An expedient and stereoselective synthesis of alkenyl nonaflates from silyl enol ethers: Optimization, scope and limitations

Article abstract of DOI:10.1002/1099-0690(200203)2002:6<1015::AID-EJOC1015>3.0.CO;2-K

The fluoride-catalysed reaction between silyl enol ethers 1 and nonafluorobutanesulfonyl fluoride (NfF) has been optimized, resulting in an expedient synthesis of the corresponding alkenyl nonaflates 3. Tetra-n-butylammonium fluoride, dried either with molecular sieves or with potassium fluoride, and potassium fluoride in the presence of dibenzo-18-crown-6 were the best and most practical catalysts for this process. The reaction allows the synthesis of a wide variety of cyclic and acyclic alkenyl nonaflates 3 in good to excellent yields. For E/Z isomeric alkenes the configuration of the double bond is essentially retained. Remarkably, enolates derived from methyl ketones also provide C-sulfonylation products 4 as a side product; the desired alkenyl nonaflates 31 and 3m could, however, be prepared in good yields by further optimization. Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.

Full text of DOI:10.1002/1099-0690(200203)2002:6<1015::AID-EJOC1015>3.0.CO;2-K

FULL PAPER
An Expedient and Stereoselective Synthesis of Alkenyl Nonaflates from
Silyl Enol Ethers: Optimization, Scope and Limitations
Ilya M. Lyapkalo,[a] Matthias Webel,[b][‡] and Hans-Ulrich Reißig*[a]
Keywords: Alkenyl nonaflates / Carbanions / Counterion / Fluorides / Silyl enol ethers
The fluoride-catalysed reaction between silyl enol ethers 1
and nonafluorobutanesulfonyl fluoride (NfF) has been optim-
ized, resulting in an expedient synthesis of the corresponding
For E/Z isomeric alkenes the configuration of the double
bond is essentially retained. Remarkably, enolates derived
from methyl ketones also provide C-sulfonylation products 4
alkenyl nonaflates 3. Tetra-n-butylammonium fluoride, dried as a side product; the desired alkenyl nonaflates 3l and 3m
either with molecular sieves or with potassium fluoride, and
potassium fluoride in the presence of dibenzo-18-crown-6
were the best and most practical catalysts for this process.
The reaction allows the synthesis of a wide variety of cyclic
and acyclic alkenyl nonaflates 3 in good to excellent yields.
could, however, be prepared in good yields by further optim-
ization.
(
Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
2002)
Introduction
TfF is a toxic gas that requires special technique and pre-
cautions[4b] and therefore has to be transformed into more
convenient but rather expensive triflating reagents.[2] NfF,
on the other hand, is a stable and easily handled liquid[7]
that is of considerable interest as a mild and efficient fluor-
inating[8] or nonaflating[8,9] reagent. However, no systematic
studies on the preparation of alkenyl nonaflates with the
aid of NfF has been reported to date.[8,10] In the context of
our interest in the application of alkenyl nonaflates in Pd-
catalysed cross-coupling reactions[5b,5d,5e] we present here
the results of our studies on their synthesis with silyl enol
ethers as starting material. For comparison, their prepara-
tion from the corresponding carbonyl compounds has been
investigated in few representative cases.
Since the discovery of their versatility in transition metal-
catalysed or -mediated cross-coupling reactions, alkenyl tri-
flates have attracted great attention. This in turn has had a
powerful impact on the development of new methods for
their synthesis. As a result, a number of new efficient triflat-
ing reagents have been introduced[1] and become commer-
cially available,[2] and have been employed in highly chemo-,
regio- and diastereoselective procedures.[3] The much
broader use of alkenyl and aryl triflates (as compared with
other perfluoroalkanesulfonates) in cross-coupling reac-
tions is reflected in the titles of relevant reviews.[4] However,
the synthetic potential of other alkenyl perfluoroalkanesul-
fonates seems to have been underestimated and still remains
to be developed. This mainly concerns 1,1,2,2,3,3,4,4,4-
nonafluorobutanesulfonic acid enol esters (hereafter abbre-
viated to alkenyl nonaflates), the successful application of
which in Heck coupling reactions,[5] including a one-flask
procedure starting from trimethylsilyl enol ethers,[5b] has al-
ready been reported. An advantage of the nonaflates is
evident from comparison of the properties of trifluorome-
thanesulfonyl fluoride (TfF) and nonafluorobutanesulfonyl
fluoride (NfF), which are industrial products[6] and pre-
cursors of all the other triflate and nonaflate derivatives.
Results and Discussion
It is noteworthy that NfF is a moderately electrophilic
reagent and cannot react directly with ketones or aldehydes
(unless they contain CϪH-activating electron-withdrawing
substituents in the α-position) in the presence of conven-
tional metal-free nitrogen bases[11] (R3N, pyridine, 2,6-di-
tert-butyl-4-methylpyridine), unlike the respective anhy-
dride.[12] Therefore, activation of the enolizable carbonyl
components by conversion into the appropriate enolates
A[13] (commonly with iPr2NLi[8] or (Me3Si)2NMet[10d]) is
required, followed by addition of NfF (Scheme 1, top) or
through the intermediacy of silyl enol ethers. Advantages
of the latter approach are:
[a]
Freie Universität Berlin, Institut für Chemie Ϫ Organische
Chemie,
Takustraße 3, 14195 Berlin, Germany
Fax: (internat.) ϩ49Ϫ(0)30/8385Ϫ5367
E-mail: hans.reissig@chemie.fu-berlin.de
[b]
Ϫ Silyl enol ethers[14] are easily available chemo-, regio-
and stereoselectively by various methods (including 1,4-ad-
ditions and DielsϪAlder and Hetero-DielsϪAlder reac-
Institut für Organische Chemie der Technischen Universität
Dresden,
01062 Dresden, Germany
See ref.[5d]
[‡]
Eur. J. Org. Chem. 2002, 1015Ϫ1025
WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 1434Ϫ193X/02/0306Ϫ1015 $ 17.50ϩ.50/0
1015

I. M. Lyapkalo, M. Webel, H.-U. Reißig
FULL PAPER
were employed, with variations in the solvent, the amounts
and concentrations of the reacting components, the reac-
tion temperature and the reaction time. Entries 1Ϫ4 and
6Ϫ8 represent optimal conditions for the corresponding
fluoride source. As the nonaflation process involves the
moisture-sensitive enolate A as a reactive intermediate,[17]
careful protection from atmospheric moisture and dryness
of the fluoride source have to be ensured. Use of the fluoro
ate complexes (entries 1Ϫ4) is beneficial, as these can be
obtained in anhydrous form. Tetra-n-butylammonium tri-
phenyldifluorosilicate, a nonhygroscopic crystalline solid,
provided the best result (entry 1), albeit giving Ph3SiF as a
side product. The analogous tin complex resulted in poorer
yields of the nonaflate 3a even when used in considerably
higher molar amounts (entry 2), and also produced toxic,
tin-containing side product(s). Tris(dimethylamino)sulfon-
Scheme 1
tions). Highly enantioselective methods for desymmetriz-
ation of cyclic ketones by use of homochiral lithium amide
bases with successive[15] or internal[16] quenching with tri-
methylsilyl chloride have been developed.
Ϫ Once generated from a silyl enol ether, enolate A inter-
acts with NfF in such a way that a fluoride anion is regener-
ated in each elementary step (Scheme 1, bottom), thus in
principle permitting the use of substoichiometric or cata-
lytic quantities of the initial fluoride source.
Ϫ The reaction routinely proceeds at high concentrations
of the reactants (1Ϫ6 mmol per mL of solvent), which al-
lows a subsequent reaction, such as a Pd-catalysed cross-
coupling, to be carried out by simple dilution with the de-
sired solvent, followed by addition of the required compon-
ents.[5b]
Optimization of the Reaction Conditions
For optimization of the reaction conditions, cyclic silyl
enol ethers 1a, 1b and 2c (Scheme 2) were used in the pres-
ence of different fluoride sources (Table 1). Thus, fluoro ate
complexes (entries 1Ϫ4) and fluoride salts (entries 5Ϫ8) Scheme 2
Table 1. Comparison of different nucleophilic activators in reactions of model trialkylsilyl enol ethers 1a, 1b and 2c with NfF to afford
the nonaflates 3a-c according to Scheme 2
Entry Starting
material Solvent
Optimized reaction conditions
NfF (equiv.) Activator Amount (equiv.) T ( °C)
Product
Time (h) (yield, %)
1
2
3
4
5
6
7
8
1a
1a
1a
1a
1a
1a
1a
1b
1a
1a
2c
THF
1.2
2.0
2.0
1.2
1.2
2.0
1.2
1.2
1.2
1.2
1.2
[nBu4N][SiPh3F2]
[nBu4N][SnPh3F2]
[(Me2N)3S][SiMe3F2]
[nBu4N][HF2][b]
CsF
0.10
0.26
0.13
0.07
0.20
0 Ǟ room temp.
0 Ǟ room temp.
0 Ǟ room temp.
0 Ǟ room temp.
25
0 Ǟ room temp.
0 Ǟ room temp.
Ϫ78 Ǟ room temp. 17
0 Ǟ room temp. 16
Ϫ78 Ǟ room temp. 16
Ϫ78 Ǟ room temp. 16
4
3a (85)
3a (55)
3a (75)
3a (77)
3a (10)
3a (60)
3a (80)
3b (85)
3a (Ϫ[e])
3a (55)
3c (32)
CH2Cl2
THF/CH2Cl2
CH2Cl2
Ϫ
PhH
THF
THF
THF
THF
THF
40
16
16
48
16
16
[a]
{[(Me2N)3PϭN]4P}ϩ FϪ 0.12
[nBu4N]F, method A[c]
[nBu4N]F, method B[d]
[nBu4N]CN
0.18
0.05
0.20
1.1
9
10
11
MeLi, 1 h[f]
MeLi, 18 h[f]
1.1
[a]
[b]
[c]
˚
1:1 ratio.
In the presence of 0.18 equiv. Me3SiSiMe3. 1 molar [nBu4N]F solution in THF dried with 4-A molecular sieves (see
ref.[29] for details). Crystalline [nBu4N]F·3H2O pre-dried in vacuum, dissolved in THF and treated with dried KF (0.7Ϫ0.9 equiv.; see
[d]
ref.[29] for details). Traces of the product detected. The silyl enol ether solution was stirred with MeLi at Ϫ5 to 0 °C for the time
[e]
[f]
stated before the addition of NfF.
