Carbon-carbon bond forming reactions with substrates absorbed non-covalently on a cellulose chromatography paper support

Article abstract of DOI:10.1039/b611824k

Reactions and purifications, including carbon-carbon bond forming reactions, can be carried out on a cellulose support on which the substrates are non-covalently absorbed. The Royal Society of Chemistry.

Full text of DOI:10.1039/b611824k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Carbon–carbon bond forming reactions with substrates absorbed non-
covalently on a cellulose chromatography paper support
Jonathon Hacon,a Amanda Morris,b Michael J. Johnston,c Stephen E. Shanahan,c Mike D. Barker,c
Graham G. A. Inglisc and Simon J. F. Macdonald*c
Received (in Bloomington, IN, USA) 16th August 2006, Accepted 11th October 2006
First published as an Advance Article on the web 6th November 2006
DOI: 10.1039/b611824k
forming reactions. We also detail new washing procedures to effect
purification on the tile and describe processing multiple substrates
on the same tile with no inter-mixing of the products.
Reactions and purifications, including carbon–carbon bond
forming reactions, can be carried out on a cellulose support on
which the substrates are non-covalently absorbed.
Ring closing metatheses have recently become a widespread and
popular methodology for forming carbon–carbon double bonds.7
An interesting recent development has been the use of metathesis
catalysts in ionic liquids to aid product isolation and the removal
of the catalyst.8 Using a standard reaction,8 we have demonstrated
an olefin metathesis on a cellulose support (Scheme 1{).
The chemical production of libraries and arrays where substrates
are attached to a solid support have found widespread use in the
last decade.1 Variations on this theme where reagents, catalysts, or
scavenger reagents are attached to a support have also become
powerful and valuable tools.2 Practically all of these techniques
require the covalent attachment of the substrate or reagent to the
support, irrespective of the nature of the support. Notable excep-
tions are elegant studies by Williams3 and Field4 where synthesis
and purification (and in some cases biological screening) are
carried out on TLC plates or on silica packed in a flash column.4
We have recently discovered that it is possible to carry out
chemical transformations and produce arrays of compounds
typically in 1–10 mg quantities, where the molecules are only
absorbed and not covalently bound onto a support such as cellulose
based chromatography paper.5 This stands in contrast to previous
reports where the substrates are covalently attached to derivatised
cellulose.6 We also demonstrated that it is possible to purify
products non-covalently absorbed on the cellulose paper by
washing with solvents which selectively remove impurities (such as
excess reagent or unreacted starting material).
Tosyl chloride (6 mg) (mp 67–69 uC) was loaded as a melt onto
a 1 cm2 tile of Whatman 17Chr chromatography paper followed
by neat diallylamine. After 10 min, the tile was washed by vigorous
shaking in dilute sodium bicarbonate solution (to remove
unreacted tosyl chloride), water, dilute acid (to remove unreacted
amine) and further water and then dried in vacuo.{ After this
purification the second reaction, the metathesis, was carried out. A
catalytic (2.5 mol%) amount of the Grubbs9 catalyst 3 was
dissolved in the commercially available ionic liquid 1-butyl-3-
imidazolium hexafluorophosphate 4 (60 mL) and then loaded onto
the tile supporting the sulfonamide 2. The tile was then sealed in a
vial and left at room temperature for 3 d.§ After a single extraction
by vigorous shaking in ether," the pyrroline product 5 (1.1 mg)
was obtained in 29% unoptimised yield in good purity by
1H-NMR with no evidence for either the ionic liquid or
contamination with the catalyst. As expected,10 there was no
indication of dissolution or degradation of the cellulose in the ionic
liquid. (We have also found that Diels–Alder reactions mediated
by Lewis acids in ionic liquids are feasible on a cellulose support.
Further details will be reported elsewhere.)
