Organogel media for on-bead screening in combinatorial catalysis

Article abstract of DOI:10.1016/j.tetlet.2005.03.125

Poly(vinylidene fluoride) (PVdF) forms thermoreversible gels with a number of dipolar aprotic solvents. Gels were prepared containing chromogenic substrates, subject to transformation by polymer-bound catalysts. When the catalysts were mixed with inactive beads and applied to the surface of the gels, the active beads were identifiable through colour changes. Active beads could also be visualised by thermographic imaging. These methods hold promise for catalyst discovery from split-and-mix combinatorial libraries.

Full text of DOI:10.1016/j.tetlet.2005.03.125

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 3923–3926
Organogel media for on-bead screening in combinatorial catalysis
Karl-Jonas Johansson,a Marc R. M. Andreae,a
Albrecht Berkesselb and Anthony P. Davisa,*
aSchool of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, UK
bInstitut fu¨r Organische Chemie der Universita¨t zu Ko¨ln, Greinstraße 4, D-50939 Cologne, Germany
Received 1 December 2004; revised 25 February 2005; accepted 17 March 2005
Available online 12 April 2005
Abstract—Poly(vinylidene fluoride) (PVdF) forms thermoreversible gels with a number of dipolar aprotic solvents. Gels were
prepared containing chromogenic substrates, subject to transformation by polymer-bound catalysts. When the catalysts were mixed
with inactive beads and applied to the surface of the gels, the active beads were identifiable through colour changes. Active
beads could also be visualised by thermographic imaging. These methods hold promise for catalyst discovery from split-and-mix
combinatorial libraries.
Ó 2005 Elsevier Ltd. All rights reserved.
High-throughput or combinatorial methodologies are
gaining importance in catalysis research.1 Parallel syn-
thesis and screening accelerates catalyst development
and allows consideration of structures which might be
ignored in the standard, iterative protocol. Most of this
work involves spatially separated catalysts, prepared
and studied in discrete reaction vessels. However, an
interesting and complementary alternative is provided
by Ôsplit-and-mixÕ combinatorial chemistry (SMCC).
Split-and-mix solid phase synthesis allows the prepar-
ation of very large libraries of resin-bound catalysts,
among which the prospects of finding exceptional and
unexpected properties should be high. Unfortunately,
locating these library members is nontrivial. The com-
pounds must be tested while still on the bead and the as-
say must be highly parallel. Realistically, it is necessary
to screen visually under conditions that allow picking of
active beads. Screening for catalysis is complicated by
product diffusion away from catalytic centres. A visual
signal (e.g., a colour change) may thus be delocalised,
preventing the identification of active beads. Solutions
to this problem have been published,2 but all have limi-
tations and improved/complementary methods are still
sought.
suspended in or placed on the gel, reactants diffuse in,
and the reaction takes place on active beads. Because
the gel limits diffusion over long distances, the products
can be confined to the neighbourhood of the bead. The
reaction must be chromogenic, or otherwise visually
detectable, but the method is otherwise fairly versatile.
The gel is clearly a critical component of the procedure.
In one variant, due to Miller, it is synthesised by poly-
merisation of specially designed precursors around the
library beads.2d Work from our group, in contrast, has
exploited commercially available agarose, which is sim-
ply dissolved in hot water and allowed to set.3
The agarose-based method is experimentally convenient,
but limited to the aqueous solvent systems required to
dissolve the polysaccharide. The scope of our strategy
could be increased significantly if suitable nonaqueous
gels (organogels) could be identified. We now report a
simple thermoreversible organogel system, which serves
this purpose, extending the methodology to polar non-
aqueous solvents.
A gel for use with our screening method should fulfil a
number of criteria. Firstly the gelling agent should be
readily accessible. Secondly, it should form gels with a
range of solvents. Thirdly these gels should have suitable
physical properties, being capable of supporting the
library and allowing bead picking under a microscope.
Fourthly, it should be unlikely to interfere with
reactions, for example, by diffusing into a bead and
interacting with catalytic centres. Initially we considered
One approach which shows promise is the use of
gel-based reaction media.2d,3 The catalyst beads are
*
Corresponding author. Tel.: +44 0117 9546334; fax: +44 0117
9298611; e-mail: anthony.davis@bristol.ac.uk
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.03.125

