Amide bond formation with a new fluorous carbodiimide: Separation by reverse fluorous solid-phase extraction

Article abstract of DOI:10.1021/ol701631m

A new fluorous carbodiimide is introduced along with a convenient procedure for amide coupling reactions. Reactions of acids and amines under standard conditions for carbodiimide couplings, followed by simple reverse fluorous solid-phase extraction (FSPE) over standard silica gel, provide the target amide products in good yields and purities. The use of HFE-7100 as a fluorous solvent is crucial for the success of the reverse FSPE.

Full text of DOI:10.1021/ol701631m

ORGANIC
LETTERS
2007
Vol. 9, No. 21
4167-4170
Amide Bond Formation with a New
Fluorous Carbodiimide: Separation by
Reverse Fluorous Solid-Phase
Extraction
Carlos del Pozo,† Adam I. Keller,† Tadamichi Nagashima,‡ and
Dennis P. Curran*,†
Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15260,
Fluorous Technologies Incorporated, UPARC, 970 William Pitt Way,
Pittsburgh, PennsylVania 15238
curran@pitt.edu
Received July 11, 2007
ABSTRACT
A new fluorous carbodiimide is introduced along with a convenient procedure for amide coupling reactions. Reactions of acids and amines
under standard conditions for carbodiimide couplings, followed by simple reverse fluorous solid-phase extraction (FSPE) over standard silica
gel, provide the target amide products in good yields and purities. The use of HFE-7100 as a fluorous solvent is crucial for the success of the
reverse FSPE.
Fluorous solid-phase extraction (FSPE) is an increasingly
popular technique for the separation of reaction mixtures
containing organic components and fluorous components.1,2
In the standard method, a mixture is charged to fluorous silica
gel, and the column is eluted with a fluorophobic solvent in
a first pass to give an organic fraction, followed by a
fluorophilic solvent in a second pass to give a fluorous
fraction. Fluorophobic solvents are polar ones, such as
methanol, acetonitrile, DMF, DMSO, and especially water.
Recognizing that the key to separation in standard fluorous
solid-phase extraction is the combination of a fluorous solid
phase with a polar liquid phase, we reversed the process
(Figure 1); the technique of reverse fluorous solid-phase
extraction combines a polar solid phase with a fluorous liquid
phase.3 The technique is attractive because the polar solid
phase is simply standard silica gel. However, the application
of reverse FSPE is currently limited because perfluorocarbons
such as FC-72 (perfluorohexanes) are used as the key
fluorous liquid component.4
Though not toxic compared to organic solvents, perfluo-
rocarbons are expensive and environmentally persistent.
Their extraordinary nonpolarity also poses problems in the
common occurrence where the fluorous reaction components
† University of Pittsburgh.
‡ Fluorous Technologies Incorporated.
(1) Introduction: Curran, D. P.; Hadida, S.; He, M. J. Org. Chem. 1997,
62, 6714-6715.
(2) Reviews: (a) Zhang, W.; Curran, D. P. Tetrahedron 2006, 62,
11837-11865. (b) Curran, D. P. In The Handbook of Fluorous Chemistry;
Gladysz, J. A., Curran, D. P., Horvath, I. T., Eds.; Wiley-VCH: Weinheim,
2004; pp 101-127. (c) Curran, D. P. Synlett 2001, 1488-1496.
Figure 1. Reverse fluorous solid-phase extraction over standard
silica gel (SiO2).
10.1021/ol701631m CCC: $37.00
© 2007 American Chemical Society
Published on Web 09/21/2007

