF/water ratios are shown in Figure S6.
% water in
DMF
% conv
HPLC
% conv
NMR
entry
mol %
anti/syna
% eeb
1
2
3
4
0
5
0
20
20
10
10
50
90
27
95
49
98
38
95
89/11
95/5
91
89
87
93
85/15
5
1
95/5
a Determined by H NMR spectroscopy in CDCl3. b Determined by
chiral HPLC, ChiralPak IA column (80:20 hexane:IPA).
the supported system also shows high diastereoselectivity and
enantioselectivity.
The aldol reaction was also carried out at 20 mol % catalyst
loading with copolymer 5e and 5b, a higher and lower molecular
weight polymer, respectively (Mn = 11.6 and 5.0 kDa), in 100%
DMF and 5 vol % water in DMF. The results obtained were
comparable to those using 5c and led to the conclusion that the
rate and stereoselectivity of the reaction (e.g., for 5e after 24 h,
96% conversion, anti/syn ratio 96/4, ee 98% (for 5c, 98%, 88/12,
and 85%) in the mixed solvent system) are comparable between
polymer samples of differing molecular weight.
To further investigate the potential of our polymer-supported
catalyst, using the most efficient conditions (5 vol % water in
DMF), the aldol reaction was carried out at a lower catalyst
loading (10 mol %). Figure 4 shows the differences in rate of
reaction when the catalyst loading was reduced from 20 to 10 mol %.
Regardless of catalyst loading, the reaction shows a distinct
enhancement in rate in the presence of water; in 100% DMF,
the rate for both catalyst loadings, throughout the 24 h is
significantly reduced, generally reaching lower final conversions,
anti/syn ratios, and enantiomeric excesses (ee) (Table 3,
entries 1 and 3). However, in DMF with 5 vol % water, though
the reaction proceeds more slowly for the first 9 h, final
conversions, anti/syn ratios, and enantiomeric excesses after 24 h
are not significantly different (Table 3, entries 2 and 4). Inter-
estingly, the reaction using 20 mol % catalyst loading in 100%
A more detailed picture was obtained by monitoring the
reaction progress at regular intervals over 9 h and then sampling
after 24 h (Figure 3). The increased rate of reaction in the
presence of water is clear from the data (after 9 h, 40% conversion
in DMF vs 74% conversion with 5 vol % water in DMF); how-
ever, it appears that 7.5 vol % water or greater has a detrimental
effect on the reaction rate of the supported L-proline (cf. after 9 h,
ca. 60% conversion for 7.5 vol % water in DMF), likely due to the
gradual reduction in polymer solubility at higher water contents.
It is interesting to note that upon supporting L-proline onto a
polymer scaffold its catalytic activity is retained, and importantly,
7237
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Table 4. Comparison of Noncatalyzed and Catalyzed Aldol
Reaction under the Same Experimental Conditions
entry
cat.
mol %
% conva
anti/synb
% eec
1
no catalyst
L-proline
L-proline
L-proline
L-proline
PS
À
1
no reaction
92/8
2
12
57
99
99
77
94
94
97
Figure 5. Graphical representation of DLS results, showing the size
(Dh,app) of copolymer 5c in different solvent systems (at 10 mg/mL).
3
5
90/10
92/8
4
10
20
À
20
1
5
90/10
no reaction
95/5
as L-proline. The enantioselectivity of the reaction was deter-
mined by HPLC, on a Chiralpak IA column, and the racemic
aldol product was run as a point of reference. These results
suggested the polymer-supported catalyst is capable of achieving
high ee values, comparable to those obtained for unsupported
L-proline. Generally, ee values of 90% or higher were found for
the reactions catalyzed by copolymers 5.