1016
Eur. J. Org. Chem. 2002, 1015Ϫ1025

Synthesis of Alkenyl Nonaflates from Silyl Enol Ethers
FULL PAPER
ium difluorotrimethylsilicate (entry 3) and tetra-n-butylam- 9Ϫ11). Highly nucleophilic tetra-n-butylammonium cyan-
monium hydrogendifluoride[18] (entry 4) showed compar- ide was expected to initiate the process generating the enol-
able efficiencies. However, the former reagent is extremely ate. This would release fluoride upon interaction with NfF
moisture-sensitive and cannot be stored as a stock solution, and induce a sequence according to Scheme 1, bottom.
whilst the latter contains relatively acidic hydrogens and However, only traces of the nonaflation product were ob-
gives good results only after in situ treatment with hexa- served (entry 9), together with recovered starting material
methyldisilane[19] by the procedure originally introduced by 1a and cyclopentanone resulting from hydrolysis. Methylli-
Vorbrüggen et al. for drying of [nBu4N]F solutions.[20]
thium is known to generate lithium enolates from silyl enol
The results with fluoride salts are shown in entries 5Ϫ8 ethers[28] in an amine-free procedure that gives unreactive
(Table 1). Cesium fluoride provided a disappointingly low tetraalkylsilane as the only side product. This procedure im-
yield of the nonaflate 3a when used without solvent (entry plies at least one equivalent of MeLi at 0 °C prior to the
5), and gave only traces of the product if the reaction was addition of the electrophile. The resulting lithium enolate
conducted in THF, which may be due to the poor solubility of cyclopentanone was treated with NfF to give nonaflate
of CsF in the reaction media.[21] Tetrakis[tris(dimethyl- 3a in moderate yield (entry 10). Analogous in situ forma-
amino)phosphoranylideneamino]phosphonium fluoride,[22] tion of the lithium enolate of cyclohexanone from tert-bu-
a highly soluble anhydrous fluoride source containing a tyldimethylsilyl enol ether 2c took considerably longer to be
dendritically branched, thermally very stable lipophilic effected (entry 11); subsequent treatment with NfF resulted
counterion, gave satisfactory yields of the product only in a low yield of nonaflate 3c.
when applied in more than 10 mol % (entry 6). Unlike this
Certain conclusions can be drawn from the data shown
phosphonium fluoride, tetra-n-butylammonium fluoride, a in Table 1 and from the following experiments. Suitably
widely used fluoride salt highly soluble in organic solvents, dried tetra-n-butylammonium fluoride (methods A or B, see
is known to be unstable in completely anhydrous form.[23] above) appears to provide a desirable compromise con-
However, both
a commercially available solution of cerning reactivity, stability in solutions, and price.
[nBu4N]F in THF containing Յ 5% water and one prepared Tetrakis[tris(dimethylamino)phosphoranylideneamino]-
from crystalline [nBu4N]F·3H2O gave low yields (15Ϫ21%) phosphonium fluoride and the studied tetra-n-butylammo-
of the nonaflate 3c in the model reaction 1c Ǟ 3c (see nium fluorosilicates can be used as purchased without any
Scheme 2), together with considerable amounts of cyclo- additional treatment, but give no appreciable benefit in re-
hexanone, indicating hydrolysis.[24] A suitable drying was action efficiency and are rather expensive.
achieved by repeated treatment of the commercially avail-
able [nBu4N]F solution in THF with powdered molecular
Synthesis of Different Nonaflates
sieves 4 A[25] (method A). The activation time and the num-
˚
ber of consecutive dryings with fresh portions of the mo-
We synthesized nonaflates 3aϪn, varying the type, num-
lecular sieves were both essential to attain satisfactory re- ber and position of substituents at the C,C-double bond in
sults (see Table 1, entry 7 and footnote). The molar quantity the starting trimethylsilyl enol ethers 1aϪn under the
of the required fluoride source (ca. 15Ϫ20 mol %) was still optimized conditions. The results are summarized in
relatively high and the drying procedure resulted in consid- Schemes 3Ϫ5. The method works well for the synthesis of
erable loss (up to 50%) of the starting [nBu4N]F solution. cyclic nonaflates formally derived from cyclic ketones
We later succeeded in evading these drawbacks by use of a (Scheme 3), and for nonaflates with a terminal ONf group,
stock solution prepared from crystalline [nBu4N]F·3H2O by arising from parent aldehyde precursors (Scheme 4). These
drying at 45Ϫ48 °C (0.01 mbar) for 15 hours followed by nonaflates are reasonably stable, colourless or yellowish hy-
dissolving in THF. The resulting solution was used in the drophobic liquids, which can easily be purified either by
presence of an excess of finely ground and thoroughly dried distillation in vacuum or by column chromatography. We
potassium fluoride (method B). According to the literature, attribute the slightly lower yields of the simplest represent-
drying of [nBu4N]F·3H2O at 40Ϫ45 °C (Ͻ 0.13 mbar) for atives Ϫ ethenyl nonaflate[29] and 1-propenyl nonaflate (3h)
ca. 48 hours affords ‘‘anhydrous’’ [nBu4N]F containing Ϫ to their higher volatility, whereas that of buta-1,3-dien-
0.1Ϫ0.3 equiv. of water and ca. 10% of [nBu4N][HF2].[26] 2-yl nonaflate 3n (Scheme 5, Table 2, entry 9) may be due
Whatever the exact constitution of the resulting [nBu4N]F, to lower stability.[30] With the exception of cyclododecanone
we believe that a fine suspension of anhydrous KF in the derivatives (1g Ǟ 3g), the reactions were clean and exclus-
solution of the predried [nBu4N]F serves as a heterogeneous ively provided the nonaflates as isolated products. Nona-
inorganic scavenger of the residual water and/or hydrogen flate 3g was not separated from the side products, which
fluoride, acting as an inexpensive and efficient source of were cyclododecyne, cyclododeca-1,2-diene and cyclodode-
‘‘anhydrous’’ fluoride in situ. The fluoride catalyst prepared canone (36%, Ͻ 1% and 20% yields, respectively according
by this method gave the best result with respect both to the to GC-MS data; see Scheme 3, bottom). Although base-in-
molar quantity used (5 mol %) and to the yield of nonaflate duced elimination of perfluoroalkanesulfonic acids from al-
3b from 4-tert-butyl(trimethylsiloxy)cyclohexene (1b) kenyl perfluoroalkanesulfonates is well documented in the
(entry 8).[27]
literature,[12a] the reaction behaviour of 1g represents the
A few experiments in the nonaflation process were per- only case in which we actually observed alkyne or allene
formed with other nucleophilic activators (Table 1, entries formation under our nonaflation conditions.
Eur. J. Org. Chem. 2002, 1015Ϫ1025
1017

I. M. Lyapkalo, M. Webel, H.-U. Reißig
FULL PAPER
Scheme 3. Method A: [nBu4N]F (0.15Ϫ0.29 equiv.), THF, NfF (1.1Ϫ2.0 equiv.), 0 °C Ǟ room temp., 16Ϫ18 h; Method B: [n-Bu4N]F
(0.04Ϫ0.05 equiv.), KF, THF, NfF (1.2 Ϫ 1.3 equiv.), Ϫ78 °C Ǟ room temp., 16Ϫ18 h
For aldehyde-derived nonaflates (Scheme 4), high degrees other enolates discussed above. As in our previous
of retention of the initial configuration of C,C-double study,[10a] we again demonstrated the cation effect on C-
bonds were observed. The E/Z ratios in products 3h, 3j and versus O-regioselectivity of enolate sulfonylation, this time
3k, as well as in buta-1,3-dien-1-yl nonaflate,[29] were close in more detail using NfF (Scheme 5). The lithium enolates
to those in the corresponding starting silyl enol ethers. were generated from the respective ketones, whereas for
Hence the nonaflation procedure had occurred without E/ other enolates, trimethylsilyl enol ethers 1lϪ1n were used as
Z isomerization of the intermediate enolate and may be precursors (Table 2). The reactions showed trends similar to
termed a stereospecific process.
those described earlier for the sulfonylation of 3,3-dimethyl-
The enolates derived from methyl ketones require separ- butan-2-one enolates with benzenesulfonyl fluoride:[10a] an
ate discussion, because a considerable amount of C-sulfon- increase in the ionic character of the enolate resulted in
ylation was observed. This formation of ketosulfones might localization of the negative charge mainly at the enolate
be due to the considerably smaller steric hindrance at the oxygen, which raised the relative proportion of alkenyl nona-
terminal enolate carbon in comparison with that in the flate as a result of charge-controlled O-nonaflation. Li-
1018
Eur. J. Org. Chem. 2002, 1015Ϫ1025

Synthesis of Alkenyl Nonaflates from Silyl Enol Ethers
FULL PAPER
thium enolates of acetone and 3,3-dimethylbutan-2-one
Use of the potassium enolate Ϫ generated by treatment
gave considerable amounts of the nonafluorobutanesulfonyl of 1m with an equimolar quantity of potassium ethoxide[31]
ketones 4 (entries 1, 4). In the reaction of the lithium enol- Ϫ distinctly shifted the ratio in favour of the alkenyl nonafl-
ate of pinacolone we found ketosulfone 6 (2%) as a by- ate 3m (cf. entries 5 and 4). The less coordinating tetra-n-
product, arising from perfluorinated cyclic sulfone 5. This butylammonium cation further enhanced the O-nonaflation
compound was present as an impurity in the NfF to an rate, rendering alkenyl nonaflates 3l and 3m either highly
extent of 5%. In other nonaflation experiments we did not predominant (entry 6) or virtually exclusive (entry 2). How-
isolate the corresponding C-sulfonylation products. Al- ever, the preparative yields were not satisfactory (entries 2,
though we cannot rigorously rule out their formation in 6, 7). 1H NMR monitoring of the reaction mixtures showed
these reactions, it seems that only the sterically less hindered considerable amounts of the parent ketones as by-products.
enolates derived from methyl ketones are able to react as C- A combination of thoroughly dried potassium fluoride (1.0
nucleophiles with 5.
equivalent) and dibenzo-18-crown-6[32] delivered a fluoride
source excluding a priori any adventitious acidic hydrogens
during the reaction, but still favouring charge control in the
nonaflation step. This provided exclusive O-nonaflation and
good yields of nonaflates 3l (entry 3) and 3m (entry 8). For
the preparation of 3m, the use of a THF/DMF mixture was
Scheme 4. Method A: [nBu4N]F (0.15Ϫ0.29 equiv.), THF, NfF
(1.1Ϫ2.0 equiv.), 0 °C Ǟ room temp., 16Ϫ18 h unless otherwise
noted
Scheme 5
Table 2. Influence of the cation on C/O-regioselectivity of enolate nonaflation with NfF according to Scheme 5 (reactions were conducted
in THF unless otherwise noted).
R
Activator for 1
(equiv.)
Cation
3:4
Yield
Entry
ratio[a]
3 (%)
4 (%)
1
2
3
4
5
6
7
8
9
Me
Me
Me
tBu
tBu
tBu
tBu
tBu
Ϫ
Liϩ
64:36
Ϫ
4l (25)
[nBu4N][HF2] (0.2) ϩ KF (2.1)[b]
dibenzo-18-crown-6 (0.2) ϩ KF (1.0)
Ϫ
nBu4Nϩ
Kϩ [d]
Liϩ
Ͼ 95: Ͻ 5[c]
Ͼ 95:Ͻ 5[c]
31:69
3l (32)
3l (68)
Ϫ
3m (47)
3m (40)
3m (39)
3m (64)
3n (61)
Ϫ
Ϫ
4m (43)
4m (17)
Ϫ
Ϫ
Ϫ
Ϫ
EtOK (1.0)
Kϩ
75:25
[nBu4N][HF2] (0.2)[e] ϩ KF (2.1)[b]
[nBu4N]F (0.2)[e] Ϫ method B
dibenzo-18-crown-6 (0.2) ϩ KF (1.0)[g]
[nBu4N]F (0.18) Ϫ method A
nBu4Nϩ
nBu4Nϩ
Kϩ[d]
93:7
Ϫ[f]
Ͼ 95:Ͻ 5[c]
Ϫ[f]
CHϭCH2
nBu4Nϩ
[a]
Determined for the crude mixture by 1H NMR after acidic workup (aqueous H3PO4, pH ϭ 2); only one of these components was
usually isolated due to different workup procedures (see Exp. Sect.). [b] KF is expected to serve here as a heterogeneous inorganic scavenger
[c]
of residual HF in the fluoride activator, analogously to Method B. No clear signals of the nonafluorobutanesulfonyl ketones 4 were
[d]
[e]
observed. As a complex with dibenzo-18-crown-6. 0.1 equiv. of the fluoride salt gave only 50Ϫ60% conversion of 1m regardless of
[f]
[g]
the reaction time. No aqueous workup was performed. In DMF/THF, 1:1 mixture.
