Ethyl difluoro(trimethylsilyl)acetate and difluoro(trimethylsilyl) acetamides - Precursors of 3,3-difluoroazetidinones

Article abstract of DOI:10.1002/ejoc.200600146

Difluoro(trimethylsilyl)acetamides 7 have been prepared from chlorodifluoroacetamides 5 by electrochemical silylation. When condensed with carbonyl compounds, they were shown to be precursors of 2,2-difluoro-3- hydroxyacetamides 8. N-(p-Methoxyphenyl)-2,2-difluoro-3-hydroxy-4- methylvaleramide (8a) has been converted into the correspending 3,3-difluoroazetidinone 9a. This new route to 3,3-difluoroazetidinones is shown to be an alternative to the one utilizing the condensation of ethyl difluoro(trimethylsilyl)acetate with imines. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejoc.200600146

FULL PAPER
DOI: 10.1002/ejoc.200600146
Ethyl Difluoro(trimethylsilyl)acetate and Difluoro(trimethylsilyl)acetamides –
Precursors of 3,3-Difluoroazetidinones[‡]
Michel Bordeau,*[a] Frédéric Frébault,[a] Mallory Gobet,[a] and Jean-Paul Picard*[a]
Keywords: Fluorine / Silicon / Amino acids / Amides / Lactams / Electrosynthesis
Difluoro(trimethylsilyl)acetamides
from chlorodifluoroacetamides
lylation. When condensed with carbonyl compounds, they
7
have been prepared
sponding 3,3-difluoroazetidinone 9a. This new route to 3,3-
difluoroazetidinones is shown to be an alternative to the one
utilizing the condensation of ethyl difluoro(trimethylsilyl)ace-
5
by electrochemical si-
were shown to be precursors of 2,2-difluoro-3-hydroxyacet- tate with imines.
amides 8. N-(p-Methoxyphenyl)-2,2-difluoro-3-hydroxy-4-
methylvaleramide (8a) has been converted into the corre-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Several general strategies have been used to synthesize
the 3,3-difluoroazetidinone skeleton (Scheme 2); they
Introduction
Azetidin-2-one structures (“β-lactams”) are present in mainly involve internal cyclization[4] either of β-amino-α,α-
many natural and nonnatural compounds that have proven difluoro esters A, with elimination of the corresponding
to be of interest as antibiotics. Although they have been alcohols,[5–7] or of 2,2-difluoro-3-hydroxy amides B, with
extracted from natural sources, great efforts have been dedi- elimination of water.[1,8,9] Access to the starting materials
cated to their synthesis both at academic and at industrial has been achieved through the condensation of the appro-
levels. The continuous need for new types of β-lactams priate Reformatsky reagents derived from an ethyl halodi-
emerged when bacterial resistance to some of these drugs fluoroacetate with either imines (to form A) or aldehydes
appeared.
or ketones to form 2,2-difluoro-3-hydroxy esters AЈ, which
Among the possible means to address this problem, the were converted into the desired amides B.
preparation of fluorinated azetidinones has been proposed
In both cases, Reformatsky reagents were produced
as one good response because the substitution of a methyl- from bromo- or iododifluoroacetates, which are very ex-
ene group by a difluoromethylene one induces changes in pensive materials. Ethyl chlorodifluoroacetate constitutes a
physical and chemical, and consequently biological, behav- more attractive starting reagent, as it is commercially avail-
ior. 3,3-Difluoro-β-lactams have been already synthesized, able and relatively cheap. We anticipated that ethyl difluoro-
and some of them are bioactive[1–3] (Scheme 1).
(trimethylsilyl)acetate (1), a recently described modified
Scheme 1. Bioactive 3,3-difluoroazetidin-2-ones.
Ruppert reagent,[10,11] might be used advantageously as it
is easily prepared from ethyl chlorodifluoroacetate and is
storable. Moreover, 1 had been demonstrated to react
efficiently with aldehydes and ketones to form 2,2-di-
fluoro-3-hydroxyacetates in good to excellent yields
(Scheme 3).[10,11]
[‡] Functionalized Ruppert Reagents, 3. Part 2: in preparation.