6
7
5c
98
57
89
95
96
89
91
92
93
98
8
5e
95/5
9
5e
5
96/4
10
11
5e
10
20
95/5
5e
96/4
a Percentage conversion determined by 1H NMR spectroscopy in
Self-Assembly of the Copolymer. As previously mentioned,
we hypothesized a gradient type profile of the two monomers in
copolymers 5. This profile is important as monomer 2 can be
considered amphiphilic and hence may undergo a selective
assembly process in mixed solvent systems to create a unique
microenvironment for the L-proline catalyst, possibly influencing
its catalytic behavior. Indeed, we wanted to further explore
whether the observed increase in rate of the supported system
in mixed solvent system was related to the formation of well-
defined aggregates. The copolymers 5c and 5e were dissolved in
DMF and in DMF with 5 vol % water and were examined by
dynamic light scattering (DLS) and microscopy (TEM and
SEM) techniques. Because of the high concentration of copolymer
used in the aldol reactions (ca. 90 mg/mL), it was not possible to
perform analysis via light scattering or light microscopy on the
actual reacting systems; thus, a lower concentration (10 mg/mL)
was used to characterize the copolymers solutions.
DLS studies of the fully dissolved copolymer 5c in DMF gave
hydrodynamic diameters of 3 nm (PD 0.204), indicating the
presence of unimers in solution. However, upon the addition
of water (at least 5 vol %) large structures with hydrodynamic
diameters (Dh,app) of 160 nm (PD 0.171) were found, confirming
a change from unimer polymer chains to chain association to
form aggregates. A graphical representation of these results is
shown in Figure 5. We also extended our analysis of copolymer
5c by changing the DMF:water ratio to larger water contents.
The difference in size between 5 and 7.5 vol % water (ca. 170 nm,
PD ca. 0.25 in both cases) was not significant, and it was not until
12 vol % water had been added that a considerable increase in Dh,app
(220 nm, PD 0.327) was found. However, it shouldbenotedthatas
the water content increased, a broadening in the polydispersity was
found. The correlation functions of the DLS runs are shown in
Figure S8. The size of the aggregates is larger than expected for
conventional micelles, as they are over 100 nm and hence may be
micelle aggregates, vesicles, or compound micelles.77
1
CDCl3 after 24 h. b Determined by H NMR spectroscopy in CDCl3
after work-up. c Determined by chiral HPLC, ChiralPak IA column
(80:20 hexane:IPA).
DMF gives a similar outcome to that of 10 mol % catalyst loading
in DMF with 5 vol % water, highlighting the importance of water
for enhancing the performance of the catalyst and allowing for
lower catalyst loadings.
To ensure the activity and selectivity observed are in fact
related to the presence of our polymer-supported catalyst,
additional reactions were carried out both with unsupported
L-proline and with a nonfunctionalized polystyrene in the reac-
tion mixture (in the absence of any catalyst). These reactions
were carried out using the same conditions as previously
discussed, in DMF with 5 vol % water for 24 h. The results are
detailed in Table 4 and demonstrate the similarity in both activity
and selectivity between our copolymer system and unsupported
L-proline at comparable loadings. However, when lower loadings
were used, i.e., 5 and 1 mol %, the difference in activity between
supported and unsupported catalysts becomes greater. The
performance of the unsupported L-proline is reduced to a greater
extent than that of the supported system at lower loadings. At the
lowest loading, 1%, unsupported L-proline reached 12% conver-
sion in 24 h compared to 57% (with excellent anti/syn ratio and
% ee) for the supported L-proline. These results further highlight
the advantage of the polymer supported system.
Characterization of Aldol Products. 1H NMR spectroscopy
was used to determine the diastereoselectivity of the reaction by
comparing the signal for the proton labeled “d” at δ 4.82 (anti-H)
and δ 5.41 (syn-H) in Figure S7. The selectivity of the polymer-
supported catalyst, when carried out in the presence of water, is
similar to that of unsupported L-proline, usually achieving
diastereomeric ratios (dr) of at least 95/5 (see Table 2, entry 4).