Eur. J. Org. Chem. 2002, 1015Ϫ1025
1019

I. M. Lyapkalo, M. Webel, H.-U. Reißig
FULL PAPER
advantageous, since the reaction in neat THF was rather lylation step, tetra-n-butylammonium fluoride was found to
sluggish. Considering the somewhat lower stability of buta- be a good fluoride source, provided that a suitable drying
1,3-dienyl nonaflate 3n,[30] we refrained from studying the procedure was performed. Potassium fluoride/dibenzo-18-
counterion effect on buta-1,3-dien-2-olate nonaflation. The crown-6 gave better results when applied to silyl enol ethers
product 3n (entry 9) was prepared and isolated in moderate derived from methyl ketones. The reaction is applicable to
yield by the general procedure without aqueous workup.[33] a wide range of silyl enol ethers derived from ketones or
The nonaflation reaction seemed to be sensitive to the aldehydes and enables alkenyl nonaflates to be synthesised
electronic effects of substituents at the C,C-double bond with a high degree of stereoretention if E/Z isomeric al-
and to steric effects of silyl substituents. Thus, attempts to kenes are involved. With the exception of compound 3g (see
prepare bisnonaflates from cis-1,2-bis(trimethylsiloxy)- Scheme 3), alkenyl nonaflates 3 were isolated as the exclus-
ethene
or
1,2-bis(trimethylsiloxy)cyclohexene
with ive products. The explicit advantage of our method is the
[nBu4N]F as a promoter have so far failed. Furthermore, use of trimethylsilyl enol ethers as starting materials. Their
none of the fluoride sources tested gave appreciable yields chemo-, regio- and stereoselective syntheses, as well as their
of nonaflates when silyl enol ethers bearing silyl groups isolation and purification, are comprehensively documented
bulkier than trimethylsilyl were employed; 1-(tert-butyldi- in the literature. The nonaflation method is suitable for the
methylsiloxy)cyclohexene (2c), 2-(tert-butyldimethylsiloxy)- synthesis of pure alkenyl nonaflates 3, but also for one-pot
cyclohexa-1,3-diene[34] and 2-(n-butyldimethylsiloxy)-3,3- coupling procedures (such as Heck or Suzuki reactions) in
dimethylbutene[5d] could not be converted into the corres- which the generated nonaflates are directly used without
ponding nonaflates. This behaviour indicated that the purification. This will be reported in a subsequent ac-
mechanism as depicted in Scheme 1 may be oversimpl- count.[5e]
ified.[35] As expected, methyllithium can also cleave the tert-
butyldimethylsilyl group upon prolonged treatment of the
Experimental Section
enol ether 2c at 0 °C (Table 1, entry 11), to generate the
respective lithium enolate, but this procedure can hardly be
applied to polyfunctional substrates as very few functional
groups are able to tolerate methyllithium under these harsh
reaction conditions.
General Methods: NMR spectra were recorded on Bruker AC 500,
WH 270 or AC 250 instruments, in CDCl3 as solvent unless stated
1
otherwise. H and 13C chemical shifts are expressed as ppm down-
field from tetramethylsilane (δ ϭ 0) used as an internal standard.
19F chemical shifts are given in ppm upfield from CFCl3 (δ ϭ 0)
used as an internal standard. 13C signals of the CF3(CF2)3 groups
are not given, as no unambiguous assignments were possible, due
to strong splitting by coupling with the 19F nuclei. Mass spectra
were registered with a Varian MAT 711 spectrometer. GS/MS ana-
lysis was performed with a HewlettϪPackard (HP) 5890 Series II
and a HP 5972 MS-Selective Detector. IR spectra were measured
with Beckman IR Acculab 4, Beckman IR 5A, PerkinϪElmer IR
1420 or PerkinϪElmer FT-IR spectrometers Nicolet 5 SXC. TLC
analysis was performed with Merck 60 F254 silica gel plates. Col-
umn chromatography (hereafter abbreviated as CC) was conducted
on 60 silica gel (40Ϫ63 µm, Fluka).
As mentioned above, generation of metal enolates is re-
quired to effect direct conversion of ketones into enol nona-
flates with NfF (see Scheme 1, top). We therefore applied
iPr2NLi followed by addition of NfF (Scheme 6). This ap-
proach was approximately as efficient as that of the ‘‘de-
tour’’ via trimethylsilyl enol ethers when applied to the syn-
thesis of the cyclic nonaflates 3c and 3f (i in Scheme 6; cf.
Scheme 3 1c Ǟ 3c or 1f Ǟ 3f). Use of DBU as base re-
quired considerably more drastic conditions[11] to effect
nonaflation of ketones with NfF and resulted in lower
yields of products. Synthesis of the model nonaflate 3c was
achieved in 48% yield by heating the reaction components
in toluene.
The lithiation, silylation and nonaflation reactions were carried out
under atmosphere of argon in heat-gun-dried reaction flasks, add-
ing the components by syringe. Solvents for reactions were dried
by standard procedures. Nonafluorobutanesulfonyl fluoride was
obtained from Bayer AG; it can also be purchased from Aldrich.
The preparation and drying of [nBu4N]F solution in THF (Aldrich
or Fluka, contains 3% water), nBu4NF·3H2O (MerckϪSchuchardt)
and KF used in Methods A and B are described by us in our pre-
ceding publication.[29] A 50% solution of [nBu4N][HF2] in CH2Cl2
(1.93 mmol·mLϪ1) was purchased from Fluka. Potassium ethoxide
was purchased from Aldrich. Dibenzo-18-crown-6 (MerckϪ
Schuchardt) was stored over CaH2 in a tightly closed vial.
Scheme 6. i. 1) iPr2NLi in THF, Ϫ78 °C; 2) NfF, Ϫ78 °C Ǟ room
temp., overnight; ii. DBU in toluene at 85 °C, NfF, 85 °C overnight
Methyl vinyl ketone (Fluka, assay ca. 95%, contains ca. 0.5% hy-
droquinone, ca. 0.5% AcOH and ca. 5% water) was treated with
P2O5 at Ϫ78 °C and recondensed under high vacuum into a cold
(Ϫ78 °C) trap containing a few crystals of hydroquinone and kept
at Ϫ18 °C. N-Ethoxycarbonyl-4-tropinone was purchased from
Acros Organics. The starting silyl enol ethers were obtained by the
standard procedures described in the literature. 1-Trimethylsiloxy-
Conclusion
We have developed a general and expedient method for
synthesis of alkenyl nonaflates 3 from trimethylsilyl enol
ethers 1 and the industrial product nonafluorobutanesul-
fonyl fluoride (NfF). For the crucial fluoride-induced desi- non-1-ene (1k, E/Z ϭ 93:7),[36] and 1-trimethylsiloxyoct-1-ene (1j,
1020
Eur. J. Org. Chem. 2002, 1015Ϫ1025

Synthesis of Alkenyl Nonaflates from Silyl Enol Ethers
FULL PAPER
E/Z ϭ 25:75)[37] were synthesized stereoselectively as described in quid (130 °C/4 mbar). The physical and spectroscopic data of 3d
the literature. The latter compound was further enriched up to 4:96
E/Z ratio by a kinetic resolution exploiting the considerably higher
reactivity of the E isomer in a hetero-DielsϪAlder reaction.[38] 8-
Methyl-3-trimethylsiloxy-8-azabicyclo[3.2.1]oct-2-ene (rac-1f) was
prepared from tropinone (Acros Organics) and isolated without
aqueous workup as described in ref.[24] Commercially available 2.5
molar nBuLi (Aldrich) was titrated by the [1,10]phenanthroline
method.[39]
agreed with those described in the literature.[10c]
3,4-Dihydronaphthalen-1-yl Nonaflate (3e): According to Method
A, silyl enol ether 1e (0.270 g, 1.24 mmol), [nBu4N]F solution
(0.36 mmol) and NfF (0.725 g, 2.40 mmol), after purification by
CC (hexane/EtOAc, 10:1), provided 0.531 g (Ͼ 99%) of 3e as a col-
ourless, viscous oil. 1H NMR (300 MHz): δ ϭ 2.51 (td, J ϭ 8.0,
5.7 Hz, 2 H, 3-H), 2.87 (t, J ϭ 8.0 Hz, 2 H, 4-H), 6.02 (t, J ϭ
5.7 Hz, 1 H, 2-H), 7.18Ϫ7.15, 7.28Ϫ7.23, 7.37Ϫ7.34 (3 m, 4 H,
Ar). 13C NMR (75.5 MHz): δ ϭ 22.3, 26.8 (2 t, C-3, C-4), 117.7,
121.3, 126.9, 127.8, 129.3 (5 d, C-5 to C-8, C-2), 128.7 (s, C-1),
136.2 (s, C-4a), 146.6 (s, C-8a). IR (film): ν˜ ϭ 3150 cmϪ1 (ϭCϪH),
2950, 2840 (CH2), 1655, 1485 (CϭC), 1420, 1145 (SO2), 1235, 1200
(CϪF). GC/MS-data: retention time: 10.5 min (81 °C). MS (EI,
70 eV): m/z (%) ϭ 428 (33) [Mϩ], 219 (24), [C4F9], 145 (54) [Mϩ Ϫ
C4F9SO2], 131 (16), 115 (100) [C9H7], 100 (12), 91 (33) [C7H7], 69
(32) [CF3], 48 (3) [SO]. C14H9F9O3S (428.3): calcd. C 39.26, H 2.12,
S 7.49; found C 39.54, H 2.08, S 7.18.
