Radical reactions with alkyl and fluoroalkyl ((fluorous) tin hydride reagents in supercritical CO2

Full text of DOI:10.1021/ja971120i

7406
J. Am. Chem. Soc. 1997, 119, 7406-7407
solution to the second problem is to use “fluorous” (highly
fluorinated) reagents6 because organofluorine compounds are
well-known to be highly soluble in supercritical CO2.2f,h,3,5b We
report herein the combination of these two solutions: typical
radical reactions can be conducted in supercritical CO2 at
moderate pressures with a recently introduced fluorous tin
hydride.6b
Radical Reactions with Alkyl and Fluoroalkyl
(Fluorous) Tin Hydride Reagents in Supercritical
CO2
Sabine Hadida,† Michael S. Super,‡ Eric J. Beckman,*,‡ and
Dennis P. Curran*,†
We investigated in some detail the simple reduction of
bromoadamantane 1 to adamantane 2 (eq 1a). All radical
reactions were conducted by initial pressurization to 850 psig
followed by heating to 90 °C (2500 psig) and pressurization to
4000 psig, and the substrate concentration was 0.05 M. It is
already known that adamantane 2 is poorly soluble under these
conditions.7 In contrast, bromoadamantane 1a is soluble in
supercritical CO2 even at 0.22 M. As expected, tris(perfluo-
rohexylethyl)tin hydride 4 was soluble under the reaction
conditions (at 0.06 M), but the mixture with tributyltin hydride
3 was not homogenous even at 7000 psig. Thus, with this
reagent it is possible that some of the reactions are not occurring
in the CO2 phase. A control experiment under the reaction
conditions also showed that tributyltin hydride 3 itself reacts
with supercritical CO2, albeit inefficiently (eq 1b). Heating of
3 and AIBN in CO2 at 90 °C for 3 h followed by evaporation
of the CO2 provided a mixture of the starting tin hydride 3 and
tributyltin formate 58 in a ratio of 3/1.9 In contrast, the fluorous
tin hydride 4 did not produce a fluorous tin formate but was
recovered unchanged.
Department of Chemistry, UniVersity of Pittsburgh
Pittsburgh, PennsylVania 15260
Department of Chemical Engineering,
UniVersity of Pittsburgh
Pittsburgh, PennsylVania 15261
ReceiVed April 8, 1997
Carbon dioxide in the supercritical state has been touted as
a suitable solvent for organic synthesis for an assortment of
environmental and practical reasons.1,2 Carbon dioxide is not
classified as a VOC (volatile organic chemical) by the EPA,
nor is it regulated in food contact applications by the FDA. In
addition, carbon dioxide is inexpensive and nonflammable and
has a readily accessible supercritical region (Pc ) 1044 psig,
Tc ) 31 °C).
Despite these and other favorable features, the use of
supercritical CO2 as a reaction solvent2 has lagged far behind
its applications in extraction, chromatography, and other separa-
tion processes.1 This is because there appear to be at least two
very broad limitations for many kinds of preparative organic
reactions. First, the use of CO2 as a solvent in polar reactions
will be limited because CO2 is nonpolar and because, even
though it is quite unreactive,2g it may not be suitable for certain
reaction classes that rely on strong nucleophiles or electrophiles.
Second, carbon dioxide is a relatively low dielectric fluid that
is incapable of dissolving large quantities of many organic
compounds at moderate pressures.3 Most organic reactions
require reagents (or catalysts), and problems arise when either
or both of the reactants and the reagents are substantially
insoluble in CO2.
(1a)
(1b)
Solutions of bromoadamantane 1a (0.05 M, 0.85 mmol), AIBN
(10%), and either tributyltin hydride 3 (1 mmol) or fluorous tin
hydride 4 (1 mmol) were heated for 3 h at 90 °C at 4000 psig
of CO2 in a standard reactor. The bromide was consumed in
both reactions. The reactor was cooled, and the CO2 was vented
through ether to collect the entrained products. The crude
product from the tributyltin hydride was purified by the “DBU
workup”10 to provide adamantane 2 in 88% yield, while the
reaction with the fluorous tin hydride 4 was purified by
partitioning between FC-72 (perfluorohexane) and benzene6b to
provide adamantane 2 in 90% yield.
Similar reduction of a primary iodide and the steroidal iodide,
bromide, and phenyl selenide 6a-c provided the corresponding
reduced product 7 in high yields, as indicated in eq 2. Yields
were comparable with tributyltin hydride and the fluorous tin
hydride, but the reactions with the fluorous tin hydride were
easier to purify, and the corresponding fluorous tin byproduct,
(C6F13CH2CH2)3SnX (8a-c, X ) I, Br, or PhSe), was isolated
in >90% yield by evaporation of the FC-72 phase. No attempt
A prospective solution to the first problem is to use radical
reactions, which require neither nucleophiles nor electrophiles
and proceed well in nonpolar solvents.4 Tanko and Blackert5a
have reported benzylic brominations in CO2, and DeSimone5b
has used CO2 as a medium for radical polymerizations. A
† Department of Chemistry.