However, when carried out in 100% DMF, the L-proline/poly-
styrene system is superior to that of unsupported L-proline,
achieving a selectivity of 75/25 or better, whereas the reaction
does not proceed with unsupported L-proline, most likely due to
lack of solubility (data not shown). This highlights how the
presence of a polymer scaffold may widen the scope of solvents
that can be used to catalyze reactions using organocatalysts such
To further investigate the results obtained by DLS, the self-
assembled structures were studied using TEM and SEM. For
characterization by TEM, samples of copolymers 5c and 5e at
10 mg/mL were prepared. These were both deposited onto TEM
grids that had previously been treated with oxygen plasma and
subsequently stained with uranyl acetate (UA) or ammonium
7238
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Figure 6. Representative TEM micrographs of 5c in DMF with different vol % of water in DMF: (A) 5 vol %, (B) 7.5 vol %, and (C) 12 vol %, all stained
with 2% AM with histograms showing the size distribution of 5c (Dav is in nm).
molybdate (AM) before imaging. Spherical structures were
revealed in both stains, though large networks/clusters of ag-
gregates were noted with UA, which were an artifact induced by
staining (Figure S9) and individual structures with AM (Figure 6
and Figure S10). Thus, the sizes obtained from the two different
stains cannot be directly compared. However, the size of the
structures found does not seem to be concentration dependent,
as similar sized structures were noted for both 10 and 2 mg/mL
solutions. Figure 6 shows the TEM micrographs of 5c in solu-
tions at different water contents stained with AM, displaying
small differences in particle size.
To confirm these observations, a second copolymer 5e was
Figure 7. Representative TEM micrograph on GO of 5e in DMF with
5 vol % water with histogram showing the size distribution of the
additionally examined by DLS and TEM. Copolymer 5e, differ-
ing only in molecular weight from 5c, showed similar self-
assembly behavior in 5 vol % water solutions. Histograms of
the size distributions are also provided. Though DLS results for
5e (210 nm, PD 0.247) differ from 5c by 50 nm, the particle size
obtained by TEM (Dav = 90 ( 20 nm) does not differ
significantly, confirming the formation of similar self-assembled
aggregates. The difference in noted sizes is most likely due to
copolymer 5e having a longer polystyrene chain.
aggregates.
TEM analysis of these aggregates did not indicate a bilayer-
type structure, and indeed minimal contrast variations inside
particles were found suggesting a solid particle, despite the large
size which is normally indicative of a vesicle-like or compound
structure. To further explore the morphology of these particles,
we imaged them using TEM on graphene oxide (GO) substrates.
GO is an excellent support for the TEM analysis of low contrast
samples as it is nearly electron transparent, and polymer nanos-
tructures can be readily imaged without staining.78 Particles
formed from copolymer 5e in 5 vol % water solutions were
analyzed using GO as a support and solid although relatively
disperse structures (ca. 75 nm) were confirmed. From Figure 7
we can clearly see that some particles are overlaid, and the
resulting contrast could possibly indicate a flattened or disklike
morphology. To explore this further, we performed some dry
TEM tomography experiments (Figure S12). The 3D projection
of the structures after reconstructing the tilt series (À45°
to +45°) into a stack indicated structures are indeed spherical,
Figure 8. SEM micrographs of 5e in a 5 vol % water solution, on an AM
stained TEM grid, which was subsequently coated with gold with the
histogram showing size distribution of 5e.
and an orthogonal slice of the stack through the structures further
confirmed that the particles are solid (Figure S13). We also were
keen to explore these nanostructures in a hydrated state using
cryo-TEM, and although the structures are much larger than
observed by dry-state TEM, the spherical morphology was
confirmed (Figure S14).
The aggregates found from copolymer 5e were further char-
acterized by SEM, and for direct comparison the stained grid
used for TEM analysis was sputter-coated with gold and then
studied by SEM. Indeed, the size obtained by SEM (110 (
24 nm) was comparable to those previously obtained by TEM
7239
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
Table 5. Recyclability of Polymer-Supported L-Proline, 5g, in
the Aldol Reaction
possibility of catalyst recovery and reuse in consecutive cycles
without reducing the activity or selectivity of the catalyst. It was
also evident that the functionalized copolymers self-assembled
into well-defined aggregates in the DMF/water solvent system
used to carry out the aldol reaction. We believe that the unique
microenvironment for L-proline inside these self-assembled
structures positively influences catalysis of the aldol reaction.