Experimental details on optimization of the nonaflation procedure
with different nucleophilic activators (Table 1) are given in ref.[5d]
Alkenyl Nonaflates 3. General Procedures. Method A: A 1  solu-
tion of ‘‘dry’’ [nBu4N]F in THF[29] was added at 0 °C, with vigor-
ous stirring, to the neat silyl enol ether 1. No dilution of the reac-
tion mixture with extra solvent was made prior to the addition of
the [nBu4N]F unless otherwise noted. Once the fluoride source had
been added, the neat NfF was injected dropwise by syringe to give
a two-phase, yellowish-to-brown reaction mixture, which soon af-
terwards became homogeneous. Stirring (at ambient temperature
for 16Ϫ18 h unless otherwise noted), removal of the volatiles in
vacuum and subsequent kugelrohr distillation of the residue af-
forded the pure products.
rac-8-Methyl-8-azabicyclo[3.2.1]oct-2-en-3-yl Nonaflate (3f): Ac-
cording to Method B, rac-1f (0.850 g, 4.00 mmol), [nBu4N]F solu-
tion (0.15 mmol) and NfF (1.57 g, 5.20 mmol) in the presence of
KF (0.200 g, 3.40 mmol) in THF (0.7 mL), after purification by
kugelrohr distillation (105Ϫ108 °C/0.3 mbar), provided 1.42 g
(84%) of 3f as an orange oil. 1H NMR (270 MHz): δ ϭ 1.57Ϫ1.68,
1.88Ϫ2.27, 2.82 (2 m, d*, J ϭ 17.5 Hz, 1 H, 4 H, 1 H, 4-H, 6-H,
7-H), 2.41 (s, 3 H, MeN), 3.39Ϫ3.50 (m, 2 H, 1-H, 5-H), 5.86
(d*, J ϭ 5.4 Hz, 1 H, 2-H); * further splitting by multiple couplings.
13C NMR (67.5 MHz): δ ϭ 29.9, 33.0, 34.3 (3 t, C-4, C-6, C-7),
Method B: A 1.05 molar [nBu4N]F solution in THF[29] was added
to a suspension of the thoroughly dried KF in THF, and the re-
sulting mixture was stirred at 0Ϫ5 °C for 15Ϫ20 min before being
cooled to Ϫ78 °C. The silyl enol ether and NfF were subsequently
added at Ϫ78 °C, and the resulting mixture was allowed to warm
to ambient temperature over 2Ϫ3 h and stirred overnight. Aqueous
workup, extraction with hexane, drying of the combined organic
phase (MgSO4), and subsequent column chromatography or kugel-
rohr distillation of the residue furnished the pure nonaflates 3.
35.0 (q, MeN), 57.4, 58.2 (2 d, C-1, C-5), 120.8 (d, C-2), 146.0 (s,
Ϫ1
˜
C-3). IR (film): ν ϭ 3500Ϫ3300 cm (ϭCϪH), 2950, 2875, 2805
(CϪH), 1420, 1145 (SO2), 1240, 1200 (CϪF). MS (EI, 80 eV): m/z
(%)
ϭ
421 (4) [Mϩ], 392 (1) [Mϩ Ϫ CH3N], 138 (12)
Cyclopent-1-enyl Nonaflate (3a): According to Method A, 1a
(0.390 g, 2.50 mmol), [nBu4N]F solution (0.45 mmol) and NfF
(0.906 g, 3.00 mmol) provided 0.725 g (80%) of 3a as a colourless
liquid (b.p. 80 °C/24 mbar). The physical and spectroscopic data of
3a agreed with those described in the literature.[10a]
[Mϩ Ϫ CF3(CF2)3SO2], 131 (1) [CF2ϭCFCF2]ϩ, 122 (5)
[Mϩ Ϫ CF3(CF2)3SO3], 110 (23) [Mϩ Ϫ CF3(CF2)3SO2 Ϫ C2H4],
96 (15), 69 (6) [CF3ϩ], 18 (100) [H2Oϩ or NH4ϩ]. C12H12F9NO3S
(421.3): calcd. C 34.21, H 2.87, N 3.32; found C 33.93, H 2.85,
N 3.50.
rac-4-tert-Butylcyclohex-1-enyl Nonaflate (3b): According to
Method B, rac-1b (0.180 g, 0.80 mmol), [nBu4N]F solution
(0.04 mmol) and NfF (0.910 g, 3.01 mmol) in THF (0.2 mL) in
the presence of KF (0.040 g, 0.70 mmol), after purification by CC
(hexane), provided 0.295 g (85%) of 3b as a colourless oil (b.p. 100
°C/2 mbar). 1H NMR (270 MHz): δ ϭ 0.89 (s, 9 H, tBu),
1.26Ϫ1.45, 1.88Ϫ2.02, 2.15Ϫ2.45 (2 m, 2 H, 2 H, 3 H, 4-H, 5-H,
6-H), 5.76 (dt, J ϭ 5.8, 2.2 Hz, 1 H, 2-H). 13C NMR (75.5 MHz):
δ ϭ 27.2 (q, CMe3), 24.1, 25.4, 28.6 (3 t, C-3, C-5, C-6), 32.1 (s,
CMe3), 42.9 (d, C-4), 118.5 (d, C-2), 149.3 (s, C-1). IR (film): ν˜ ϭ
3100Ϫ3000 cmϪ1 (ϭCϪH), 2965, 2930, 2875, 2850 (CϪH), 1420,
1145 (SO2), 1240, 1200 (CϪF). GC/MS-data: retention time:
9.02 min (79 °C). MS (EI, 70 eV): m/z (%) ϭ 435 (9) [Mϩ Ϫ 1], 379
(19), 218 (10), 131 (10), 83 (16), 69 (35, CF3ϩ), 57 (100) [C4H9ϩ],
55 (21), 41 (19). C14H17F9O3S (436.3): calcd. C 38.54, H 3.93, S
7.34; found C 38.53, H 3.94, S 7.19.
Nonaflation of 1-Trimethylsiloxycyclododec-1-ene (1g): The reaction
was performed according to Method A, by treatment of 1g (1.27 g,
5.00 mmol; E/Z ϭ 82:18) with [(Me2N)3S][SiMe3F2] (0.65 mmol,
0.25 molar in THF/CH2Cl2) and NfF (3.02 g, 10.0 mmol), followed
by kugelrohr distillation at 120 °C/0.001 mbar to afford a mixture
of products (1.34 g). Repeated fractional distillation at 80 °C/
2 mbar furnished a fraction (0.890 g) enriched in nonaflate 3g but
complete separation was not achieved, and the content of the re-
sulting mixture of products, established by GC-MS, was: (E)-3g
(29%), (Z)-3g (14%), cyclododecyne[40] (36%), cyclododeca-1,2-di-
ene[40] (Ͻ 1%) and cyclododecanone (20%).
(E)-Cyclododec-1-enyl Nonaflate (3g): 1H NMR (300 MHz): δ ϭ
1.20Ϫ1.74, 2.10Ϫ2.47 (2 m, 20 H, 3-H Ϫ 12-H), 5.49 (t, J ϭ 8.3 Hz,
1 H, 2-H). 13C NMR (75.5 MHz): δ ϭ 22.0, 22.9, 24.0, 24.2, 24.3,
24.7, 25.3, 25.4, 26.0, 26.4 (10 t, C-3 Ϫ C-12), 124.0 (d, C-2), 148.6
(s, C-1).
(Z)-Cyclododec-1-enyl Nonaflate (3g): 1H NMR (300 MHz): δ ϭ
1.20Ϫ1.74, 2.10Ϫ2.47, (2 m, 20 H, 3-H Ϫ 12-H), 5.37 (t, J ϭ
7.7 Hz, 1 H, 2-H). 13C NMR (75.5 MHz): δ ϭ 22.1, 22.5, 23.9,
24.0, 24.2, 24.3, 24.4, 24.9, 25.4, 26.1 (10 t, C-3 Ϫ C-12), 123.0 (d,
C-2), 150.5 (s, C-1).
Cyclohex-1-enyl Nonaflate (3c): According to Method A, 1c (1.70 g,
10.0 mmol), [nBu4N]F solution (1.50 mmol) and NfF (6.04 g,
20.0 mmol) provided 1.54 g (71%) of 3c as a colourless liquid (120
°C/24 mbar). The physical and spectroscopic data of 3c agreed with
those described in the literature.[10a]
Cyclohept-1-enyl Nonaflate (3d): According to Method A, 1d
Prop-1-enyl Nonaflate (3h): According to Method A, 1h (0.355 g,
(0.920 g, 5.00 mmol), [nBu4N]F solution (0.90 mmol) and NfF 2.72 mmol; E/Z ϭ 88:12), [nBu4N]F solution (0.45 mmol) and NfF
(3.02 g, 10.0 mmol) provided 1.40 g (71%) of 3d as a colourless li-
Eur. J. Org. Chem. 2002, 1015Ϫ1025
(0.906 g, 3.00 mmol), after purification by kugelrohr distillation at
1021

I. M. Lyapkalo, M. Webel, H.-U. Reißig
FULL PAPER
70 °C/20 mbar, provided 0.582 g (63%) of 3h as a colourless oil (E/
9), 22.5, 26.5, 28.7, 28.8, 28.9, 31.7 (6 t, C-3 Ϫ C-8), 122.5 (d, C-
1
Z ϭ 77:23). (E)-3h: H NMR (300 MHz): δ ϭ 1.73 (dd, J ϭ 7.0, 2), 136.2 (d, C-1). The following signals can be assigned to the
1
1.8 Hz, 3 H, Me), 5.29 (qd, J ϭ 7.0, 5.7 Hz, 1 H, 2-H), 6.58 (dq*,
nonaflate (Z)-3k: H NMR (300 MHz): δ ϭ 0.92 (t*, J ϭ 6.0 Hz,
J ϭ 5.7, 1.8 Hz, 1 H, 1-H); * the signal partially overlaps with 3 H, 9-H), 5.21 (td, J ϭ 7.7, 5.5 Hz, 1 H, 2-H), 6.54 (mc*, 1 H, 1-
signals of the Z isomer. 13C NMR (75.5 MHz): δ ϭ 9.6 (q, C-3),
H); * the signal partially overlaps with the signals of the E isomer.
115.1 (d, C-2), 136.5 (d, C-1). (Z)-3h: 1H NMR (300 MHz): δ ϭ IR (film): ν ϭ 3150 cmϪ1 (ϭCϪH), 2960, 2930, 2860 (CϪH), 1665
1.69 (dd, J ϭ 7.2, 1.8 Hz, 3 H, Me), 5.79 (dq, J ϭ 7.2, 4.6 Hz, 1 (CϭC), 1430, 1145 (SO2), 1240, 1205 (CϪF). GC/MS-data: (E)-3k,
H, 2-H), 6.52 (mc*, 1 H, 1-H); * the signal partially overlaps with retention time: 6.99 min (140 °C). MS (EI, 70 eV): m/z (%) ϭ 424
˜
the signals of the E isomer. 13C NMR (75.5 MHz): δ ϭ 11.7 (q, C-
(Ͻ 1) [Mϩ], 219 (8) [C4F9ϩ], 131 (21), 95 (23), 82 (38), 69 (100)
3), 117.5 (d, C-2), 136.7 (d, C-1). IR (film): ν˜ ϭ 3125 cmϪ1 [CF3ϩ], 55 (63) [C4H9ϩ], 43 (67) [C3H7ϩ], 29 (37) [C2H5ϩ]. (Z)-3k,
(ϭCϪH), 2935 (CϪH), 1675 (CϭC), 1390, 1145 (SO2), 1240, 1200 retention time: 6.49 min (135 °C). MS (EI, 70 eV): m/z (%) ϭ 424
(CϪF). GC/MS-data: (E)-3h, retention time: 1.71 min (72 °C). MS (38) [Mϩ], 219 (10) [C4F9ϩ], 131 (23), 95 (23), 82 (38), 69 (100)
(EI, 70 eV): m/z (%) ϭ 340 (15) [Mϩ], 276 (8) [Mϩ Ϫ SO2], 219 (8)
[CF3ϩ], 55 (48) [C4H9ϩ], 43 (80) [C3H7ϩ], 29 (38) [C2H5ϩ].
