Fluorous Mitsunobu reagents and reactions

Article abstract of DOI:10.1016/S0040-4020(02)00205-3

A fully fluorous Mitsunobu reaction procedure is introduced. This employs both existing [(C₆F₁₃CH₂CH₂C₆H ₄)₂PPh] and new [C₈F₁₇CH₂CH₂C₆H ₄PPh₂] fluorous phosphines and a new fluorous azodicarboxylate (C₆F₁₃CH₂CH₂OC(O)N=NCOOCH ₂CH₂C₆F₁₃). A procedure involving parallel reactions with representative nucleophiles and alcohols under typical Mitsunobu conditions in THF followed by rapid solid phase extraction (spe) over fluorous silica provides clean products in excellent yields. The fluorous fraction containing the oxidized phosphine oxide and the reduced hydrazide can be readily separated and the starting reagents can be regenerated by appropriate redox reactions in high yield for reuse.

Full text of DOI:10.1016/S0040-4020(02)00205-3

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
TETRAHEDRON
Pergamon
Tetrahedron 58 )2002) 3855±3864
Fluorous Mitsunobu reagents and reactions
Sivaraman Dandapani and Dennis P. Curranp
Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260, USA
Received 26 December 2001; accepted 15 February 2002
AbstractÐA fully ¯uorous Mitsunobu reaction procedure is introduced. This employs both existing [)C6F13CH2CH2C6H4)2PPh] and new
[C8F17CH2CH2C6H4PPh2] ¯uorous phosphines and a new ¯uorous azodicarboxylate )C6F13CH2CH2OC)O)NvNCOOCH2CH2C6F13). A
procedure involving parallel reactions with representative nucleophiles and alcohols under typical Mitsunobu conditions in THF followed
by rapid solid phase extraction )spe) over ¯uorous silica provides clean products in excellent yields. The ¯uorous fraction containing the
oxidized phosphine oxide and the reduced hydrazide can be readily separated and the starting reagents can be regenerated by appropriate
redox reactions in high yield for reuse. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
¯uorous phosphines to perform a ¯uorous Mitsunobu reac-
tion. After a standard solution phase Mitsunobu coupling,
all the reagent-derived products can be separated from the
desired organic products by rapid ®ltration through ¯uorous
silica gel. The ¯uorous byproducts can be separated from
each other and reconverted to the starting reagents for reuse.
We have recently learned that Dobbs and McGregor-
Johnson have also made the same ¯uorous azodicarboxylate
and used it in some Mitsunobu reactions. They used this
reagent with triphenylphosphine and separated the ¯uorous
hydrazine by the complementary method of liquid±liquid
extraction.9 This leaves the substitution product contami-
nated with triphenylphospine oxide; however, highly ¯uori-
nated phosphines that could complement a liquid±liquid
separation are already known.2,6
Fluorous methods for synthesis and separation are poised to
move from a relatively small, specialized community into
the mainstream of organic synthesis.1 Early ¯uorous tech-
niques were based on liquid±liquid separations, and these
techniques are increasing in importance in catalysis and
other areas.2 Typically, a large number of ¯uorines )39
and up) is used to ensure high solubility of the ¯uorous
component in a suitable highly ¯uorinated solvent. In
1997, we reported the ®rst solid±liquid separations based
on ¯uorous silica gel )silica gel with a ¯uorocarbon bonded
phase).3 Solid±liquid separations are operationally con-
venient, and relatively few ¯uorines are needed to effect a
simple binary separation of a mixture into `¯uorous' and
`non-¯uorous' )that is, organic) fractions.4 The ef®ciency
of the separations over ¯uorous reverse phase silica gel
enables the use of ¯uorous reagents and catalysts with
lower molecular weight and improved solubility in organic
solvents.5 A recent `New Tools in Synthesis' article high-
lights ¯uorous silica gel and its uses in ¯uorous synthesis
techniques.6a
Eq. )1) shows a generic Mitsunobu reaction of an acidic
pronucleophile 1 and an alcohol 2 promoted by diethylazo-
dicarboxylate 3 )DEAD) and triphenylphosphine 4 )TPP).10
This reaction produces the coupled product 5 along with
dicarboethoxyhydrazine 6 )DCEH) and triphenylphosphine
oxide 7 )TPPO). The lack of atom economy detracts
somewhat from large scale applications, but the Mitsunobu
reaction has been enduringly popular for small scale and
The usefulness of ¯uorous methods in mainstream synthesis
depends on the availability of a diverse assortment of
proven reagents, catalysts, tags, protecting groups and
scavengers. A partialilst of useful¯uorous compounds
includes a wide assortment of ¯uorous phosphines2 and tin
reagents7 as well as an increasing assortment of oxygen8 and
nitrogen4,6b protecting groups.