The cellulose paper (typically cut into 10 6 10 mm ‘‘tiles’’) can
be regarded as an ‘‘immobilised solvent’’. Thus washing sequences
that purify a crude reaction mixture dissolved in an organic solvent
(such as washing out excess amine with dilute aqueous acid after
an amide coupling), have been shown to be equally applicable to
the washing of an ‘‘immobilised solvent’’—in this case the
cellulose. The advantages of the cellulose include its flexibility,
simplicity of handling and processing large numbers of tiles,
exceptionally low cost (100 6 1 cm2 tiles cost 20 p), ease of
‘‘encoding’’ (simply by writing on the tile with a pencil or by
cutting the tiles into different shapes) and ease of analysing the
contents of a tile (by snipping off a small corner of the tile and
placing the fragment in acetonitrile prior to LC-MS analysis). The
potential of this technique appears considerable.5
We next explored a Wadsworth–Emmons reaction on the
cellulose support. As cellulose is composed of polysaccharides
derived from glucose, we planned to use sodium ethoxide as base
for this reaction. If an equilibrium is set up between the ethoxide
and alkoxides of the saccharide, in principle both may act as a
base. Triethylphosphonoacetate was loaded onto a tile followed by
an excess of benzaldehyde (10 equiv.) (Scheme 2{).
The sodium ethoxide in ethanol (ca. 7 equiv.) was then applied
and the tile allowed to stand at room temperature in a sealed vial
for 1 h. The tile was then washed sequentially by vigorous shaking
with sodium bisulfite solution (to remove unreacted aldehyde),
water, 8% sodium hydrogen carbonate solution, water and dried.
Extraction of the product in the usual fashion gave a good yield
Having previously described the conversion of amines into
amides, sulfonamides and carbamates and demonstrated some
simple deprotections,5 we describe here carbon–carbon bond
aFormerly a pupil of Meridian School, Royston, Hertfordshire, UK
bFormerly a pupil of Bishops’ Hatfield Girls’ School, Hatfield,
Hertfordshire, UK
1
(1.9 mg, 82%) of pure E-cinnamate as determined by H-NMR.
One of the advantages of this cellulose supported technology is
the ease with which the paper can be cut into different shapes. One
example of this is described here where different substrates were
cMedicinal Chemistry, ri CEDD, Gunnels Wood Road, Stevenage, UK
SG1 2NY. E-mail: simon.jf.macdonald@gsk.com
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 625–627 | 625

View Article Online
Scheme 1
of E-ethyl cinnamate (1.8 mg, 78%), E-ethyl 4-methylcinnamate
(1.5 mg, 60%) and E-ethyl 4-methoxycinnamate (2.3 mg, 85%)
were obtained in high purity and with no cross-contamination as
1
determined by H-NMR.
Purification of the products for the Wadsworth–Emmons
reactions involved washing the tiles or comb with aqueous sodium
bisulfite and sodium hydrogen carbonate. There are other
numerous potential washing proceduresI for the selective removal
of products, reagents or by-products aside from those described
here and elsewhere.5 One option is the use of ‘‘phase switching’’
methodologies where following its use in a reaction, a reagent is
degraded allowing its removal by a washing procedure.11 A further
option is the use of fluorous reagents and/or solvents as described
by Curran et al.12 which would broaden the range of reactions that
might be used with this technology. We illustrate here the use and
Scheme 2
used on the same tile which had been cut into the shape of a
‘‘comb’’ (Fig. 1). This has advantages in streamlining the work-up
procedures.
As an example, we carried out the Wadsworth–Emmons
reaction above using three different benzaldehydes as substrates
on the same tile (Scheme 3{).
Each ‘‘tooth’’ of the comb supported a different substrate—
benzaldehyde on tooth A, tolualdehyde on tooth B and para-
methoxybenzaldehyde on tooth C. In each case the reagents were
applied to the centre of each tooth. There is no contamination
between each tooth as the substrates and reagents remain confined
to the tooth region with no or very little spread to the ‘‘backbone’’
of the comb. The multiple reactions and work-ups were carried out
as before with the whole comb being washed to remove unreacted
aldehyde and phosphonate by-products. Finally, before extraction
of the products with ethyl acetate, the teeth were cut off (at the top
of the tooth) and placed in separate vials. In this way, good yields
removal of
a fluorous tin reagent used in the selective
monotosylation of phenyl-1,2-ethanediol 12 (Scheme 4).13
The fluorous tin oxide 13 was dissolved in perfluorohexanes and
loaded onto a tile which was then dried. The substrate and other
reagents were loaded in dichloromethane. Following reaction, the
tile was dried and washed with dilute hydrochloric acid and
perfluorohexanes prior to extraction of the product with
dichloromethane. The product was not contaminated with any
tin reagents as determined by 19F-NMR. Alternatively the fluorous
tin reagent could be selectively recovered in good yield by washing
with perfluorohexanes.