C13H17F9O3S (424.3): calcd. C 36.80, H 4.04, S 7.56; found C
[C4F9ϩ], 131 (33), 119 (12) [C2F5ϩ], 100 (23) [C2F4ϩ], 69 (100)
[CF3ϩ], 57 (6), 41 (88) [C3H5ϩ], 29 (75). Ϫ (Z)-3h, retention time: 37.07, H 4.07, S 7.21.
1.84 min (72 °C). MS (EI, 70 eV): m/z (%) ϭ 340 (38) [Mϩ], 276
3,3-Dimethylbuten-2-yl Nonaflate (3m): According to Method B,
(15) [Mϩ Ϫ SO2], 219 (12) [C4F9ϩ], 131 (38), 119 (13) [C2F5ϩ], 100
(25) [C2F4ϩ], 69 (100) [CF3ϩ], 57 (6), 41 (92) [C3H5ϩ], 29 (77).
C7H5F9O3S (340.2): calcd. C 24.72, H 1.48, S 9.43; found C 25.04,
H 1.50, S 9.43.
3m was prepared from 1m (0.170 g, 1.00 mmol), [nBu4N]F solution
(0.20 mmol) and NfF (0.400 g, 1.30 mmol) in the presence of KF
(0.120 g, 2.10 mmol) in THF (0.3 mL). No aqueous workup was
performed. Instead, the reaction mixture was diluted with pentane,
filtered through celite and concentrated in vacuum (up to
2-Methylprop-1-enyl Nonaflate (3i): According to Method A, 1i
(0.600 g, 4.17 mmol), diluted with THF (4 mL), [nBu4N][SiPh3F2] 100 mbar), and the residue was subjected to CC (pentane) to give
solution (0.50 mmol; 0.25 molar solution in THF) and NfF (1.81 g, the nonaflate 3m (0.150 g, 39%) as a colourless liquid. The physical
6.00 mmol), after 21 h at room temp. and purification by kugelrohr and spectroscopic data of 3m agreed with those described in the lit-
distillation at 80 °C/10 mbar, provided 1.34 g (90%) of 3i as a col-
ourless oil. 1H NMR (300 MHz): δ ϭ 1.69 (d, J ϭ 1.3 Hz, 3 H, E-
Me), 1.75 (d, J ϭ 1.1 Hz, 3 H, Z-Me), 6.42 (mc, 1 H, 1-H). 13C
NMR (75.5 MHz): δ ϭ 15.6, 19.0 (2 q, 2 Me), 125.9 (s, C-2), 130.8
(d, C-1). The physical and spectroscopic data of 3i were consistent
with those described in the literature.[10a]
erature.[10a]
Buta-1,3-dien-2-yl Nonaflate (3n): According to Method A, 1n
(0.355 g, 2.50 mmol), [nBu4N]F solution (0.45 mmol) and NfF
(0.906 g, 3.00 mmol), after purification by kugelrohr distillation at
25 °C/20 mbar, provided 0.538 g (61%) of 3n as a colourless, volatile
1
liquid. H NMR (270 MHz): δ ϭ 5.17 (d, J ϭ 3.5 Hz, 1 H, 1-H),
Oct-1-enyl Nonaflate (3j): According to Method A, 1j (0.500 g,
2.50 mmol; E/Z ϭ 4:96), [nBu4N]F solution (0.45 mmol) and NfF
5.28 (dd, J ϭ 3.5, 1.3 Hz, 1 H, 1-H), 5.41 (dq, J ϭ 11.1, 0.7 Hz, 1
H, 4-H), 5.65 (d, J ϭ 17.1 Hz, 1 H, 4-H), 6.29 (dd, J ϭ 17.1,
(0.906 g, 3.00 mmol), after purification by kugelrohr distillation at 11.1 Hz, 1 H, 3-H). 13C NMR (75.5 MHz): δ ϭ 107.0 (t, C-1), 118.7
100 °C/1 mbar, provided 0.893 g (87%) of 3j as a colourless oil (E/ (t, C-4), 129.1 (d, C-3), 152.1 (s, C-2). IR (film): ν˜ ϭ 3050 cmϪ1
Z ϭ 8:92). (Z)-3j: 1H NMR (300 MHz): δ ϭ 0.82Ϫ1.00, 1.21Ϫ1.48 (ϭCϪH), 1650 (CϭC), 1425, 1145 (SO2), 1240, 1205 (CϪF). GC/
(2 m, 4 H, 7 H, 4-H Ϫ 8-H), 2.19 (q, J ϭ 7.5 Hz, 2 H, 3-H), 5.23 MS-data: retention time: 4.16 min (74 °C). MS (EI, 70 eV): m/z
(td, J ϭ 7.5, 5.7 Hz, 1 H, 2-H), 6.57 (d, J ϭ 5.7 Hz, 1 H, 1-H). 13
C
(%) ϭ 352 (12) [Mϩ], 219 (10) [C4F9ϩ], 153 (100), 131 (27), 119
NMR (75.5 MHz): δ ϭ 13.9 (q, C-8), 22.5, 24.1, 28.4, 28.7, 31.5 (13) [C2F5ϩ], 100 (20) [C2F4ϩ], 69 (100) [CF3ϩ], 53 (50) [C4H5ϩ],
1
(5 t, C-3 Ϫ C-7), 120.3 (d, C-2), 135.5 (d, C-1). (E)-3j: H NMR 41 (35) [C3H5ϩ], 27 (16) [C2H3ϩ]. C8H5F9O3S (352.2): calcd. C
(300 MHz): δ ϭ 0.82Ϫ1.00, 1.58Ϫ1.71 (2 m, 4 H, 7 H, 4-H Ϫ 8-
H); 2.05 (q, J ϭ 7.8 Hz, 2 H, 3-H), 5.78 (td, J ϭ 12.1, 7.8 Hz, 1
H, 2-H), 6.56 (d*, 1 H, 1-H); * the coupling constant could not be
determined owing to very low intensity. 13C NMR (75.5 MHz): δ ϭ
27.28, H 1.43, S 9.10; found C 27.09, H 1.57, S 8.13.
Cation Effect on C- vs. O-Nonaflation of NfF of Enolates Derived
from Methyl Ketones. 1-(Nonafluorobutanesulfonyl)propan-2-one
(4l): A solution of n-butyllithium in hexane (2.44 molar, 4.9 mL,
12.0 mmol) was added at Ϫ78 °C to a solution of iPr2NH (1.58 g,
16.0 mmol) in THF (19 mL). The cooling bath was removed, the
resulting mixture was allowed to warm to room temp. and then
recooled to Ϫ78 °C, neat acetone (0.590 g, 10.0 mmol) was added
dropwise, and the resulting mixture was stirred for 0.5 h at Ϫ78 °C.
Then neat NfF (4.54 g, 15.0 mmol) was added to the generated
solution of lithium propen-2-olate. The reaction mixture was stirred
at Ϫ78 °C for 2 h, gradually allowed to warm to room temp. and
stirred for an additional 15 h. The resultant brownish solution was
then poured into a vigorously stirred mixture of pentane (100 mL)
and ice/aq. H3PO4 (100 mL, pH ϭ 2), the aqueous phase was ex-
tracted with pentane (3ϫ20 mL), and the combined organic phase
was dried (MgSO4). 1H NMR monitoring after removal of volatiles
13.8 (q, C-8), 20.4, 26.5, 28.5, 29.0, 31.5 (5 t, C-3 Ϫ C-7), 122.5 (d,
Ϫ1
˜
C-2), 136.2 (d, C-1). IR (film): ν ϭ 3120 cm (ϭCϪH), 2960,
2935, 2860 (CϪH), 1669 (CϭC), 1430, 1130 (CϪF), 1240, 1205
(SO2). GC/MS-data: (Z)-3j, retention time: 9.67 min (167 °C). MS
(EI, 70 eV): m/z (%) ϭ 410 (Ͻ 1) [Mϩ], 274 (10), 218 (19), 131 (23),
109 (21), 81 (40), 69 (100) [CF3ϩ], 55 (52) [C4H9ϩ], 43 (52) [C3H7ϩ],
29 (24) [C2H5ϩ]. (E)-3j: retention time: 10.02 min (170 °C). MS (EI,
70 eV): m/z (%) ϭ 410 (Ͻ 1) [Mϩ], 274 (25), 218 (28), 131 (30), 109
(12), 100 (13), 81 (26), 69 (100) [CF3ϩ], 55 (43) [C4H9ϩ], 43 (32)
[C3H7ϩ], 29 (17) [C2H5ϩ]. C12H15F9O3S (410.3): calcd. C 35.13, H
3.68, S 7.81; found C 35.55, H 3.79, S 7.32.
Non-1-enyl Nonaflate (3k): According to Method A, 1k (0.535 g,
2.50 mmol; E/Z ϭ 93:7), [nBu4N]F solution (0.45 mmol) and NfF
(0.906 g, 3.00 mmol), after purification by kugelrohr distillation at in vacuum (up to 100 mbar) showed a ratio of 3l/4l of 64:36. The
80 °C/1 mbar, provided 0.827 g (78%) of 3k as a colourless oil (E/ residue was subjected to CC (gradient elution: hexane Ǟ hexane/
1
Z ϭ 95:5). (E)-3k: H NMR (300 MHz): δ ϭ 0.87 (t, J ϭ 6.0 Hz, Et2O/AcOH, 200:10:1 Ǟ hexane/Et2O/AcOH, 50:10:1) to give
3 H, 9-H), 1.27Ϫ1.46 (m, 10 H, 4-H Ϫ 8-H), 2.03 (qd, J ϭ 7.7, 0.850 g (25%) of the pure sulfonyl ketone 4l as yellowish crystals,
1.2 Hz, 2 H, 3-H), 5.76 (dt, J ϭ 11.9, 7.7 Hz, 1 H, 2-H), 6.53 (br.
m.p. 53Ϫ57 °C. The compound 4l existed in CDCl3 solution as a
d, J ϭ 11.9 Hz, 1 H, 1-H). 13C NMR (75.5 MHz): δ ϭ 14.0 (q, C-
90:10 mixture of keto and enol forms. Keto tautomer, 1H NMR
1022
Eur. J. Org. Chem. 2002, 1015Ϫ1025

Synthesis of Alkenyl Nonaflates from Silyl Enol Ethers
FULL PAPER
(270 MHz): δ ϭ 2.49 (s, 3 H, 3-H), 4.32 (br. s, 2 H, 1-H). 13C NMR sulfonyl ketone 4m (0.380 g, 17%) and 3,3-dimethyl-1-
(67.5 MHz): δ ϭ 31.5 (q, C-3), 61.7 (t, C-1), 191.5 (s, C-2). Enol
(1,1,2,2,3,3,4,4-octafluorobutane-1-sulfonyl)butan-2-one (6,
1
tautomer, H NMR (270 MHz): δ ϭ 2.16 (s, 3 H, 3-H), 5.25 (br. s, 0.051 g, 2%). Compound 6 (yellowish oil, slowly crystallized upon
1 H, 1-H), 10.16 (br. s, 1 H, OH). 13C NMR (67.5 MHz): δ ϭ 22.8 storage in a refrigerator) existed in CDCl3 solution as an 84:16
(q, C-3), 87.5 (d, C-1); C-2 signal was not observed due to its low mixture of keto and enol forms. Keto tautomer, 1H NMR
Ϫ1
˜
intensity. IR (KBr): ν ϭ 3445 cm (br., OH), 2980, 2930 (CϪH),
(270 MHz): δ ϭ 1.23 (s, 9 H, CMe3), 4.45 (t, J ϭ 1.2 Hz*, 2 H, 1-
1730 (CϭO), 1360, 1140 (SO2), 1205 (br., CϪF). MS (EI, 80 eV): H), 6.09 [tt, 2J(1HϪ19F) ϭ 51.8, 3J(1HϪ19F) ϭ 5.0 Hz, 1 H, CF2H].