3924
K.-J. Johansson et al. / Tetrahedron Letters 46 (2005) 3923–3926
small-molecule gelling agents, on which there is exten-
sive literature,4 but the above criteria proved difficult
to match. For example, gelator 15 possesses useful prop-
erties but must be prepared from expensive 1R,2R-trans-
cyclohexanediamine while diamide 2 forms gels, which
are too soft.6 Moreover, although these compounds
are prepared in one step, their purification can be
troublesome.
NO2
NO2
O
OH
NO2
+
C
MeNO2
catalyst
Green
Yellow
Yellow
4
7
8
Scheme 2. Catalyst = 1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2-a]pyrim-
idine, polymer-bound (Aldrich).
O
O
C14H29
C11H23
NH
NH
aminomethyl polystyrene) in the ratio 1:10. Typically,
the aldehyde 4 (100 mg) and PVdF (1 g) were added to
acetonitrile (9 mL) and heated to 100 °C for ca. 30 min
in a pressure tube with stirring. The liquid was decanted
into a flat-bottomed vessel,10 and ethyl cyanoacetate
(1 mL), malononitrile (0.2 mL) or nitromethane (1 mL)
was added with stirring. When the gel had set, the
appropriate model library was spread on the surface
and observed under a microscope. The results are illus-
trated in Figures 1–3. For all three reactions a clear dis-
tinction could be observed between catalytic beads and
the inactive controls. In the case of reactions A and B,
the active beads clearly took on the colour of the prod-
ucts 5 and 6, respectively. In the case of reaction C, the
active beads turned brown, possibly suggesting further
reaction/decomposition. In no case could we see transfer
of colour between active and inactive beads, even
though they were in close contact. Beads could be re-
moved from the gel without difficulty. These findings
demonstrate that the gel can deliver substrates (and pre-
sumably solvent) to the beads, and that it provides the
right physical environment for the screening experiment.
Material transfer between beads is clearly slow, as found
in our earlier work.3
NH
O
NH
O
F F
n
C11H23
C14H29
1
2
3
We therefore turned to an alternative approach, the use
of nonpolar polymers.7 A literature search suggested
two possibilities, poly(vinyl chloride) (PVC) and poly-
vinylidene fluoride (PVdF, 3). PVC forms gels with a
range of solvents, but these materials acquire sticky sur-
faces which interfered with bead picking and seemed to
block substrate transfer. PVdF, however, proved more
accommodating. It is reported to dissolve in ketone or
ester solvents at high temperatures and set to gels on
cooling.8 In our hands, it also succeeded with a number
of other solvents. At the level of 10% w/v, it formed ther-
moreversible gels with acetone, dioxane, acetonitrile,
nitromethane and THF.9 The gels were physically resil-
ient without being sticky, and seemed good candidates
for the screening system. To test their suitability, we em-
ployed three reactions chosen (a) for their chromogenic-
ity and (b) for their susceptibility to catalysis by
commercially-available resins. 1-Pyrenecarboxaldehyde
4 is coloured light green. Knoevenagel condensations
with malononitrile (reaction A) and ethyl cyanoacetate
(reaction B) give products 5 (red) and 6 (yellow). Both
reactions could be catalysed by polymer-bound piper-
azine (Scheme 1), proceeding smoothly in the absence
of PVdF with toluene as solvent. The Henry reaction
of 4 with nitromethane (reaction C) could be catalysed
by a polymer-bound guanidine base, giving 7 with 8 as
a side product (Scheme 2). Both 7 and 8 are yellow.
Straightforward visual screening is of course limited by
the need for chromogenic or fluorogenic substrates.
An elegant alternative is the use of thermal imaging,
as pioneered by Taylor and Morken.2a,11 In its original
For tests on gelated reaction mixtures, model Ôcatalyst
librariesÕ were prepared by mixing the piperazine or
guanidine beads with inactive controls (acetylated
O
NC
CN
O
NC
B
OEt
CN
A
CH2(CN)2
catalyst
CO2Et
catalyst
Figure 1. PVdF/MeCN gel containing 4 and malononitrile (reaction
A), overlaid with model Ôcatalyst libraryÕ composed of resin-bound
piperazine and inactive controls (1:10). Micrograph taken 2 h after
application of beads. Four active beads (red) are clearly visible. Five
inactive controls are also present. The beads appear in groups of 3 on
the gel surface.
Green
Red
Yellow
5
4
6
Scheme 1. Catalyst = piperazinomethyl polystyrene (Nova biochem).