Herein we describe new ¯uorous dialkylazodicarboxyates
and we use these in combination with known and new
Keywords: ¯uorous; Mitsunobu; solid phase extraction; triphenylphos-
phine; diethylazodicarboxylate.
p
Corresponding author. Tel.: 11-412-624-8240; fax: 11-412-624-9861;
e-mail: curran@pitt.edu
ꢀ1†
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020)02)00205-3

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3856
S. Dandapani, D. P. Curran / Tetrahedron 58 %2002) 3855±3864
phosphine and a ¯uorous azodicarboxylate. A number of
¯uorous analogs of triphenylphosphine )abbreviated as
FTPP) were already available in our group,5 and these
provided a ready departure point for Mitsunobu chemistry.
Based on retention times on a ¯uorous hplc column, we
initially selected phosphine 8 PhP)C6H4CH2CH2C6F13)2
)Fig. 1) with one unsubstituted phenylring and two substi-
tuted rings bearing per¯uorohexylgroups insualted by
ethylene spacers. Fluorous phosphines with longer chains
or with three ¯uorous phenyl rings are also available, but
hplc studies5 suggested that the longer retentions and added
molecular weight of these analogs was not needed for solid
phase extraction )spe) separations. The molecular weight of
8 is 954.
The ¯uorous phosphine 8 was paired with the classical
DEAD reagent 3 and this combination was then tested for
its applicability in Mitsunobu reactions. Esteri®cation of
3,5-dinitrobenzoic acid and N-alkylation of phthalimide
with methanol, ethanol and isopropanol were conducted,
and the results of this series of experiments are shown in
Fig. 1.
Reactions with 3,5-dinitrobenzoic acid 9 were conducted by
adding a solution of FTPP 8 )1 equiv.) and the alcohol
)2 equiv.) in ether to a solution of DEAD 3 )1 equiv.) and
3,5-dinitrobenzoic acid )1 equiv.) in ether.16 This order of
addition is hereafter called Procedure A. Reactions with
phthalimide 10 were conducted by adding a solution of
DEAD 3 )1 equiv.) in THF to a solution of phthalimide
F
)1 equiv.), TPP 8 )1 equiv.), and the alcohol )2 equiv.).
Figure 1. Mitsunobu reactions with ¯uorous phosphine 8 and DEAD.
This order of addition is called Procedure B. After
12±16 h, the solvent was evaporated and the crude reaction
mixture ),200 mg) was taken up in methanoland charged
onto 2 g of homemade17 ¯uorous silica gel. Elution
with 10 mL of 80% MeOH/water gave a mixture of the
Mitsunobu adducts 11 or 12 along with dicarboethoxy-
hydrazine )DCEH) 6. While the ¯uorous solid phase
extraction completely removed the ¯uorous phosphine
oxide 13 PhP)O))C6H4CH2CH2C6F13)2 from 11 or 12,
chromatography was needed to remove the hydrazine 6.
The isolated yields of 11 or 12 after normalsilca gel
chromatography are shown in the bottom of Fig. 1 and
range from 75 to 91%. These results clearly show that
¯uorous phosphine 8 is effective in typicalMitsunobu
reactions.
discovery chemistry because of its scope, stereoselectivity
and mild reaction conditions. The separation of the reagent-
derived byproducts from the desired coupled products is
almost invariably the most time- and resource-consuming
part of the Mitsunobu reaction.