m/z (%) ϭ 340 (Ͻ 1) [Mϩ], 325 (3) [Mϩ Ϫ CH3], 219 (11) 13C NMR (125 MHz): δ ϭ 25.5 (q, CMe3), 45.9 (s, CMe3), 55.4 (t,
[CF3(CF2)3]ϩ, 200 (Ͻ 1) [C4F8]ϩ, 181 (Ͻ 1) [C4F7]ϩ, 169 (1) C-1), 107.4 [dtt, 1J(13CϪ19F) ϭ 255, 2J(13CϪ19F) ϭ 32 Hz, CF2H],
[CF3(CF2)2]ϩ, [CF2ϭCFCF2]ϩ, (41) 199.8 (s, C-2). 19F NMR (470 MHz):
131 (13) 121 Ϫ137.49 [d**,
[Mϩ Ϫ CF3(CF2)3], 119 (4) [CF3CF2]ϩ, 105 (1), 100 (8) [C2F4]ϩ, 69 2J(1HϪ19F) ϭ 51.8 Hz, 2 F, CF2H], Ϫ129.36, Ϫ122.19 (2 s**, 2 F,
δ
ϭ
(29) [CF3ϩ], 57 (6) [Mϩ Ϫ CF3(CF2)3SO2], 42 (100) [CH2ϭCϭO]ϩ.
2 F, C2F2, C3F2), Ϫ112.48 (br. t, J ϭ 12.8 Hz, 2 F, C1F2); * long
C7H5F9O3S (340.2): calcd. C 24.72, H 1.48; found C 24.92, H 1.34. range coupling with 19F nuclei; ** further complex splitting by
19FϪ19F and/or 1HϪ19F couplings. Enol tautomer, 1H NMR
1-(Nonafluorobutanesulfonyl)-3,3-dimethylbutan-2-one (4m): As de-
scribed above, a solution of n-butyllithium in hexane (2.40 molar,
5.0 mL, 12.0 mmol), iPr2NH (1.58 g, 16.0 mmol), 3,3-dimethylbu-
tan-2-one (1.00 g, 10.0 mmol) and NfF (4.54 g, 15.0 mmol) in THF
(19 mL) at Ϫ78 °C provided 4m, which was isolated by CC (gradi-
ent elution: hexane Ǟ hexane/Et2O/AcOH, 200:10:1) as colourless
crystals, m.p. 55Ϫ57 °C in 47% yield (1.80 g). The compound ex-
isted in CDCl3 solution as an 82:18 mixture of keto and enol forms.
Keto tautomer, 1H NMR (270 MHz): δ ϭ 1.231 (s, 9 H, CMe3),
4.47 (t, J ϭ 1.2 Hz*, 2 H, 1-H). 13C NMR (125 MHz): δ ϭ 25.4
(q, CMe3), 45.9 (s, CMe3), 55.5 (t, C-1), 199.8 (s, C-2) . 19F NMR
(470 MHz): δ ϭ Ϫ126.37 (t**, J ϭ 13.8 Hz, 2 F, C3F2), Ϫ121.69
(m, 2 F, 2 F, C2F2), Ϫ112.38 (br. t, J ϭ 13.8 Hz, 2 F, C1F2), Ϫ81.18
(tt, J ϭ 9.7, 2.2 Hz, 3 F, CF3); * long range coupling with 19F
nuclei; ** further complex splitting by 19FϪ19F and/or 1HϪ19F
couplings. Enol tautomer, 1H NMR (270 MHz): δ ϭ 1.227 (s, 9 H,
CMe3), 5.30 (s, 1 H, 1-H), 10.33 (s, 1 H, OH). 13C NMR
(125 MHz): δ ϭ 27.1 (q, CMe3), 38.8 (s, CMe3), 84.2 (d, C-1), 188.9
(s, C-2). 19F NMR (470 MHz): δ ϭ Ϫ126.50 (t*, J ϭ 14.0 Hz, 2 F,
C3F2), ca. Ϫ121.7** (m, 2 F, 2 F, C2F2), Ϫ115.29 (t*, J ϭ 13.7 Hz,
2 F, C1F2), Ϫ81.22 (tt, J ϭ 9.7, 2.4 Hz, 3 F, CF3); * further complex
splitting by 19FϪ19F couplings. ** the exact chemical shift could
not be determined owing to shielding by the corresponding signal
of the keto tautomer. IR (KBr): ν˜ ϭ 3420 cmϪ1 (br., OH), 2980,
2925 (CϪH), 1930, 1720 (CϭO), 1365, 1140 (SO2), 1235, 1205
(CϪF). MS (EI, 80 eV): m/z (%) ϭ 382 (3) [Mϩ], 367 (3)
[Mϩ Ϫ CH3], 325 (4) [Mϩ Ϫ tBu], 219 (2) [CF3(CF2)3]ϩ, 177 (3),
163 (100) [Mϩ Ϫ CF3(CF2)3], 131 (4) [CF2ϭCFCF2]ϩ, 121 (3), 119
(2) [CF3CF2]ϩ, 105 (2), 100 (2) [C2F4]ϩ, 99 (3)
[Mϩ Ϫ CF3(CF2)3SO2], 85 (3) [tBuCO]ϩ, 69 (12) [CF3ϩ], 57 (67)
[tBuϩ]. C10H11F9O3S (382.3): calcd. C 31.42, H 2.90; found C
31.22, H 2.65.
(270 MHz): δ ϭ 1.22 (s, 9 H, CMe3), 5.29 (s, 1 H, 1-H), 6.07 [tt,
3
2J(1HϪ19F) ϭ 51.8, J(1HϪ19F) ϭ 5.1 Hz, 1 H, CF2H], 10.32 (s, 1
H, OH). 13C NMR (125 MHz): δ ϭ 27.1 (q, CMe3), 38.7 (s, CMe3),
84.3 (d, C-1), 188.7 (s, C-2); CF2H signal was not observed due to
its low intensity or shielding by the corresponding signal of the keto
2
tautomer. 19F NMR (470 MHz): δ ϭ Ϫ137.53 [d*, J(1HϪ19F) ϭ
51.8 Hz, 2 F, CF2H], Ϫ129.74, Ϫ122.33 (2 s*, 2 F, 2 F, C2F2, C3F2),
Ϫ115.30 (t*, J ϭ 12.7 Hz, 2 F, C1F2); * further complex splitting
1
by 19FϪ19F and/or HϪ19F couplings.
Nonaflations of 1l and 1m Promoted by Tetrabutylammonium Hy-
drogen Difluoride: As described for Method B, 1 mmol scale in
THF (0.3 mL), in the presence of KF (0.120 g, 2.10 mmol),
[nBu4N][HF2] (0.105 mL) in place of [nBu4N]F and NfF (0.400 g,
1.30 mmol). After acidic aqueous workup, pentane extraction, dry-
ing and evaporation as described above, the contents of the organic
1
phases were analysed by H NMR. CC (pentane) provided 0.110 g
(32%) of 3l or 0.150 g (40%) of 3m as pure, colourless liquids.
Dibenzo-18-crown-6/KF as a Fluoride Promoter. Propen-2-yl Non-
aflate (3l): Potassium fluoride (0.058 g, 1.00 mmol) was dried at
230 °C (0.04 mbar) for 1 h. After this had cooled to room temp.,
dibenzo-18-crown-6 (0.072 g, 0.20 mmol) was added, and the mixed
powder was heated at 120 °C (0.04 mbar) for 10 min, after which
THF (1 mL) and silyl enol ether 1l (0.134 g, 1.00 mmol) were added
consecutively at room temp. The reaction mixture was stirred for
10 min at room temp., cooled to 0 °C, and NfF (0.404 g,
1.30 mmol) was added dropwise. After warming up to room temp.,
the mixture was stirred for 70 h. Acidic aqueous workup as de-
scribed in the procedure for the synthesis of 3m/4m, filtration
through celite, pentane extraction, drying and evaporation (up to
100 mbar) gave crude 3l (no signals of the ketone 4l were detected
by 1H NMR). Pure nonaflate 3l was isolated by CC (pentane) in
68% yield (0.231 g) as a colourless, volatile liquid. 1H NMR
(270 MHz): δ ϭ 2.10 (d, J ϭ 1.1 Hz, 3 H, 3-H), 4.95 (dq, J ϭ
3.4, 1.1 Hz, 1 H, 1-H), 5.09 (d, J ϭ 3.4 Hz, 1 H, 1-H). 13C NMR
Reaction between Potassium 3,3-Dimethylbut-1-en-2-olate and NfF:
Potassium ethoxide (0.510 g, 6.00 mmol) was dried at 100 °C
(0.04 mbar) for 1 h, after which THF (10 mL) was added. Silyl enol
ether 1m (1.04 g, 6.00 mmol) was added dropwise at Ϫ40 °C to the
resulting suspension. The mixture was allowed to warm to 5 °C
over 1 h, resulting in a nearly homogeneous, bright yellow solution
of the potassium enolate. After the mixture had been cooled to
Ϫ70 °C, neat NfF (2.20 g, 7.20 mmol) was added, and the reaction
mixture was allowed to warm to 10 °C over 2 h and stirred at this
temperature for 7 h. The resultant yellow solution was then poured
into a vigorously stirred mixture of pentane (100 mL) and ice/aq.