K.-J. Johansson et al. / Tetrahedron Letters 46 (2005) 3923–3926
3925
Figure 4. Thermographic image of polymer-bound DMAP beads on
PVdF/MeCN gel containing PhCH2OH, Et3N and Ac2O. The beads
appear yellow, indicating temperatures of up to 0.7 °C above
background.
Figure 2. PVdF/MeCN gel containing 4 and ethyl cyanoacetate
(reaction B), overlaid with beads as for Figure 1. Micrograph taken
24 h after application of beads.
be used in visual screening of resin beads for catalytic
activity. Preliminary results suggest that the gels may
also facilitate screening by IR thermography. In future
work we hope to apply this methodology to the discov-
ery of novel molecular catalysts through combinatorial
chemistry.
Acknowledgements
This research was supported by the European Commis-
sion (Human Potential Network contract HPRN-CT-
2000-00014) and the EPSRC (GR/R42757/01).
Figure 3. PVdF/MeCN gel containing 4 and nitromethane (reaction
C), overlaid with model Ôcatalyst libraryÕ composed of resin-bound
1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2-a]pyrimidine and inactive con-
trols (1:10). Micrograph taken 2 h after application of beads.
References and notes
1. Ding, K. L.; Du, H. F.; Yuan, Y.; Long, J. Chem. Eur. J.
2004, 10, 2872; Berkessel, A. Curr. Opin. Chem. Biol. 2003,
7, 409; de Vries, J. G.; de Vries, A. H. M. Eur. J. Org.
Chem. 2003, 799; Gennari, C.; Piarulli, U. Chem. Rev.
2003, 103, 3071; Reetz, M. T. Angew. Chem., Int. Ed. 2002,
41, 1335; Shimizu, K. D.; Snapper, M. L.; Hoveyda, A. H.
Chem. Eur. J. 1998, 4, 1885.
2. (a) Taylor, S. J.; Morken, J. P. Science 1998, 280, 267; (b)
Berkessel, A.; Herault, D. A. Angew. Chem., Int. Ed. 1999,
38, 102; (c) Copeland, G. T.; Miller, S. J. J. Am. Chem.
Soc. 1999, 121, 4306; (d) Harris, R. F.; Nation, A. J.;
Copeland, G. T.; Miller, S. J. J. Am. Chem. Soc. 2000, 122,
11270; (e) De Muynck, H.; Madder, A.; Farcy, N.; De
version, the method involved flotation of a library on a
chlorinated (i.e., relatively dense) solvent containing re-
agent and substrate. Catalytic beads were raised in tem-
perature by the heat of reaction, and were detected by an
IR camera. We could see several advantages to perform-
ing this experiment on PVdF gels. Firstly, the beads
would be immobilised and easily selected. Secondly,
there would be no need for an especially dense solvent.
Thirdly, heat conduction away from the beads may be
reduced, leading to greater sensitivity. As a preliminary
trial, we added triethylamine (16 lL) and benzyl alcohol
(10 lL) to molten PVdF/MeCN gel (100 lL). The mix-
ture was spread on glass and allowed to set. Acetic anhy-
dride (10 lL) was applied to the surface, and appeared
to be absorbed by the gel in ꢀ3 min. Polymer-bound
DMAP12 was spread on the surface, which was then im-
aged with an IR camera.13 Initially the beads could not
be detected, but after ca. 30 s they appeared against the
background (Fig. 4). The imaging data implied temper-
ature rises of up to 0.7 °C, due to the exothermic
DMAP-catalysed acetylation of the alcohol. Control
experiments with inactive beads showed no detectable
temperature variations.
¨
´
´
Clercq, P. J.; Perez-Payan, M. N.; Ohberg, L. M.; Davis,
A. P. Angew. Chem., Int. Ed. 2000, 39, 145; (f) Krattiger,
P.; McCarthy, C.; Pfaltz, A.; Wennemers, H. Angew.
Chem., Int. Ed. 2003, 42, 1722; (g) Clouet, A.; Darbre, T.;
Reymond, J.-L. Angew. Chem., Int. Ed. 2004, 43, 4612.
3. Muller, M.; Mathers, T. W.; Davis, A. P. Angew. Chem.,
¨
Int. Ed. 2001, 40, 3813.
4. Terech, P.; Weiss, R. G. Chem. Rev. 1997, 97, 3133; van
Esch, J. H.; Feringa, B. L. Angew. Chem., Int. Ed. 2000,
39, 2263.
5. Hanabusa, K.; Yamada, M.; Kimura, M.; Shirai, H.
Angew. Chem., Int. Ed. Engl. 1996, 35, 1949.
6. Tomioka, K.; Sumiyoshi, T.; Narui, S.; Nagaoka, Y.; Iida,
A.; Miwa, Y.; Taga, T.; Nakano, M.; Handa, T. J. Am.
Chem. Soc. 2001, 123, 11817.
In conclusion, we have found that PVdF forms gels with
a range of polar aprotic solvents and that these gels may
7. Guenet, J. M. Thermoreversible Gelation of Polymers and
Biopolymers; Academic: New York, 1992; Osada, Y.;