Further work will explore the extension of these L-proline systems
for applications in well-defined micellar catalysis.
cycle
24 h conva
anti/syna
eeb
recovery (%)
1
2
3
95
99
94
97/3
95/5
95/5
89
90
88
88
85
90
1
a Determined by H NMR spectroscopy, in CDCl3. b Determined by
chiral HPLC, ChiralPak 1A column (80:20 hexane:IPA).
’ ASSOCIATED CONTENT
(90 ( 20 nm). A representative SEM micrograph is shown in
Figure 8, together with a histogram of the particle size distribution.
Zeta Potential. The zeta potential of the well-defined aggre-
gates, specifically copolymer 5e, was evaluated to explore the
location of the L-proline functionalities, whether they are con-
tained within the core or are present at the surface of the
aggregates. The achieved zeta values at concentrations 2 mg/mL and
10 mg/mL were positive, 26.0 mV and 32.5 mV, respectively. These
values suggest that the L-proline functionalities are not at the
surface of the aggregates, as carboxylic acid groups on the surface
of the aggregates would have resulted in negative zeta potentials.
This finding supports our hypothesis that L-proline catalysis takes
place within the contained microenvironment of the polymeric
aggregates.
S
Supporting Information. Further characterization data
b
for the monomer, CTA, and copolymers and copolymerization
data along with characterization data for the nanostructures.
This material is available free of charge via the Internet at http://
pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: dal28@cam.ac.uk (D.A.L.), r.k.o-reilly@warwick.ac.uk
(R.K.O.).
Recycling. The recycling potential of our polymer-supported
catalytic system was studied using the previously discussed aldol
reaction between 4-nitrobenzaldehyde and cyclohexanone, in the
5 vol % water/DMF solvent mixture. The copolymer 5g
(10.2 kDa, Dh,app = 143 nm, PD = 0.202) employed for the
recycling study was recovered from the reaction mixture via the
dropwise addition of aqueous LiBr (50 mL, 4 wt %), which
quenches the reaction and simultaneously, due to the insolubility
of polystyrene in water, causes the copolymer to precipitate out
of solution. Then, the precipitated polymer was filtered, washed
with additional water, and dried in the vacuum oven at 40 °C.
The aldol product was then extracted from the aqueous solution
in the same manner as previously described. Once dried, the
polymer was redissolved in DMF/5% water, and a second aldol
cycle was carried out. The amount of reactants added to the new
cycle was adjusted according to the amount of polymer recovered
from the previous cycle. This adjustment was done to keep the
aldol reaction at 10 mol % catalyst loading throughout the
different cycles. Table 5 shows the recycling efficiency of our
polymer-supported system. The catalyst was successfully used in
3 cycles without losing significant activity or selectivity, recover-
ing ∼90% of the catalyst in each cycle. Attempts to recycle the
lower molecular weight polymer, 5b, led to less efficient re-
covery; hence, although the molecular weight does not affect the
catalysis rate, it can influence the ease of recyclability.
’ ACKNOWLEDGMENT
The EPSRC and Royal Society are thanked for financial
support, as are the University of Cambridge and the University
of Warwick through the research development fund. D.A.L.
thanks Homerton College and the Department of Chemistry
for support. T.P.S. and T.H.E. thank the NIH-NCRR COBRE
(#P20RR017716) and NIST, U.S. Department of Commerce
(#70NANB7H6178), for financial support. The statements here-
in do not reflect the views of NIH or NIST. The authors also
acknowledge Courtney Ondeck and Dr. Amanda Evans for initial
work in this area and Prof. van Herk (TU Eindhoven) for
assistance and useful discussions with regard to reactivity ratios.
Malvern instruments are acknowledged for assistance with zetasizer
measurements. The GPC equipment used in this research was
obtained through Birmingham Science City: Innovative Uses for
Advanced Materials in the Modern World (West Midlands
Centre for Advanced Materials Project 2), with support from
Advantage West Midlands (AWM) and part funded by the
European Regional Development Fund (ERDF). Finally, we
acknowledge the University of Delaware Keck Microscopy
Facility for use of their TEM and Vitrobot and the Center for
Molecular and Engineering Thermodynamics for use of their
DLS and SLS instruments.