H3PO4 (100 mL, pH ϭ 2), the aqueous phase was extracted with
pentane (3ϫ20 mL), and the combined organic phase was dried
(MgSO4). 1H NMR monitoring after removal of volatiles in va-
cuum (up to 100 mbar) showed the 3m/4m ratio to be 75:25. The
residue was subjected to CC (gradient elution: pentane Ǟ hexane/
(67.5 MHz): δ ϭ 20.1 (q, C-3), 105.4 (t, C-1), 153.3 (s, C-2). IR
Ϫ1
˜
(film): ν ϭ 3000Ϫ2935 cm
(CϪH), 1679 (CϭC), 1422, 1145
(SO2), 1240, 1207 (CϪF). MS (EI, 80 eV): m/z (%) ϭ 340 (5) [Mϩ],
276 (6) [Mϩ Ϫ SO2], 219 (7) [CF3(CF2)3]ϩ, 131 (100) [CF2ϭ
CFCF2]ϩ, 119 (7) [CF3CF2]ϩ, 100 (10) [C2F4]ϩ, 81 (13) [C2F3]ϩ,
77 (8) [FCH2C(Me)ϭOH]ϩ, 75 (13) [C3H4FO]ϩ, 73 (50) [Oϭ
CHC(Me)ϭOH]ϩ, 69 (58) [CF3ϩ], 64 (4) [SO2ϩ], 57 (6)
[Mϩ Ϫ CF3(CF2)3SO2], 43 (3) [MeCϵO]ϩ, 42 (16) [CH2ϭCϭO]ϩ,
41 (5) [CH2ϭCHCH2]ϩ, 40 (75) [MeCϵCH]ϩ or [CH2ϭCϭCH2]ϩ,
39 (4) [C3H3]ϩ, 38 (10) [C3H2]ϩ or [F2ϩ], 31 (18) [CH3O]ϩ or [CFϩ],
28 (8) [COϩ] or [C2H4]ϩ, 27 (51) [C2H3]ϩ, 26 (8) [C2H2]ϩ. HRMS:
calcd. for C7H5F9O3S 339.98157; found 339.98333.
3,3-Dimethylbut-2-enyl Nonaflate (3m): A mixture of KF (0.058 g,
Et2O/AcOH, 200:10:1) to give the pure nonaflate 3m (1.07 g, 47%), 1.00 mmol) and dibenzo-18-crown-6 (0.072 g, 0.20 mmol) was
Eur. J. Org. Chem. 2002, 1015Ϫ1025 1023

I. M. Lyapkalo, M. Webel, H.-U. Reißig
FULL PAPER
dried as described above, and a mixture of DMF/THF 1:1 (1 mL) [Mϩ Ϫ CF3(CF2)3SO2], 180 (9) [Mϩ Ϫ CF3(CF2)3SO3], 168 (28)
was added. After this had been cooled to 0 °C, silyl enol ether 1m
[Mϩ Ϫ CF3(CF2)3SO2 Ϫ CO], 154 (100) [C8H12NO2]ϩ, 140 (4)
[C7H10NO2]ϩ, [CF2ϭCFCF2]ϩ,
131 (5) 126 (4)
(0.170 g, 1.00 mmol) was added, and the resulting suspension was
stirred at 0Ϫ5 °C for 10 min. Neat NfF (0.440 g, 1.40 mmol) was [C8H12NO2 Ϫ C2H4]ϩ, 124 (4) [C8H12NO2 Ϫ C2H4 Ϫ H2]ϩ, 100 (4)
added dropwise, and the resulting mixture was allowed to warm to
[C2F4ϩ], 96 (17) [C5H6NO]ϩ, 82 (19) [C5H8N]ϩ, 80 (7) [C5H6N]ϩ,
room temp. and stirred at this temperature for 70 h. Workup and 79 (7) [C5H5N]ϩ, 69 (17) [CF3ϩ], 68 (8) [C4H6N]ϩ, 67 (6)
purification as described above provided 0.244 g (64%) of nonafl-
ate 3m.
[C4H5N]ϩ, 28 (20 ) [HCϵNH]ϩ or [COϩ] or [C2H4]ϩ, 27 (5)
[C2H3]ϩ or [HCN]ϩ, 26 (5) [C2H2]ϩ or [CNϩ]. C14H14F9NO5S
(479.3): calcd. C 35.08, H 2.94, N 2.92; found C 35.07, H 2.74,
N 2.72.
Synthesis of Nonaflates from Ketones. Deprotonation with LDA.
Cyclohex-1-enyl Nonaflate (3c): A solution of n-butyllithium in hex-
ane (2.2 molar, 2.32 mL, 5.10 mmol) was added at Ϫ78 °C to a
solution of iPr2NH (0.516 g, 5.10 mmol) in THF (80 mL). After
this had remained for 1 h at Ϫ78 °C, cyclohexanone (0.490 g,
5.00 mmol) was added, followed by stirring for a further 1 h at Ϫ78
°C. Neat NfF (3.02 g, 10.0 mmol) was added dropwise, and the
reaction mixture was allowed to warm up to room temp. overnight
and then poured into the saturated aq. NH4Cl. The water phase
was extracted with EtOAc, and the combined organic phase was
washed consecutively with saturated aq. NaCl and water, dried
(Na2SO4) and filtered. The volatiles were removed in vacuum, and
the residue was subjected to kugelrohr distillation to furnish 3c
(1.23 g, 65%).
rac-8-Methyl-8-azabicyclo[3.2.1]oct-2-en-3-yl Nonaflate (3f): A so-
lution of n-butyllithium in hexane (2.5 molar, 5.6 mL, 14.0 mmol)
was added to a solution of iPr2NH (1.54 g, 15.0 mmol) in THF (12
mL) at Ϫ78 °C. The cooling bath was removed, the resulting mix-
ture was allowed to warm to room temp. and then once more co-
oled to Ϫ78 °C, and a solution of 8-methyl-8-azabicyclo[3.2.1]oc-
tan-3-one (tropinone) (1.43 g, 10.2 mmol) in THF (10 mL) was ad-
ded. After warming to Ϫ50 °C for 0.5 h, it was again cooled to
Ϫ78 °C, and neat NfF (6.04 g, 20.0 mmol) was added. The reaction
mixture was stirred at Ϫ78 °C for 2 h, allowed gradually to warm
to room temp. and stirred for additional 15 h. The resultant brown
solution was then poured into a vigorously stirred mixture of hex-
ane (100 mL) and ice/satd. aq. NaHCO3 (100 mL), the aqueous
phase was extracted with hexane (3ϫ30 mL), and the combined
organic phase was dried (Na2SO4). Filtration, removal of volatiles
in vacuo and subsequent CC (gradient elution: hexane Ǟ hexane/
Et2O, 20:1 Ǟ hexane/Et2O/Et3N, 8:2:1) gave 3f (3.72 g, 86%).
Nonaflation in the Presence of DBU. Cyclohex-1-enyl Nonaflate
(3c): A solution of cyclohexanone (0.245 g, 2.50 mmol) and DBU
(0.456 g, 3.00 mmol) in toluene (5 mL) was warmed, and neat NfF
(0.906 g, 3.00 mmol) was added at 85 °C over 3.5 h, followed by
stirring at this temperature overnight. Direct kugelrohr distillation
of the reaction mixture gave 3c (0.439 g, 48%).
Acknowledgments
Generous support by the Deutsche Forschungsgemeinschaft and
the Alexander von Humboldt foundation (research fellowship for
I. M. L.) is most gratefully acknowledged. We thank Bayer AG,
Leverkusen, for generous donations of nonafluorobutanesulfonyl
fluoride and Dr. R. Zimmer for help during preparation of this ma-
nuscript.
[1]
N,N-Bis(trifluoromethanesulfonyl)aniline (N-phenyltriflimide)
is perhaps the most popular tool for transfer of a triflyl group
to an enolate oxygen, although N-(5-chloro-2-pyridyl)triflimide
and N-(2-pyridyl)triflimide appear to be more efficient at low
temperature; see ref.[4b] and references cited therein.
[2]
According to recent catalogues of chemical suppliers, the prices
of these aryl triflimides are approximately 5Ϫ15 times higher
than that of Tf2O.
[3]
˜
W. J. Scott, M. R. Pena, K. Swärd, S. J. Stoessel, J. K. Stille, J.
Org. Chem. 1985, 50, 2302Ϫ2308; see also ref.[4b]
[4] [4a]
W. J. Scott, J. E. McMurry, Acc. Chem. Res. 1988, 21,
[4b]
[4c]
47Ϫ54.
K. Ritter, Synthesis 1993, 735Ϫ762.
S. Cacchi,
R. Brückner, J. Suffert,
J. F. Hartwig, Angew. Chem. 1998,
[4d]
Pure Appl. Chem. 1996, 68, 45Ϫ52.
Synlett 1999, 657Ϫ679.
[4e]
110, 2154Ϫ2177; Angew. Chem. Int. Ed. 1998, 37, 2047Ϫ2067.
rac-8-Ethoxycarbonyl-8-azabicyclo[3.2.1]oct-2-en-3-yl
Nonaflate
[5] [5a]
S. Bräse, A. de Meijere, Angew. Chem. 1995, 107,
(3o): As described above, a solution of n-butyllithium in hexane
(2.24 molar, 5.3 mL, 11.9 mmol), iPr2NH (1.43 g, 14.3 mmol), N-
carbethoxy-4-tropinone (1.87 g, 9.48 mmol, added dropwise as a
solution in 1.5 mL THF followed by stirring at Ϫ78 °C before the
addition of NfF) and NfF (4.30 g, 14.3 mmol) in THF (20 mL) at
Ϫ78 °C provided 3o, which was isolated by CC (gradient elution:
hexane Ǟ hexane/Et2O, 20:1 Ǟ hexane/Et2O, 8:1 Ǟ hexane/Et2O,
4:1) in 94% yield (4.27 g) as a yellow oil. 1H NMR (270 MHz): δ ϭ
1.24 (t, J ϭ 7.1 Hz, 3 H, OCH2Me), 1.69Ϫ1.81, 1.98Ϫ2.06, 2.11,
2.16Ϫ2.31, 3.02 (2 m, br. d, J ϭ 17.2 Hz, m, br. m, 1 H, 2 H, 1 H,
1 H, 1 H, 4-H, 6-H, 7-H), 4.13 (br. q, J ϭ 7.1 Hz, 2 H, OCH2Me),
4.40Ϫ4.66 (br. m, 2 H, 1-H, 5-H), 6.09 (br. d, J ϭ 5.4 Hz, 1 H, 2-
H). 13C NMR (67.5 MHz): δ ϭ 14.5 (q, OCH2Me), 28.9, 34.8, 36.5
(3 br. t, C-4, C-6, C-7), 51.8, 51.9 (2 d, C-1, C-5), 61.4 (t,
OCH2Me), 123.2 (br. d, C-2), 154.1 (s, C-3 or CϭO); one of the
[5b]
2741Ϫ2743; Angew. Chem. Int. Ed. Engl. 1995, 34, 2545.
M. Webel, H.-U. Reißig, Synlett 1997, 1141Ϫ1142. [5c] S. Bräse,
[5d]
Synlett 1999, 1654Ϫ1656.
nische Universität Dresden, 2000.
M. Webel, Dissertation, Tech-
[5e]
I. M. Lyapkalo, M.
Webel, H.-U. Reißig, manuscript in preparation.
TfF and NfF are routinely obtained by electrochemical fluor-
ination of MeSO2Cl or MeSO3H and 2,5-dihydrothiophene
1,1-dioxide,[8] respectively.
Nonafluorobutanesulfonyl fluoride is slightly cheaper per
mmol than Tf2O.
R. Zimmer, M. Webel, H.-U. Reißig, J. Prakt. Chem. 1998, 340,
274Ϫ277 and references cited therein.
For synthesis of aryl nonaflates from aryl trimethylsilyl ethers
[6]
[7]
[8]
[9]
[9a]
or phenols, see:
H. Niederprüm, P. Voss, V. Beyl, Justus
[9b]
Liebigs Ann. Chem. 1973, 20Ϫ32.