3926
K.-J. Johansson et al. / Tetrahedron Letters 46 (2005) 3923–3926
Gong, J. P.; Tanaka, Y. J. Macromol. Sci.—Polym. Rev.
2004, C44, 87.
11. Thermal imaging has also been applied to reaction arrays.
For examples, see: Reetz, M. T.; Becker, M. H.; Kuhling,
K. M.; Holzwarth, A. Angew. Chem., Int. Ed. 1998, 37,
2647; Holzwarth, A.; Schmidt, P. W.; Maier, W. E.
Angew. Chem., Int. Ed. 1998, 37, 2644; Reetz, M. T.;
8. Cho, J. W.; Song, H. Y.; Kim, S. Y. Polymer 1993, 34,
1024; Tazaki, M.; Wada, R.; Okabe, M.; Homma, T. J.
Appl. Polym. Sci. 1997, 65, 1517; Kim, B. S.; Baek, S. T.;
Song, K. W.; Park, I. H.; Lee, J. O.; Nemoto, N. J.
Macromol. Sci.—Phys. 2004, B43, 741.
9. Typical procedure: PVF pellets (Aldrich, m.w. 530,000,
cat. no. 34,707-8; 500 mg) were placed in a pressure tube
with MeCN (5 mL) and heated at 100 °C for 30 min with
stirring. The resulting homogeneous solution was cooled
to room temperature and formed a clear gel within
ꢀ5 min. The gel could be melted and reset as desired.
10. These experiments employed shallow beakers, ꢀ5 cm in
diameter, with tight fitting lids to prevent evaporation of
solvent.
Becker, M. H.; Liebl, M.; Furstner, A. Angew. Chem., Int.
¨
Ed. 2000, 39, 1236; Berkessel, A.; Ashkenazi, E.; Andreae,
M. R. M. Appl. Catal. A: General 2003, 254, 27.
12. 4-(N-Benzyl-N-methylamino)-pyridine, polymer bound;
Aldrich cat. no. 35,988-2.
13. IR-thermography was performed using a ThermaCAM
SC 1000 camera (Inframetrics) with ThermaGram 95Ò
(Vers. 1.60.07) and DynaMite 97Ò (Vers. 1.01) software
from Thermoteknix Systems Ltd., Cambridge, UK. No
emissivity correction was applied. Detectable wavelength
range of the camera is 3.4–5.4 lm.