Further applications of this technology that might be envisaged
including derivation of the cellulose to provide a different
environment for chemical reactions or alternative washing
protocols. This is analogous to the derivation of resins or silica
with, for example, long perfluorinated alkyl chains to provide a
fluorous environment suitable for fluorous substrates.12
In summary, we have described carbon–carbon bond forming
reactions where the substrates are absorbed non-covalently on a
cellulose support. We have also demonstrated that multiple
substrates can be processed on a suitably shaped comb without
cross-contamination. Finally we have demonstrated the use and
removal of a fluorous reagent in a cellulose environment. We hope
these and previous5 examples demonstrate the promise of this
powerful but cheap technology which we have found exceptionally
Fig. 1 Illustration and dimensions of the ‘‘comb’’. Each ‘‘tooth’’ of the
comb (labelled A, B and C) supports a different reaction but the work up
procedures can be carried out on the whole comb before cutting off each
tooth and extracting the product.
Scheme 3
Scheme 4
626 | Chem. Commun., 2007, 625–627
This journal is ß The Royal Society of Chemistry 2007

View Article Online
{ At this point the sulfonamide 2 (5.3 mg) can be isolated from the tile (by
boiling in ethyl acetate for 3 min) in good purity as determined by 1H-
NMR.
simple to use. In fact all the practical work described here and
previously5 was carried out by 16 year old school students.
§ In an attempt to increase the rate of the reaction, it was heated in a
microwave; however the tile was pyrolised to an ash within seconds! It is
known that microwaves heat ionic liquids with unusual efficiency.14
" For a good recovery of product from the tile multiple extractions with
ether are required. From previous work, hot ethyl acetate is an effective
extraction solvent, but in this case ether was preferred to avoid
simultaneous extraction of ionic liquid.
I These include aqueous ethanolamine solutions to remove boronic acids,
non-polar organic solvents such as hexane to remove highly lipophilic
substances and aqueous fluoride solutions to remove tin reagents.
Notes and references
{ N-Tosyl-3-pyrroline (5). Tosyl chloride (6 mg, 0.03 mmol) was melted and
applied using a Gilson pipette onto the centre of a 1 cm2 tile of Whatman
17Chr chromatography paper. Diallylamine (45 mL, 0.36 mmol) was added
in a similar fashion. After standing in a sealed vial for 10 min, the tile was
then vigorously shaken in 8% sodium hydrogencarbonate solution (5 mL)
for 30 s, water (5 mL) for 1 min, 0.5 M hydrochloric acid (5 mL) for 1 min
and finally water (5 mL) for 1 min. After drying in vacuo, a solution of
Grubb’s catalyst (benzylidene-bis(triphenylphosphine)ruthenium dichlor-
ide) (0.3 mg, 10 mol%, 0.003 mmol) in 1-butyl-3-methylimidazolium
hexafluorophosphate) (60 mL) (requires warming to obtain a solution) was
applied to the centre of the tile which was then allowed to stand in a sealed
vial for 3 d at room temperature. The product was then extracted with
diethyl ether (5 mL) for 10 min. The ether was evaporated under a stream
of dry nitrogen followed by drying in vacuo, affording pure pyrroline 7 by
1H-NMR (1.9 mg, 29%). 1H-NMR (CDCl3, 250 MHz) d 7.66 (d, J = 8 Hz,
2H), 7.25 (d, J = 8 Hz, 2H), 5.58 (s, 2H), 4.04 (s, 4H), 2.37 (s, 3H).
E-Ethyl cinnamate (7) Using a Gilson pipette, triethylphosphonacetate
(3.14 mL, 0.016 mmol) and benzaldehyde (13.4 mL, 0.16 mmol) were
sequentially loaded onto the centre of a 1 cm2 tile of Whatman 17Chr
chromatography paper. Finally, sodium ethoxide in ethanol (21 wt%,
33.5 mL, 0.1 mmol) was added and the tile then sealed in a vial at room
temperature for 1 h. The tile was then vigorously shaken in saturated
sodium bisulfite solution (freshly prepared, 5 mL) for 5 min, then water
(10 mL) for 10 min, 8% sodium hydrogencarbonate solution (5 mL) for
5 min and finally water (10 mL) for 10 min. The tile was then dried at high
vacuum and finally extracted with boiling ethyl acetate (5 mL). Solvent
removal by evaporation under a stream of dry nitrogen followed by drying
in vacuo afforded pure cinnamate by H-NMR (1.9 mg, 82%). H-NMR
(CDCl3, 250 MHz) d 7.75 (d, J = 16 Hz, 1H), 7.57 (m, 2H), 7.43 (m, 3H),
6.50 (d, J = 16 Hz, 1H), 4.32 (quartet, J = 7 Hz, 2H), 1.41 (t, J = 7 Hz, 3H).