Part 1: Ref.[10]
[a] Laboratoire de chimie organique et organométallique, Uni-
versité Bordeaux 1,
33405 Talence Cedex, France
E-mail: m.bordeau@lcoo.u-bordeaux1.fr
j-p.picard@lcoo.u-bordeaux1.fr
Eur. J. Org. Chem. 2006, 4147–4154
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4147

M. Bordeau, F. Frébault, M. Gobet, J.-P. Picard
FULL PAPER
Scheme 2. Reformatsky routes to 3,3-difluoroazetidin-2-ones.
Scheme 3. Access to 2,2-difluoro-3-hydroxyacetates from 1.
fluoroacetic acid or its sodium salt by described meth-
ods.[14,15] Yields were slightly higher from the salt than from
the free acid, because of the volatility of the acid
(Scheme 6).
Results
We extended this study to the condensation of 1 with an
imine (namely a Schiff base) and obtained the correspond-
ing ethyl 3-amino-2,2-difluoroacetate 2.[12] The reaction re-
quired the use of 1 equiv. of fluoride ion, however, imines
being less nucleophilic in character than carbonyl com-
pounds. This β-amino ester was easily transformed into the
already known[5] azetidinone 3 on treatment with a base as
shown in Scheme 4.
However, we have not yet been able to extend the scope
of this reaction to other types of imines. Use of N-benzyl-
idenebenzylamine, for instance, resulted in desilylation of 1
(forming ethyl difluoroacetate) and silylation of the starting
imine at the benzylic position (compound 4).[13] This could
be indicative of a relatively basic character of the intermedi-
ate anionic species (Scheme 5).
Scheme 6. Syntheses of chlorodifluoroacetamides 5.
Direct condensation of chlorodifluoroacetyl chloride (6)
with these amines by described procedures[16–18] gave the
corresponding amides, but in low yields, probably because
Scheme 5. Attempted addition of 1.
We thus turned to a reversed strategy in which the azetid- of the low boiling point of the acyl chloride (bp. 26–28 °C
inone precursors were 2,2-difluoro-3-hydroxyacetamides, at atmospheric pressure).
which required the prior synthesis of a series of difluoro(tri-
The chloro amides were therefore electrochemically si-
methylsilyl)acetamides 7, a new class of functionalized Rup- lylated by the aluminium sacrificial anode technique[10] as
pert reagents. The amides were prepared by condensation in the case of the chloro ester to yield the corresponding
of the appropriate primary amines with either chlorodi- trimethylsilyl amides 7 (Scheme 7).
Scheme 4. Synthesis of azetidinone 3.
4148
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4147–4154

Difluoro(trimethylsilyl)acetyl Precursors of 3,3-Difluoroazetidinones
FULL PAPER
An alternate synthesis of 8a by direct amidation of the
corresponding hydroxy ester (Scheme 10) has been re-
ported.[8,19]
Scheme 10. Synthesis of 8a from the corresponding 3-hydroxy ester.
Scheme 7. Synthesis of difluoro(trimethylsilyl)acetamides 7.
Hydroxy amides 8 were subjected to cyclization reaction
in the presence of the Mitsunobu reagent [diethyl azodicar-
boxylate (DEAD)/triphenylphosphane] to yield the ex-
pected azetidinones 9. It should be pointed out that cycliza-
tion of 3-hydroxy amides as a synthetic route to β-lactams
is not very popular.[18,20–23]
This strategy was actually efficient only in the case of 8a
(Scheme 11), whereas 8b and 8c were recovered unchanged.
Use of diisopropyl diazodicarboxylate (DIAD),[24] reported
to be more active in Mitsunobu reactions, did not give bet-
ter results.
We also treated 5 with chlorotrimethylsilane and magne-
sium in THF according to Uneyama’s procedure for aryl
esters.[11b] Only 7a was obtained in comparable yield. This
result parallels Uneyama’s observation that aliphatic esters
could not be silylated in this way.
The GC/MS data of amides 7a–c each present an m/z
·
peak at [M – 51], indicating the loss of the [HCF2 ] fragment
(Scheme 8). This could be the result of a C Ǟ N or a C Ǟ
O silyl migration induced by heat on the GC column or in
the injection chamber to form either the N-silyldifluoro-
acetamide or the O-silyliminodifluoroacetate.