’ REFERENCES
(1) MacMillan, D. W. C. Nature 2008, 455, 304–308.
(2) Gaunt, M. J.; Johansson, C. C. C. Chem. Rev. 2007, 107,
5596–5605.
(3) Raj, M.; Singh, V. K. Chem. Commun. 2009, 6687–6703.
(4) List, B.; Lerner, R. A.; Barbas, C. F. J. Am. Chem. Soc. 2000,
122, 2395–2396.
’ CONCLUSIONS
We report the use of RAFT copolymerization to controllably
incorporate an amino acid (L-proline) containing monomer into
a polystyrene backbone to yield copolymers with tunable catalyst
loading. The organocatalytic activity and selectivity of these
RAFT derived polymers were carefully studied in a model aldol
reaction, resulting in conversions, diastereoselectivities, and
enantioselectivities close to those achieved in unsupported
L-proline catalysis but with higher activity and selectivity at low
loadings when compared to unsupported L-proline. The major
advantage of our polymer-supported catalyst system is the
(5) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. 1971,
10, 496–497.
(6) List, B. Tetrahedron 2002, 58, 5573–5590.
(7) List, B.; Hoang, L.; Martin, H. J. Proc. Natl. Acad. Sci. U. S. A.
2004, 101, 5839–5842.
(8) Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Chem. Rev.
2007, 107, 5471–5569.
7240
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241

Macromolecules
ARTICLE
(9) Movassaghi, M.; Jacobsen, E. N. Science 2002, 298, 1904–1905.
(10) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615–1621.
(11) Notz, W.; Tanaka, F.; Watanabe, S.-i.; Chowdari, N. S.; Turner,
J. M.; Thayumanavan, R.; Barbas, C. F. J. Org. Chem. 2003, 68,
9624–9634.
(46) Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H. Macro-
molecules 2007, 40, 4446–4455.
(47) Goto, A.; Fukuda, T. Prog. Polym. Sci. 2004, 29, 329–385.
(48) Lowe, A. B.; McCormick, C. L. Prog. Polym. Sci. 2007, 32, 2
83–351.
(12) List, B.; Pojarliev, P.; Castello, C. Org. Lett. 2001, 3, 573–575.
(13) Mase, N.; Thayumanavan, R.; Tanaka, F.; Barbas, C. F. Org.
Lett. 2004, 6, 2527–2530.
(14) Bui, T.; Barbas, C. F. Tetrahedron Lett. 2000, 41, 6951–6954.
(15) Hoang, L.; Bahmanyar, S.; Houk, K. N.; List, B. J. Am. Chem.
Soc. 2002, 125, 16–17.
(49) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58,
379–410.
(50) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2006,
59, 669–692.
(51) Moad, G.; Rizzardo, E.; Thang, S. H. Polymer 2008, 49,
1079–1131.
(16) Notz, W.; Tanaka, F.; Barbas, C. F. Acc. Chem. Res. 2004, 37,
580–591.
(52) Perrier, S.; Takolpuckdee, P. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 5347–5393.
(17) Mase, N.; Nakai, Y.; Ohara, N.; Yoda, H.; Takabe, K.; Tanaka,
F.; Barbas, C. F. J. Am. Chem. Soc. 2006, 128, 734–735.
(18) Hayashi, Y.; Sumiya, T.; Takahashi, J.; Gotoh, H.; Urushima, T.;
Shoji, M. Angew. Chem., Int. Ed. 2006, 45, 958–961.
(19) Cobb, A. J. A.; Shaw, D. M.; Longbottom, D. A.; Gold, J. B.; Ley,
S. V. Org. Biomol. Chem. 2005, 3, 84–96.
(53) Willcock, H.; O’Reilly, R. K. Polym. Chem. 2010, 1, 149–157.
(54) Mori, H.; Iwaya, H.; Endo, T. Macromol. Chem. Phys. 2007,
208, 1908–1918.