are somewhat polar. This limitation can be overcome in
principle by blending FC-72 with organic solvents. But FC-
72 is not miscible with most organic solvents, so the range
of solvent blends that can be used for reverse FSPE is limited.
of solubility and gelling. To investigate these problems, we
prepared the four new fluorous carbodiimides shown in
Figure 2 and briefly investigated their properties. Carbodi-
We sought to expand the reverse FSPE technique by
replacing the nonpolar perfluorocarbons with more polar
hydrofluoroethers (HFEs, RfOR).5,6 Solvents such as HFE-
7100 (perfluorobutyl methyl ether, C4F9OCH3)7 are less
expensive than perfluorocarbons and are much less persistent.
Compared to organic solvents, they are still relatively safe.8
Most importantly, from the performance standpoint they are
more polar and are more readily blended with organic
solvents. Accordingly, they have the potential to significantly
increase the range of applications of reverse FSPE.
Figure 2. Fluorous carbodiimides prepared for preliminary solubil-
ity studies.
To probe the usefulness of HFE-7100 in reverse FSPE,
we synthesized new fluorous carbodiimides (FCDIs) and
investigated their use in amide coupling reactions. Among
the many reagents for amide coupling, carbodiimides have
shown enduring popularity because of their broad scope and
ease of use.9 Separation of the derived urea byproducts is
often problematic, and many solutions have been offered.10
Ureas are polar and often streak on silica. Accordingly, the
separation of an inherently polar fluorous urea from an amide
reaction product is a challenging process because the purpose
of the fluorous tag is to move the urea more rapidly off the
column in a reverse FSPE.
Palomo and colleagues described several fluorous carbo-
diimides and introduced a protocol for separation based on
liquid-liquid extraction with FC-72.11 Strong fluorous acids
were used to help draw the otherwise poorly soluble fluorous
ureas into FC-72. In preliminary experiments with several
fluorous carbodiimides and ureas, we encountered problems
imides 1a and 1b were soluble in dichloromethane (a
common solvent for carbodiimide reactions), but their derived
ureas gelled the reaction mixtures. These gelating properties
might be interesting in other settings, but they are not very
attractive for SPE because concentrated reaction mixtures
cannot be pipetted or even poured; they can only be diluted.
The urea derived from 1c did not gel dichloromethane, but
1c itself was unattractive because it was not very soluble in
dichloromethane.
In contrast, carbodiimide 2 was soluble in dichloromethane
and other common solvents, and its derived urea 6 did not
gel (or even precipitate from) dichloromethane. Carbodiimide
2 is made starting from readily available alcohol 3 by the
straightforward sequences of steps shown in Scheme 1 (see
Scheme 1. Synthesis of Urea 6 and Carbodiimide 2
(3) Matsugi, M.; Curran, D. P. Org. Lett. 2004, 6, 2717-2720.
(4) (a) Gladysz, J. A.; Emnet, C. In The Handbook of Fluorous Chemistry;
Gladysz, J. A., Curran, D. P., Horvath, I. T., Eds.; Wiley-VCH: Weinheim,
2004; pp 11-23. (b) Ryu, I.; Matsubara, H.; Emnet, C.; Gladysz, J. A.;
Takeuchi, S.; Nakamura, Y.; Curran, D. P. In Green Reaction Media in
Organic Synthesis; Blackwell: Ames, IO, 2005; pp 59-124.
(5) Examples of applications of HFEs as reaction solvents: (a) Mizuno,
M.; Goto, K.; Miura, T.; Matsuura, T.; Inazu, T. Tetrahedron Lett. 2004,
45, 3425-3428. (b) Fukuyama, T.; Arai, M.; Matsubara, H.; Ryu, I. J. Org.
Chem. 2004, 69, 8105-8107.
(6) Examples of applications of HFEs as liquid-liquid extraction
solvents: Yu, M. S.; Curran, D. P.; Nagashima, T. Org. Lett. 2005, 7, 3677-
3680. (b) Curran, D. P.; Bajpai, R.; Sanger, E. AdV. Synth. Catal. 2006,
348, 1621-1624. (c) Chu, Q.; Yu, M.; Curran, D. P. Tetrahedron 2007, in
press.
(7) (a) HFE solvents are produced by 3M under the tradename Novec:
http://www.3m.com/Product/information/Novec-Engineered-Fluid.html. Most
solvents are mixtures of isomers. For example, HFE-7100 is a mixture of
perfluorobutyl methyl ether and perfluoroisobutyl methyl ether (n- and
isobutyl groups). HFE-7100 is also available from Sigma-Aldrich. (b) DPC
owns an equity interest in Fluorous Technologies, Inc.
(8) (a) Wallington, T. J.; Nielsen, O. J. In Handbook of EnVironmental
Chemistry: Organofluorines; Neilson, A. H., Ed.; Springer-Verlag: Berlin,
2002; Vol. 3, pp 85-102. (b) Most HFEs are not classified as a volatile
organic compound (VOC) and are approved for use under the US EPA
Significant New Alternatives Program (SNAP): http://www.epa.gov/ozone/
snap/regulations.html.
(9) (a) Podlech, J. In Houben-Weyl Methods of Organic Chemistry.
Synthesis of Peptides and Peptidomimetics; Goodman, M., Felix, A.,
Moroder, L., Toniolo, C., Eds.; Thieme-Verlag: Stuttgart, 2001; Vol. E22a,
pp 517-533. (b) Reagents for Glycoside, Nucleotide, and Peptide Synthesis;
Crich, D., Ed.; Wiley: New York, 2005.
Supporting Information for details). O-Allylation provides
4, which is converted to 5 by a sequence of hydroboration
and oxidation to give an alcohol, bromination, and then
(11) (a) Palomo, C.; Aizpurua, J. M.; Loinaz, I.; Fernandez-Berridi, M.
J.; Irusta, L. Org. Lett. 2001, 3, 2361-2364. (b) Aizpurua, J. M.; Palomo,
C.; Loinaz, I. In The Handbook of Fluorous Chemistry; Gladysz, J. A.,
Curran, D. P., Horvath, I. T., Eds.; Wiley-VCH: Weinheim, 2004; pp 459-
461.
(10) For examples, see: (a) Sauer, D. R.; Kalvin, D.; Phelan, K. M.
Org. Lett. 2003, 5, 4721-4724. (b) Lannuzel, M.; Lamothe, M.; Perez, M.
Tetrahedron Lett. 2001, 42, 6703-6705. (c) Jamieson, C.; Congreve, M.
S.; Emiabata-Smith, D.; Ley, S. V. Synlett 2000, 1603-1607.
4168
Org. Lett., Vol. 9, No. 21, 2007