The ‘‘comb’’ array of cinnamates: E-ethyl cinnamate, E-ethyl 4-methyl-
cinnamate, E-ethyl 4-methoxycinnamate. To each tooth of the comb, was
added triethylphosphonoacetate (3.11 mL, 0.013 mmol). To tooth A was
then added benzaldehyde (13.4 mL, 0.13 mmol), tooth B p-tolualdehyde
(15.6 mL, 0.13 mmol) and tooth C p-methoxybenzaldehyde (16.1 mL,
0.13 mmol). To each tooth was then added sodium ethoxide (33.5 mL of a
21% solution in ethanol). The comb was then placed in an ethanol rich
atmosphere for 30 min. The comb was then worked-up as before except the
teeth were cut off at the top of the tooth before extraction of the product.
The products obtained in good purity are E-ethyl cinnamate (1.8 mg, 78%),
E-ethyl 4-methylcinnamate (1.5 mg, 60%) and E-ethyl 4-methoxycinnamate
(2.3 mg, 85%).
1 B. A. Bunin, The Combinatorial Index, Academic Press, New York,
1998.
2 S. V. Ley and I. R. Baxendale, Nat. Rev. Drug Discovery, 2002, 1, 573.
3 L. Williams, Chem. Commun., 2000, 435; L. Williams and O. Bergersen,
J. Planar Chromatogr.–Mod. TLC, 2001, 14, 318; L. Williams, R. Sjovik
and M. L. Falk-Pedersen, J. Planar Chromatogr.–Mod. TLC, 2004, 17,
244.
4 B. Mukhopadhyay, P. Cura, K. P. R. Kartha, C. H. Botting and
R. A. Field, Org. Biomol. Chem., 2005, 3, 3468; B. Mukhopadhyay,
S. V. Maurer, N. Rudolph, R. M. van Well, D. A. Russell and
R. A. Field, J. Org. Chem., 2005, 70, 9059.
5 S. E. Shanahan, D. D. Byrne, G. G. A. Inglis, M. Alum and
S. J. F. Macdonald, Chem. Commun., 2002, 2554.
6 R. Frank, Tetrahedron, 1992, 48, 9217; M. Lebl, Biopolymers, 1998, 47,
397; U. Reineke, R. Volkmer-Engert and J. Schneider-Mergener, Curr.
Opin. Biotechnol., 2001, 12, 59.
7 A. Furstner, Angew. Chem., Int. Ed., 2000, 39, 3012.
8 D. Semeril, H. Olivier-Bourbigou, C. Bruneau and P. H. Dixneuf,
Chem. Commun., 2002, 146; R. C. Buijsman, E. van Vuuren and
J. G. Sterrenburg, Org. Lett., 2001, 3, 3758.
9 T. Trnka and R. H. Grubbs, Acc. Chem. Res., 2001, 34, 18.
10 R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J. Am.
Chem. Soc., 2002, 124, 4974.
11 An example is the use of di-tert-butylazodicarboxylate in the Mitsunobu
reaction. Following reaction, treatment with acid allows simple removal
of the reagent by washing.G. W. Starkey, J. J. Parlow and D. L. Flynn,
Bioorg. Med. Chem. Lett., 1998, 8, 2385; M. Kiankarima, R. Lowe,
J. R. McCarthy and J. P. Whitten, Tetrahedron Lett., 1999, 40, 4497.
12 Fluorous Chemistry: Tetrahedron Symposium-in-Print Number 91
Tetrahedron, ed. J. A. Gladysz and D. P. Curran, 2002 58, pp. 3823–
4131.
1
1
13 B. Bucher and D. P. Curran, Tetrahedron Lett., 2000, 41, 9617.
14 R. S. Varma and V. V. Namboodiri, Chem. Commun., 2001, 643–644.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 625–627 | 627