‡ Department of Chemical Engineering.
(1) (a) Eckert, C. A.; Knutson, B. L.; Debenedetti, P. G. Nature 1996,
383, 313. (b) Brennecke, J. F. Chem. Ind.-London 1996, 831. (c) Phelps,
C. L.; Smart, N. G.; Wai, C. M. J. Chem. Educ. 1996, 73, 1163.
(2) Leading references: (a) Savage, P. E.; Gopalan, S.; Mizan, T. I.;
Martino, C. J.; Brock, E. E. AIChE J. 1995, 41, 1723. (b) Kaupp, G. Angew.
Chem., Int. Ed. Engl. 1994, 33, 1452. (c) Weinstein, R. D.; Renslo, A. R.;
Danheiser, R. L.; Harris, J. G.; Tester, J. W. J. Phys. Chem. 1996, 100,
12337. (d) Jessop, P. G.; Ikariya, T.; Noyori, R. Science 1995, 269, 1065.
(e) Jessop, P. G.; Hsiao, Y.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc.
1996, 118, 344. (f) Burk, M.; Feng, S.; Gross, M. F.; Tumas, W. J. Am.
Chem. Soc. 1995, 117, 8277. (g) For example, even Lewis acid reactions
can be conducted in CO2. Pernecker, T.; Kennedy, J. P. Polymer Bull. 1997,
33, 19. (h) Leitner and co-workers are concurrently reporting that perfluo-
ralkyl-substituted phosphanes are soluble ligands for the catalysis in CO2.
Baumann, W.; Kainz, J.; Koch, D.; Leitner, W. Angew. Chem., Int. Ed.
Engl. In press. We thank Professor A. Pfalz for informing us of this work
and Prof. W. Leitner for a preprint.
(6) (a) Horvath, I. T.; Rabai, J. Science 1994, 266, 72. (b) Curran, D. P.;
Hadida, S. J. Am. Chem. Soc. 1996, 118, 2531. (c) Dimagno, S. G.; Dussault,
P. H.; Schultz, J. A. J. Am. Chem. Soc. 1996, 118, 5312. (d) Curran, D. P.;
Hoshino, M. J. Org. Chem. 1996, 61, 6480. (e) Studer, A.; Hadida, S.;
Ferrito, R.; Kim, S.-Y.; Ja¨ger, P.; Wipf, P.; Curran, D. P. Science 1997,
275, 823.
(3) (a) Luque de Castro, M. D.; Valcarcel, M.; Tena M. T. Analytical
Supercritical Fluid Extraction; Springer-Verlag: Berlin, 1994; p 41. (b)
McHugh, M. A.; Krukonis V. J. Supercritical Fluid Extraction: Principle
and Practice; Butterworths: Stoneham, MA, 1994.
(4) Giese, B. Radicals in Organic Synthesis: Formation of Carbon-
Carbon Bonds; Pergamon Press: New York, 1986. (b) Curran, D. P. In
ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 4, pp 715 and 779. (c) Jasperse, C. P.;
Curran, D. P.; Fevig, T. L. Chem. ReV. 1991, 91, 1237. (d) Motherwell,
W.; Crich, D. Free Radical Chain Reactions in Organic Synthesis; Academic
Press: London, 1992.
(7) McFann, G. J.; Johnston, K. P.; Hurter, P. N.; Hatton, T. A. IEC
Res. 1993, 32, 2336.
(8) The formate was identified by comparison with authentic material
prepared by the direct reaction of bis(tributyltin)oxide with formic acid.
Ohara, M.; Okawara, R. J. Organomet. Chem. 1965, 3, 484.
(9) No tin formate was formed in an experiment conducted without
AIBN.
(5) (a) Tanko, J. M.; Blackert, J. F. Science 1994, 263, 203. (b)
DeSimone, J. M.; Guan, Z.; Elsbernd, C. S. Science 1992, 257, 945.
(10) Curran, D. P.; Chang, C.-T. J. Org. Chem. 1989, 54, 3140.
S0002-7863(97)01120-7 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7407
was made to recover the tributyltin products after DBU workup.
No evidence for carboxylated products was obtained in these
experiments.
The standard Giese reaction shown in eq 5 was also
conducted. Addition of iodoadamantane 1b to acrylonitrile (5
equiv) provided the radical adduct 15 in 81% yield after flash
chromotography. Also formed in this experiment was a acetone-
soluble material that we assume is polyacrylonitrile.5b When
1.5 equiv of acrylonitrile was used, compound 15 was isolated
in 70% yield, and the formation of the acetone-soluble material
was not observed.