To avoid chromatography, an assortment of separation-
friendly Mitsunobu reagents have been introduced. Both
polymer-bound phosphines11 and azodicarboxylates12 have
been used, though large excesses are needed for convenient
rates and conversions. And the two polymeric reagents
presumably cannot be used together since their interaction
would be prevented by phase isolation. Suitably substituted
reagents can be removed by acid±base extractions )either
directly13 or after chemicalactivation 14) or by decompo-
sition to volatile products. Residual reagents and byproducts
can also be polymerized for removal by ®ltration in a tech-
nique called `impurity annihilation'.15 These methods allow
solution phase reactions, but they require extra reaction
chemistry to effect the separation. This potentially adds
time, limits scope, and renders dif®cult or impossible the
recycling of the reagents. The ¯uorous Mitsunobu reactions
and reagents described herein offer an appealing solution to
the nagging separation problems of the Mitsunobu reaction
without compromising the reaction itself.
The new ¯uorous DEAD reagent 17 )hereafter called
`FDEAD', pronounced `eff-dead') required to complement
the ¯uorous phosphine was synthesized as shown in Scheme
1. 2-Per¯uorohexylethanol 14 was treated with 1,10-carbo-
nyldiimidazole )1.2 equiv.) at room temperature in THF.
Quenching with water and extraction with ether provided
a crude product )presumably imidazolide 15), which was
directly reacted with hydrazine hydrochloride and triethyl-
amine. The ¯uorous hydrazine 16 was isolated as a white
powder in 85% yield after standard chromatography. Oxi-
dation of 16 with iodosobenzene diacetate18 was not clean,
but 17 was produced smoothly on exposure of 16 to bromine
and pyridine in dichloromethane.19 Standard workup and
concentration of the dichloromethane gave the FDEAD
reagent 17 as a yellow solid )molecular weight 810) in
quantitative yield. The purity of 17 was excellent as judged
2. Results and discussion
A fully ¯uorous Mitsunobu reaction requires both a ¯uorous

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
S. Dandapani, D. P. Curran / Tetrahedron 58 %2002) 3855±3864
3857
To test the reactivity of the FDEAD reagent 17 in Mitsunobu
reactions and to investigate the possibility of separating
the ¯uorous phosphine oxide and hydrazine byproducts,
we reacted it with 3,5-dinitrobenzoic acid 9, ethanoland
FTPP 8 )Fig. 2) by Procedure A. After about 16 h, the
solvent was evaporated, the crude reaction mixture
),200 mg) was taken up in MeOH and loaded on to 2 g
of ¯uorous silica. Elution with 80% MeOH gave the
Mitsunobu adduct 11a in 92% yield free of any impurities
1
as judged by GC-MS analysis and H NMR spectroscopy.
Further elution with ether gave a mixture of the ¯uorous
phosphine oxide 13 and the ¯uorous hydrazine 16
)FDCEH). This ¯uorous byproduct mixture was readily
separated on normal silica gel by eluting with 3:2 hexane/
EtOAc. The less polar hydrazine 16 emerged ®rst and was
isolated in 80% yield; the more polar phosphine oxide 13
followed in 86% yield.
Scheme 1.
The ease of separation of the ¯uorous reagent-based
byproducts 13 and 16 is handy because the reagents can
then be recycled. Both the reduction of the ¯uorous phos-
phine oxide 13 )with alane20) to return 8 and the oxidation of
the ¯uorous hydrazine 16 to regenerate 17 are clean, high
yielding reactions.
1
by H, 13C and 19F NMR spectroscopy, and hence it was
used without further puri®cation in the following Mitsunobu
F
reactions. A solid sample of DEAD 17 has been stored at
ambient temperature for severalmonths without apparent
decomposition, and a solution of 17 in CDCl3 )0.02 M)
1
did not deteriorate over 2 weeks as assayed by H NMR
spectroscopy.