(55) Mori, H.; Iwaya, H.; Nagai, A.; Endo, T. Chem. Commun. 2005,
4872–4874.
(56) Mori, H.; Kato, I.; Matsuyama, M.; Endo, T. Macromolecules
2008, 41, 5604–5615.
(20) Veverkovꢀa, E.; Strasserovꢀa, J.; Sebesta, R.; Toma, S. Tetrahedron:
Asymmetry 2010, 21, 58–61.
(57) Mori, H.; Matsuyama, M.; Endo, T. Macromol. Chem. Phys.
2009, 210, 217–229.
(58) Evans, A. C.; Skey, J.; Wright, M.; Qu, W.; Ondeck, C.;
Longbottom, D. A.; O’Reilly, R. K. J. Polym. Sci., Part A: Polym. Chem.
2009, 47, 6814–6826.
(59) Evans, A. C.; Lu, A.; Ondeck, C.; Longbottom, D. A.; O’Reilly,
R. K. Macromolecules 2010, 43, 6374–6380.
(60) Blackmond, D.; Armstrong, A.; Coombe, V.; Wells, A. Angew.
Chem., Int. Ed. 2007, 46, 3798–3800.
(21) Bellis, E.; Kokotos, G. Tetrahedron 2005, 61, 8669–8676.
(22) Tang, Z.; Jiang, F.; Cui, X.; Gong, L.-Z.; Mi, A.-Q.; Jiang, Y.-Z.;
Wu, Y.-D. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 5755–5760.
(23) Tang, Z.; Yang, Z.-H.; Chen, X.-H.; Cun, L.-F.; Mi, A.-Q.; Jiang,
Y.-Z.; Gong, L.-Z. J. Am. Chem. Soc. 2005, 127, 9285–9289.
(24) Kristensen, T. E.; Hansen, T. Eur. J. Org. Chem. 2010, 3179–3204.
(25) Benaglia, M.; Celentano, G.; Cozzi, F. Adv. Synth. Catal. 2001,
343, 171–173.
(26) Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi, A.; Celentano, G.
(61) Brogan, A. P.; Dickerson, T. J.; Janda, K. D. Angew. Chem., Int.
Adv. Synth. Catal. 2002, 344, 533–542.
Ed. 2006, 45, 8100–8102.
(27) Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi, A.; Celentano, G.
J. Mol. Catal. A: Chem. 2003, 204, 157–163.
(62) Gruttadauria, M.; Giacalone, F.; Noto, R. Adv. Synth. Catal.
2009, 351, 33–57.
(28) Font, D.; Jimeno, C.; Pericas, M. A. Org. Lett. 2006, 8,
4653–4655.
(63) Liu, Y.-X.; Sun, Y.-N.; Tan, H.-H.; Tao, J.-C. Catal. Lett. 2008,
120, 281–287.
(29) Gruttadauria, M.; Giacalone, F.; Marculescu, A. M.; Lo Meo, P.;
Riela, S.; Noto, R. Eur. J. Org. Chem. 2007, 4688–4698.
(30) Gruttadauria, M.; Salvo, A. M. P.; Giacalone, F.; Agrigento, P.;
Noto, R. Eur. J. Org. Chem. 2009, 2009, 5437–5444.
(31) Liu, Y. X.; Sun, Y. N.; Tan, H. H.; Liu, W.; Tao, J. C.
Tetrahedron: Asymmetry 2007, 18, 2649–2656.
(64) Zotova, N.; Franzke, A.; Armstrong, A.; Blackmond, D. G.
J. Am. Chem. Soc. 2007, 129, 15100–15101.
(65) Dhaon, M. K.; Olsen, R. K.; Ramasamy, K. J. Org. Chem. 1982,
47, 1962–1965.
(66) Neises, B.; Steglich, W. Angew. Chem., Int. Ed. 1978, 17,
522–524.
(32) Font, D.; Bastero, A.; Sayalero, S.; Jimeno, C.; Pericas, M. A.
Org. Lett. 2007, 9, 1943–1946.