(2)
(5)
In contrast, reduction of 9-iodoanthracene with the fluorous
tin hydride 4 provided anthracene in 71% yield along with 10%
yield of 9-anthracenecarboxylic acid (eq 3). It thus appears that
there is some potential to carboxylate reactive radicals in
supercritical CO2, and carboxylation could presumably be
increased by raising the CO2 density. This is important because
carboxylation is rare in radical chemistry (while decarboxylation
is common), and because the use of supercritical CO2 as a
reagent has heretofore been difficult.2e,f,11
The simplicity of these experiments belies their significance.
Although the favorable features of using supercritical CO2 as a
reaction medium for organic synthesis have been widely cited
for over a decade, it is still not very clear what kinds of reactions
to run in supercritical CO2 and how to run them. Tin hydrides
and related reagents are now commonly used in synthesis to
make C-H, C-C, and (to a lesser extent) C-N and C-O
bonds. Although further studies are surely needed, wholesale
porting of this reaction class to supercritical CO2 now seems
like a viable possibility. Furthermore, the use of highly CO2-
soluble fluorous reagents and catalysts2b should prove to be a
valuable strategy to transport other reaction classes to CO2. In
addition, there are now fluorous protecting groups available,6e
and these could be used to confer additional solubility on the
reaction substrates themselves. The solubility problem can also
be viewed from the other side: fluorous chemistry is providing
useful new options in both traditional and parallel synthesis,6e,14
but one of the main problems in this field is choosing a reaction
solvent. Supercritical CO2 provides a valuable addition to the
existing complement of solvents available to dissolve both
fluorous and organic compounds.
(3)
Reduction of 1,1-diphenyl-6-bromo-1-hexene 10 under the
standard conditions with the fluorous tin hydride 4 provided
the 5-exo cyclized product 11 in 87% isolated yield along with
7% of the reduced product 12 (eq 4). Interestingly, reduction
of 10 under the standard conditions in liquid benzotrifluoride
(1 atm) provided only the cyclized product 11, which was
isolated in 75% yield.12 The absence of the reduced product is
expected since all the relevant rate constants are known,13 and
hydrogen transfer cannot compete with cyclization under these
conditions. Although effects of phase separation cannot be
firmly ruled out, we speculate that some reduced product 12
forms in supercritical CO2 because diffusion in supercritical
solvents is faster than in liquids.1,2 This would allow bimo-
lecular reduction to compete more efficiently with cyclization.
Significantly, the reduction of 10 with tributyltin hydride 3
produced neither 11 nor 12 but returned 10 and 3 along with
some tin formate 5. Although additional solubility studies are
needed, this failure may be due to the insolubility of 10 and 3.
Reduction of the aryl iodide 13 with 4 provided 14 in 99% yield
along with 99% of the tin iodide 8a.
Ultimately, the biggest advantage of coupling fluorous
reagents with synthetic chemistry may be at the separation stage.
At the end of many reactions, the problem of separating products
from spent and unspent reagents remains. The use of environ-
mentally friendly reaction solvents like supercritical CO2 makes
little sense if the reactions are followed by standard extractions
or chromatographies with traditional organic solventssextractions
and chromatographies invariably require more solvent volumes
than the reactions that precede them. However, it seems
probable that the large differences in solubility in supercritical
CO2 between fluorous and organic compounds can be translated
into practical separation procedures. In the long run, the CO2
should serve as both the reaction and the separation solvent.
Acknowledgment. We thank the National Institutes of Health for
funding this work.
Supporting Information Available: A representative experimental
procedure for the reductions, a summary of the solubility experiments,
and representative spectra of crude products (7 pages). See any current
masthead page for ordering and Internet access instructions.
JA971120I
(4)
(12) Heating of 10 and 4 for 12 h at 90 °C with AIBN returns mostly
recovered starting materials, presumably because 10 and 4 are not miscible.
This suggests that reactions with 4 are occurring in the CO2 phase. Heating
of 10 and 4 with a minimal amount of benzotrifluoride (estimated
concentration 0.6 M) provided the following yields: 11, 62%; 12, 6%; 10,
21%.
(11) Anionic carboxylations under electrochemical conditions: Sullen-
berger, E. F.; Dressman, S. F.; Michael, A. C. J. Phys. Chem. 1994, 98,
5347.
(13) Curran, D. P.; Hadida, S.; Horner, J. H.; Martinez, F. N.; Newcomb,
M. Tetrahedron Lett. 1997, 38, 2783.
(14) Curran, D. P. ChemtractssOrg. Chem. 1996, 9, 75.