We next conducted four experiments to test the compati-
bility of the ¯uorous Mitsunobu reagents 8 and 17 with each
other and with the regular organic reagents triphenyl-
phosphine and DEAD. 3,5-Dinitrobenzoic acid was coupled
with MeOH by reacting all four possible combinations of
¯uorous )FTTP 8 and FDEAD 17) and organic reagents )TPP
4, and DEAD 3) by Procedure A. After 12 h, the solvent was
evaporated and each reaction mixture was individually
subjected to an identical¯uorous soild phase extraction.
The organic fraction from each experiment was concen-
1
trated and analyzed by H NMR spectroscopy. These four
NMR spectra are reproduced in Fig. 3.
In the reaction with both standard reagents )TPP and
DEAD), the spe has no effect and the product 11a is isolated
mixed with TPPO 7 and DCEH 6. With ¯uorous azo-
dicarboxylate 17 and TPP 4, the ¯uorous hydrazine 16
)FDCEH) is retained, but the TPPO 7 passes through with
the product 11a. In the reverse combination )FTPP 8 and
DEAD 3), the ¯uorous phosphine oxide )FTPPO) 13 is
retained but the DCEH 6 passes through. However, in the
experiment where both ¯uorous reagents )FTPP 8 and
FDEAD 17) were used, both derived ¯uorous products
F
)FTPPO 13 and DCEH 16) were retained and the pure
product 11a eluted. These results provide compelling
evidence of the simplicity and effectiveness of the separa-
tion of ¯uorous compounds from organic ones on ¯uorous
silica.
Several parallel experiments were conducted to explore the
scope of these ¯uorous reagents with a representative selec-
tion of simple alcohols and nucleophiles. These parallel
experiments are enabled by the ¯uorous spe since all the
reaction mixtures can be puri®ed in just a few minutes. Nine
experiments combining three nucleophiles )3,5-dinitro-
benzoic acid, phthalimide and N-)t-butoxycarbonyl)±p-
toluene sulfonamide21) with three different alcohols
)MeOH, allyl alcohol and p-¯uorobenzylalcohol) were
Figure 2. Fluorous Mitsunobu reaction with separation of ¯uorous
byproducts and recycling.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
3858
S. Dandapani, D. P. Curran / Tetrahedron 58 %2002) 3855±3864
Figure 3. Mitsunobu reactions of 3,5,-dinitrobenzoic acid and methanol with all possible combinations of TPP, DEAD, FTPP and FDEAD. )a) Using TPP and
DEAD, )b) Using TPP and FDEAD, )c) Using FTPP and DEAD, )d) Using FTPP and FDEAD.
conducted as usual)Procedure A was used for the acid and
B was used for the imide and sulfonamide). After ¯uorous
spe, the solvent was evaporated and the organic product was
cases the GC chromatogram exhibited only a single peak.
In the sixth case )entry 3), the product was 91% pure. The
two products from p-¯uorobenzylaclohol)entries 3, 6)
exhibited single resonances in their 19F NMR spectra; no
CF3 or CF2 resonances from the ¯uorous products were
evident. However, the results with N-)t-butoxycarbonyl)-
p-toluenesulfonamide )entries 7±9) were not fully
satisfactory. The reaction with MeOH )entry 7)
occurred in 100% conversion, but the reaction with allyl
alcohol was only about 41% complete )entry 8). No con-
version at all was observed with p-¯uorobenzylaclohol
)entry 9).
1
weighed and analyzed by GC-MS and H NMR spectro-
scopy. The products from p-¯uorobenzylalcohol reactions
were also analyzed by 19F NMR spectroscopy with the ¯uor-
ine atom serving as an internalstandard. Resutsl are
summarized in Table 1, entries 1±9.
The results of the six reactions with 3,5-dinitrobenzoic acid
and phthalimide )entries 1±6) were highly satisfactory;
yields of products ranged from 75±93% and in ®ve of six