(67) Ma, Q.; Wooley, K. L. J. Polym. Sci., Part A: Polym. Chem. 2000,
38, 4805–4820.
(33) Miao, W. S.; Chan, T. H. Adv. Synth. Catal. 2006, 348, 1711–1718.
(34) Aprile, C.; Giacalone, F.; Gruttadauria, M.; Marculescu, A. M.;
Noto, R.; Revell, J. D.; Wennemers, H. Green Chem. 2007, 9, 1328–1334.
(35) Bellis, E.; Kokotos, G. J. Mol. Catal. A: Chem. 2005, 241,
166–174.
(68) Benaglia, M.; Chen, M.; Chong, Y. K.; Moad, G.; Rizzardo, E.;
Thang, S. H. Macromolecules 2009, 42, 9384–9386.
(69) Benaglia, M.; Chiefari, J.; Chong, Y. K.; Moad, G.; Rizzardo, E.;
Thang, S. H. J. Am. Chem. Soc. 2009, 131, 6914–6915.
(70) Skey, J.; O’Reilly, R. K. Chem. Commun. 2008, 4183–4185.
(71) Fineman, M.; Ross, S. D. J. Polym. Sci., Part A: Polym. Chem.
1950, 5, 259–262.
(36) Sanda, F.; Endo, T. Macromol. Chem. Phys. 1999, 200, 2651–
2661.
(37) Lokitz, B. S.; Stempka, J. E.; York, A. W.; Li, Y.; Goel, H. K.;
Bishop, G. R.; McCormick, C. L. Aust. J. Chem. 2006, 59, 749–754.
(38) Skey, J.; Hansell, C. F.; O’Reilly, R. K. Macromolecules 2010,
43, 1309–1318.
(72) van Herk, A. M. J. Chem. Educ. 1995, 72, 138–140.
(73) Gallow, K. C.; Jhon, Y. K.; Tang, W.; Genzer, J.; Loo, Y.-L.
J. Polym. Sci., Part B: Polym. Phys. 2011, 49, 629–637.
(74) Seno, K.-I.; Tsujimoto, I.; Kanaoka, S.; Aoshima, S. J. Polym. Sci.,
Part A: Polym. Chem. 2008, 46, 6444–6454.
(39) Skey, J.; O’Reilly, R. K. J. Polym. Sci., Part A: Polym. Chem. 2008,
46, 3690–3702.
(75) Font, D.; Sayalero, S.; Bastero, A.; Jimeno, C.; Pericas, M. A.
Org. Lett. 2007, 10, 337–340.
(76) Zotova, N.; Franzke, A.; Armstrong, A.; Blackmond, D. G.
J. Am. Chem. Soc. 2007, 129, 15100–15101.
(40) Casolaro, M.; Bottari, S.; Cappelli, A.; Mendichi, R.; Ito, Y.
Biomacromolecules 2004, 5, 1325–1332.
(41) Kristensen, T. E.; Vestli, K.; Fredriksen, K. A.; Hansen, F. K.;
Hansen, T. Org. Lett. 2009, 11, 2968–2971.
(42) Kristensen, T. E.; Vestli, K.; Jakobsen, M. G.; Hansen, F. K.;
Hansen, T. J. Org. Chem. 2010, 75, 1620–1629.
(77) Zhang, L.; Eisenberg, A. J. Am. Chem. Soc. 1996, 118, 3168–3181.
(78) Williams, P. E.; Moughton, A. O.; Patterson, J. P.; Khodabakhsh,
S.; O’Reilly, R. K. Polym. Chem. 2011, 2, 720–729.
(43) Monterio, M. Macromolecules 2010, 43, 1159–1168.
(44) Barner-Kowollik, C.; Perrier, S. J. Polym. Sci., Part A: Polym.
Chem. 2008, 46, 5715–5723.
(45) Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le,
T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.;
Rizzardo, E.; Thang, S. H. Macromolecules 1998, 31, 5559–5562.
7241
dx.doi.org/10.1021/ma201256m |Macromolecules 2011, 44, 7233–7241