The relative intensities of these [M+ – 51] peaks vary with
the nature of the R group and can be related to the elec-
tronic density on the nitrogen or oxygen atom.
As in the case of the silyl esters, compounds 7 were con-
densed with an aldehyde, namely isobutyraldehyde, in the
presence of a catalytic amount of fluoride anion to form
the desired 2,2-difluoro-3-hydroxyacetamides 8 (Scheme 9).
Scheme 11. Synthesis of 3,3-difluoroazetidinone 9a from amide 8a.
The mechanism of this transformation is based on the
formation of an ylide between triphenylphosphane and
DEAD and on deprotonation of NH of the amide function
by the ylide in preference to OH. Finally, the oxygen atom
of the hydroxy group attacks the phosphorus atom to form
diethyl hydrazodicarboxylate and the azetidinone
(Scheme 12). The failure to cyclize 8b or 8c is probably due
to the proton on the nitrogen atom being less acidic than
the one of the hydroxy group.
Although a previous synthesis of 9a had already been
reported,[18,23] no physicochemical data for this azetidinone
had been given, so its full structural description was estab-
lished in this work by 1H, 13C, and 19F NMR spectroscopy,
Scheme 9. Synthesis of 2,2-difluoro-3-hydroxyacetamides 8.
No asymmetric induction had occurred, as hydroxy ester complemented by partial simulation of the spectra to ascer-
8b was obtained as a 1:1 mixture of two diastereomers. tain coupling constants.
Scheme 8. Thermal behavior of 7 and relative magnitude of the loss of the CHF2 fragment.
Eur. J. Org. Chem. 2006, 4147–4154
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4149

M. Bordeau, F. Frébault, M. Gobet, J.-P. Picard
FULL PAPER
Scheme 12. Proposed Mitsunobu reaction mechanism.
cence indicator 60 F254). Specific rotations of chiral compounds
dissolved in chloroform were obtained at 20 °C with a P3001 Krüss
Optronic automatic digital polarimeter and a 10 cm long tube.
Conclusions
Modified Ruppert’s reagents 1 and 7 provide original
routes to 3-amino-2,2-difluoro ester 2 and 2,2-difluoro-3- Slow additions were performed with a Razel A-99 automatic sy-
ringe. Dichloromethane and DMF were distilled from calcium hy-
dride, THF and Et2O from sodium/benzophenone ketyl, and
TMSCl from magnesium turnings. Electrosyntheses were per-
formed by use of the described equipment and procedures.[10]
hydroxy amides 8a–c and to 3,3-difluoroazetidinones 3 and
9a. Since these cyclic compounds are excellent precursors
of β-amino acids,[23] after N-deprotection and basic hydrol-
ysis, this strategy is very interesting, as fluorinated β-hy-
droxy amides can be converted through a two-step process
into the corresponding fluorinated β-amino acids, ready to
be incorporated into polypeptides chains (Scheme 13).
Synthesis of Ethyl 2,2-Difluoro-2-(trimethylsilyl)acetate (1): This si-
lyl ester was synthesized as in the literature.[10] Spectroscopic data
are identical to those reported in the literature[10] and are sup-
1
plemented as: B.p. 55 °C (20 Torr). H NMR (CDCl3): δ = 0.16 (s,
3
3
9 H, Si–CH3), 1.27 (t, 3 H, J = 7 Hz, CH2CH3), 4.24 (q, 2 H, J
= 7 Hz, CH2CH3) ppm. 13C NMR (CDCl3): δ = –5.23 (Si–CH3),
1
13.92 (CH2–CH3), 62 (CH2–CH3), 120.95 (t, JC,F = 268.9 Hz,
CF2), 166.24 (t, JC,F = 25.7 Hz, C = O) ppm. 19F NMR (CDCl3):
2
δ = –123.5(s) ppm. IR (KBr): ν = 2959, 1756 (C=O), 1255, 1114,
˜
1052 cm–1. MS (EI): m/z (%) = 196 (1) [M+·], 181 (5), 153 (7), 125
(8), 103 (10), 81 (17), 77 (38), 73 (100), 55 (15), 45 (21).