L. R. Subramanian, A.
Garcia Martinez, A. H. Fernandez, R. M. Alvarez, Synthesis
[9c]
1984, 481Ϫ485.
M. Rottlaender, P. Knochel, J. Org. Chem.
carbon signals (C-3 or CϭO) was not observed due to overlapping
1998, 63, 203Ϫ208.
For occasional preparations, see:
U. Reißig, Chem. Ber. 1982, 115, 3687Ϫ3696.
Ϫ1
˜
or strong broadening. IR (film): ν ϭ 3600Ϫ3400 cm (ϭCϪH),
[10]
[10a]
E. Hirsch, S. Hünig, H.-
3075, 2985, 2920, 2880 (CϪH), 1710 (br., CϭO), 1420, 1145 (SO2),
1240, 1200 (CϪF). MS (EI, 80 eV): m/z (%) ϭ 479 (40) [Mϩ], 450
(5) [Mϩ Ϫ Et], 434 (3) [Mϩ Ϫ OEt], 406 (Ͻ 3) [Mϩ Ϫ CO2Et], 391
(5) [Mϩ Ϫ CF4], 378 (7) [Mϩ Ϫ CH2NCO2Et], 314 (4)
[Mϩ Ϫ CH2NCO2Et Ϫ SO2], 219 (5) [CF3(CF2)3]ϩ, 196 (21)
[10b]
U. Hert-
enstein, S. Hünig, H. Reichelt, R. Schaller, Chem. Ber. 1986,
119, 699Ϫ721. [10c] L. R. Subramanian, H. Bentz, M. Hanack,
[10d]
Synthesis 1973, 293Ϫ294.
Rosauer, A. Lansky, A. Adams, O. Reiser, A. de Meijere, Eur.
K. Voigt, P. von Zezschwitz, K.
1024
Eur. J. Org. Chem. 2002, 1015Ϫ1025

Synthesis of Alkenyl Nonaflates from Silyl Enol Ethers
FULL PAPER
D. Seebach, A. K. Beck, T. Mukhopadhyay, E. Thomas, Helv.
Chim. Acta 1982, 65, 1101Ϫ1133.
D. P. Cox, J. Terpiniski, W. Lawrynowicz, J. Org. Chem. 1984,
49, 3216Ϫ3219.
It is [nBu4N]F that catalyses the nonaflation, as the reference
control experiment conducted in the presence on KF without
[nBu4N]F under otherwise identical conditions showed no con-
version of 1b to 3b at all.
G. Stork, P. F. Hudrlik, J. Am. Chem. Soc. 1968, 90,
4464Ϫ4465.
I. M. Lyapkalo, M. Webel, H.-U. Reißig, Eur. J. Org. Chem.
2001, 4189Ϫ4194.
We observed partial decomposition of 3n when it was passed
through a pad of silica gel. The instability of the closely related
3-methylbuta-1,3-dien-2-yl triflate has been reported: L. Luan,
J.-S. Song, R. M. Bullock, J. Org. Chem. 1995, 60, 7170Ϫ7176.
For the facile generation of potassium enolates from trimethyl
silyl enol ethers by equimolar amounts of potassium ethoxide,
see: W. Yu, Z. Jin, Tetrahedron Lett. 2001, 42, 369Ϫ372.
In an early study, we applied stoichiometric amounts of 18-
crown-6 in an attempt to modify the reactivity of potassium
3,3-dimethyl-buten-2-olate;[10a] We found only a few examples
of crown ether/KF-induced O-methanesulfonylation or O-acyl-
ation of silyl enol ethers: U. Hertenstein, S. Hünig, H. Reichelt,
R. Schaller, Chem. Ber. 1982, 115, 261Ϫ287. R. A. Olofson, J.
Cuomo, Tetrahedron Lett. 1980, 21, 819Ϫ822. W. Ando, H.
Tsumaki, Tetrahedron Lett. 1982, 23, 3073Ϫ3076.
The implied C-nonaflation product 4n would be expected to
be rather unstable and prone to rapid polymerization, since it
contains nucleophilic and electrophilic moieties.
The observed inertness of the tert-butyldimethylsilyl enol ethers
in fluoride-induced nonaflation with NfF provides evidence of
high chemoselectivity of the reaction, which may be useful in
synthesis when the selective transformation of a trimethyl silyl
enol functionality in the presence of a TBDMS ether is re-
quired.
[10e]
[25]
[26]
[27]
J. Org. Chem. 1998, 1521Ϫ1534.
F. Bellina, D. Ciucci, R.
[10f]
Rossi, P. Vergamini, Tetrahedron 1999, 55, 2103Ϫ2112.
T.
Okauchi, T. Yano, T. Fukamachi, J. Ichikawa, T. Minami, Tet-
rahedron Lett. 1999, 40, 5337Ϫ5340.
[11]
Recently, some steroid 3-ketones were converted into poly- and
perfluoroalkanesulfonates by use of the respective alkanesul-
fonyl fluorides in the presence of DBU; see: Z. Zhu, W. Tian,
Q. Liao, Tetrahedron Lett. 1996, 37, 8553Ϫ8556. Since the con-
ditions reported are rather drastic (toluene, 90 °C, 3 h) this
procedure can hardly be applied to the synthesis of acyclic enol
nonaflates, owing to the ease of concomitant DBU-induced
NfOH elimination to afford alkynes and/or allenes.[12a]
[28]
[29]
[30]
[12] [12a]
P. J. Stang, M. Hanack, L. R. Subramanian, Synthesis
1982, 85Ϫ126. [12b] M. Hanack, G. Auchter, J. Am. Chem. Soc.
1985, 107, 5238Ϫ5245.
[13]
Because of the strongly electron-withdrawing effect of per-
fluoroalkyl substituents, and hence the substantial positive
charge on sulfur RFSO2X, the reaction is subject to charge con-
trol in most cases, resulting in RFSO2-attack on the negatively
charged carbonyl oxygen to give the desired alkenyl sulfonates.
For a systematic study of the counterion effect on the regiose-
lectivity of enolate sulfonylation, see ref.[10a]
[31]
[32]
[14]
For the review articles, see: [14a] P. Brownbridge, Synthesis 1983,
[14b]
1Ϫ28; ibid. 85Ϫ104.
I. Fleming, A. Barbero, D. Walter,
Chem. Rev. 1997, 97, 2063Ϫ2192.
[15] [15a] K. Sugasawa, M. Shindo, H. Noguchi, K. Koga, Tetrahed-
[15b]
ron Lett. 1996, 37, 7377Ϫ7380.
I. Vaulont, H.-J. Gais, N.
[33]
[34]
Reuter, E. Schmitz, R. K. L. Ossenkamp, Eur. J. Org. Chem.
1998, 805Ϫ826.
[16] [16a]
H. Izawa, R. Shirai, H. Kawasaki, H.-d. Kim, K. Koga,
[16b]
Tetrahedron Lett. 1989, 30, 7221Ϫ7224.
T. Momose, M.
Toshima, N. Toyooka, Y. Hirai, C. H. Eugster, J. Chem. Soc.,
Perkin Trans. 1 1997, 1307Ϫ1313.
An equilibrium between the enolate and the anionic silicate
complex formed with Me3SiF has been observed; see: R.
Noyori, I. Nishida, J. Sakata, J. Am. Chem. Soc. 1983, 105,
1598Ϫ1608.
[17]
[35]
The sulfur atom in nonafluorobutanesulfonyl fluoride is sur-
rounded with highly electronegative substituents and so may
be prone to form intermediate five-coordinated S-centred ate
complexes followed by CϪS bond cleavage: CF3(CF2)3SO2F ϩ
[18]
[19]
For the preparation and drying see: D. Landini, H. Molinari,
M. Penso, A. Rampoldi, Synthesis 1988, 953Ϫ955.
This procedure implies the intermediacy of highly nucleophilic
trimethyl silicide anion (Me3SiϪ), which may interact with
other functionalities in a substrate; see: A. Mori, A. Fujita, K.
Ikegashira, Y. Nishihara, T. Hiyama, Synlett 1997, 693Ϫ694.
H. Vorbrüggen, K. Krolikiewicz, Acc. Chem. Res. 1995, 28,
509Ϫ520.
Recently, CsF was reported to give high yields of the bis-non-
aflates from bistrimethylsilyl enol ethers derived from substi-
tuted cyclohexan-1,3-diones: see ref.[5c] Fused salts:
CsFϪCsOH and CsFϪCsCl have been found to be very effici-
ent for nucleophilic activation of silyl enol ethers and silyl ket-
ene acetales; see: J. Busch-Petersen, Y. Bo, E. J. Corey, Tetra-
hedron Lett. 1999, 40, 2065Ϫ2068.
R. Schwesinger, R. Link, G. Thiele, H. Rotter, D. Honert, H.-
H. Limbach, F. Männle, Angew. Chem. 1991, 103, 1376Ϫ1378,
Angew. Chem. Int. Ed. Engl. 1991, 30, 1372.
H. Vorbrüggen, K. Krolikiewicz, Tetrahedron Lett. 1984, 25,
1259Ϫ1262.
We believe that this occurs because of the low reactivity of
NfF, as a similar acylation procedure with more reactive acyl
fluorides in the presence of commercially available catalytic
[nBu4N]F in THF resulted in high yields of the carboxylic acid
enol esters; see: D. Limat, M. Schlosser, Tetrahedron 1995, 51,
5799Ϫ5806.
F
Ϫ Ǟ [CF3(CF2)3SO2F2]Ϫ Ǟ [CF3(CF2)2CF2]Ϫ ϩ SO2F2. This
kind of precomplexation followed by fragmentation may inter-
fere the enolate nonaflation process when fluoride-induced en-
olate generation is hampered by bulky substituents at a Si
centre, as in tert-butyldimethylsilyl enol ethers. For the forma-
tion of a stable five-coordinated ate complex from a sulfonic
ester, see: C. W. Perkins, S. R. Wilson, J. C. Martin, J. Am.
Chem. Soc. 1985, 107, 3209Ϫ3218. Fluoride-induced CϪS
bond cleavage in 2,2,2-trifluoro(1-trifluoromethyl)ethyl nonafl-
ate, resulting in 1,1,2,2,3,3,4,4,4-nonafluorobutane, is docu-
mented in the literature: M. Hanack, J. Ullmann, J. Org. Chem.
1989, 54, 1432Ϫ1435.
S. Matsuzawa, Y. Horiguchi, E. Nakamura, I. Kuwajima, Tet-
rahedron 1989, 45, 349Ϫ362.
P. F. Hudrlik, A. M. Hudrlik, A. K. Kulkarni, J. Am. Chem.
Soc. 1985, 107, 4260Ϫ4264.
C. Hippeli, N. Basso, F. Dammast, H.-U. Reißig, Synthesis
1990, 26Ϫ27.
S. C. Watson, J. F. Eastham, J. Organomet. Chem. 1967, 9,
165Ϫ168.
[20]
[21]
[36]
[37]
[38]
[39]
[40]
[22]
[23]
[24]
K. M. Brummond, E. A. Dingess, J. L. Kent, J. Org. Chem.
1996, 61, 6096Ϫ6097.
Received September 18, 2001
[O01450]
Eur. J. Org. Chem. 2002, 1015Ϫ1025
1025