Gabriel synthesis. Reaction of amine 5 with isopropyliso-
cyanate provides urea 6.
Fluorous urea 6 was conveniently transformed to FCDI 2
in a biphasic reaction by taking advantage of the insolubility
of F-26 urea 6 in FC-72. The urea was first dissolved in
CH2Cl2, and FC-72 was added, resulting in two phases.
Triphenylphosphine (3 equiv), bromine (3 equiv), and
triethylamine (5 equiv) were added, and the reaction mixture
was stirred vigorously until it was dark red. FCDI 2 was the
only reaction component soluble in FC-72, so the FC-72 layer
was separated and concentrated. The residue was filtered
through celite to afford the pure carbodiimide 2 as a faint
yellow oil in 93% yield.
Potential solvent systems for the reverse FSPE procedure
were evaluated in a series of TLC experiments on standard
silica TLC plates by using fluorous urea 6 and amide 7 as a
model reaction mixture. These experiments are described in
the thesis of A. Keller,12 and two of the more informative
ones are illustrated in Figure 3.
Figure 4. Acid 8 and amine 9 components in the amide coupling
reactions.
resulting mixture was cooled to 0 °C under argon. After 15
min, the carboxylic acid 8 (0.11 mmol) was added and the
mixture was warmed to rt and, 3 h later, concentrated to
dryness. The crude product was dissolved in a small amount
of THF (0.4 mL) and loaded onto a 2 g SiO2 cartridge. Initial
washing with HFE-7100/EtOAc (3:1) provided a fluorous
fraction,13 and subsequent elution with EtOAc provided a
product fraction, which was concentrated and analyzed.
With a few exceptions, the yields of products 7 were above
70% and the purities were 95% or better. Some of the trace
Figure 3. Standard silica gel TLC experiments with fluorous urea
6 and model amide 7.
Elution of the model mixture with a standard 3:1 blend of
hexane/ethyl acetate showed the urea spot at Rf 0.38 and the
amide at 0.21. Replacing the hexane with HFE-7100
increased the Rf of the urea spot to 0.54 while at the same
time planting the amide spot on the baseline. These contrast-
ing effects can be attributed to the fluorous nature of HFE-
7100, which promotes the mobility of 6 and retards that of
7. The same effect cannot be generated by FC-72 because it
cannot be mixed with ethyl acetate to provide a suitable
solvent blend due to immiscibility.
Table 1. Crude Yields and Purities of Amides from FCDI
Coupling Followed by Reverse FSPEa
entry
acid 8
amine 9
amide 7
yield (%)
purity (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
a
a
a
a
a
b
b
b
b
b
c
c
c
c
c
a
b
c
d
e
a
b
c
d
e
a
b
c
aa
ab
ac
ad
ae
ba
bb
bc
bd
be
ca
cb
cc
cd
ce
87
71
86
58
87
86
73
73
61
70
68 (57)
62 (45)
75
95b
99b
95c
95c
99c
Following some preliminary development experiments,12
a small library of 15 amides was made by coupling acids 8
and amines 9 with FCDI 2 followed by reverse FSPE and
direct analysis of the crude product for fluorous content by
19F or 1H NMR. The acids 8a-e and amines 9a-e used are
shown in Figure 4, and the results of the couplings are
summarized in Table 1.
98b
98b
95c
95c
99c
85b (99)b,d
80b (96)b,d
97c
In the standard procedure, the amine 9 (0.12 mmol) was
added to FCDI 2 (0.14 mmol) in CH2Cl2 (2 mL), and the
d
e
57
74
95c
97c
a
Procedure: FCDI and amine were dissolved in DCM (2 mL) under
(12) Keller, A., Ph.D. Thesis, University of Pittsburgh, 2007. The thesis
is archived and openly available from the University of Pittsburgh Library
Electronic Thesis Deposition (ETD) site at http://etd.library.pitt.edu/ETD/
available/etd-04122007-142347/.
b
argon at 0 °C. Acid was added, and the mixture was stirred at rt for 3 h.
By 19F NMR. By H NMR. After flash chromatography.
c
1
d
Org. Lett., Vol. 9, No. 21, 2007
4169

1
impurities were fluorous as ascertained by H or 19F NMR,
and these typically consist of small amounts of the urea,
although the presence of other products cannot be ruled out
(see below). Even better purities can be obtained by
conducting the separation chromatographically (collecting
and analyzing fractions); however, the purpose of these
experiments was to assess the scope of a fixed SPE procedure
across a range of substrates. For this purpose, the general
procedure worked well.
Scheme 2. Separation in a Difficult Coupling Reaction
Somewhat lower yields were obtained in the reactions
involving tetrahydroisoquinoline 9d as the amine component
(entries 4, 9, and 14), but the purities of the products were
still very good. However, coupling of acid 8c with amines
9a and 9b gave products in only 85% and 80% purity,
respectively. These products are unique by TLC analysis in
that they are considerably less polar than the other amide
products; presumably they are too close to the fluorous region
for adequate SPE separation. The partially pure products were
further purified by standard flash chromatography (with
fraction collection and TLC analysis) in 3:1 HFE-7100/
EtOAc to provide the amide 9a (57% yield, 99% pure) and
9a,b (45% yield, 96% pure).
More difficult amide coupling reactions than those shown
in Table 1 are not expected to proceed in near-quantitative
yields, and accordingly, crude mixtures will contain other
byproducts derived from the carbodiimide besides the urea.
N-Methylindole-3-carboxylic acid 8d is known to be a
recalcitrant participant in carbodiimide-mediated amide
coupling reactions,10a so we investigated the coupling of this
acid with fluorous carbodiimide 2 and diisopropylcarbodi-
imide (DIC) under conditions where high yields of amide
product were not expected. The crude mixtures were purified
by flash chromatography with HFE-7100/EtOAc or hexane/
ethyl acetate. The results of these experiments are sum-
marized in Scheme 2.
The crude mixture from the fluorous reaction showed four
clear spots on TLC. The most polar spot was the target
product 7db, which was isolated in pure form in 17% yield.
The other spots, in reverse order of elution, where identified
as the fluorous urea 6 (21%), the N-acyl urea 10 (15%), and
the carbodiimide dimer 11 (36%). The N-acyl urea resulting
from an O- to N-acyl shift is a well-known dead end product
in carbodiimide couplings, and carbodiimide dimers have
also been observed.14
In contrast, the reaction mixture with standard DIC showed
only two spots on TLC, one of which proved to be the N-acyl
urea 12 (47%), while the other was a mixture of amide 7db
and diisopropyl urea (DIU).
These results show the potential options for manipulation
of the retention times of urea byproducts by using fluorous
solvents. In the case at hand, the standard urea DIU and the
target product coincidentally have about the same Rf, but
the side products from the fluorous reaction are easily
separated from 7db. It is unlikely that amide products could
also coincidentally have about the same Rf as the fluorous
urea or another fluorous byproduct. However, if this occurs,
then separation can probably be established by changing the
fluorous solvent of the elution mixture to an organic one,
thereby significantly changing the relative retention factors
of the fluorous compared to nonfluorous components.
In summary, a new fluorous carbodiimide 2 has been
introduced, and a general procedure for amide coupling
reactions has been deployed. Reactions of acids and amines
under standard conditions for carbodiimide couplings, fol-
lowed by simple reverse fluorous solid-phase extraction over
standard silica gel, provide the target amide products in good
yields and purities. The use of HFE-7100 as a fluorous
solvent is crucial for the success of the FSPE. The results
suggest that HFE-7100 and related solvents have excellent
potential to modulate relative retention factors of organic
and fluorous components in both solid-phase extraction and
chromatographic applications.
Acknowledgment. We thank the National Institute of
General Medical Science of the National Institutes of Health
for funding (P50GM067082 and R44GM067326). C.P.
thanks the Ministerio de Educacio´n y Ciencia of Spain for a
Ramo´n y Cajal contract.
Supporting Information Available: Experimental pro-
cedures for reagent synthesis and amide coupling along with
representative spectra of the products following reverse
FSPE. This material is available free of charge via the
Internet at http://pubs.acs.org.
(13) The fluorous urea can be recovered in about 80% (after purification)
for reuse.
(14) Hartke, K.; Rossbach, F. Angew. Chem., Int. Ed. 1968, 7, 72-74.
OL701631M
4170
Org. Lett., Vol. 9, No. 21, 2007