Addition of 1 to a Schiff Base to Provide Amino Ester 2: Under
anhydrous nitrogen static pressure, N-benzylideneaniline (3.63 g,
20 mmol), KF (1.27 g, 21.5 mmol), and HMPA (20 mL) were intro-
duced into a round-bottomed flask fitted with a condenser and a
magnetic stirring bar. The mixture was stirred at 50 °C, 1 (3.92 g,
20 mmol) was added, and stirring was continued for 12 h. The or-
ganic phase was extracted with diethyl ether, the resulting ethereal
solution was separated, and the diethyl ether was evaporated. The
crude product was purified by chromatography on silica gel (pen-
tane/Et2O, 19:1). Amino ester 2 was recovered in 25% yield.
Scheme 13. Possible route to 3-amino-2,2-difluoroaliphatic acids
(and esters) from 2,2-difluoro-3-hydroxyaliphatic amides 8 by well-
known β-lactam chemistry transformations.
Moreover, zinc is not particularly recommended in bio-
logical media because it can cause undesired effects. In this
context, our route to fluorinated β-lactam derivatives offers
an interesting alternative. Further studies are in progress.
Ethyl 2,2-Difluoro-3-phenyl-3-(phenylamino)propionate (2): M.p.
Experimental Section
General: 1H and 13C NMR spectra were recorded with a Bruker
3
104 °C. 1H NMR (CDCl3): δ = 1.26 (t, J = 7.2 Hz, 3 H, Ha), 4.28
(q, 3J = 7.2 Hz, 2 H, Hb), 4.46 (d, 3J = 9. 8 Hz, 1 H, NH), 5.10
3
3
3
AC 250 FT (1H NMR: 250 MHz, 13C NMR: 62.9 MHz), a Bruker (ddd, JH,F = 7.7 Hz, JH,F = 18.8 Hz, JH,H = 9.8 Hz, He), 6.65
n
m
3
3
3
DPX 300 (1H NMR: 300 MHz, 13C NMR: 75 MHz) or a Bruker
DPX 400 (1H NMR: 400 MHz, 13C NMR: 100 MHz) spectrometer
(d, J = 7.6 Hz, 2 H, Hk), 6.72 (t, J = 7.6 Hz, 2 H, Hl), 7.14 (t, J
= 7.6 Hz, 1 H, Ho) and 7.60 (m, 5 H, Harom in Ph-C) ppm. 13C
with use of CDCl3 (δ = 77.7 ppm) or residual CHCl3 (δ = 7.27 ppm) NMR (CDCl3): δ = 13.86 (Ca), 60.12 (dd, 2JC,F = 26.9 Hz, JC,F
=
=
2
1
1
as internal standards. 19F NMR spectra were obtained in CDCl3
with a Bruker AP 200 spectrometer (188 MHz), with PhCF3 as the
internal standard (δ = –63 ppm versus FCCl3). IR spectra were
recorded with a Perkin–Elmer Paragon 1000 FT/IR spectrometer.
Coupled GC/MS analyses were performed with a Finnigan inte-
grated spectrometer. Liquid chromatography was performed on sil-
ica gel (0.040–0.063 mm) and progress of the reactions was moni-
tored by TLC on silica gel plates (thickness 0.2 mm with fluores-
22.5 Hz, Ce), 63.19 (Cb), 114.29 (dd, JC,F = 257.4 Hz, JC,F
256.3 Hz, Cd), 115.18 (Ck), 119.14 (Co), 128 (Cl), 145.33 (Cj), 128–
2
129 (Cg, Ch and Ci), 133.84 (Cf) and 163.54 (dd, JC,F = 30.7 Hz,
2JC,F = 32.9 Hz, Cc) ppm. 19F NMR (CDCl3): δ = –109.55 (dd, Fm
3
3
2
or Fn) and –119.55 (ddd, JH,F = 7.7 Hz, JH,F = 18.8 Hz, JF,F
=
258 Hz, Fn or Fm) ppm. From the 1H and 19F NMR spectra, fluor-
ine atoms are observed to be diastereotopic, owing to the presence
of a close chiral center. IR (KBr): ν = 3354 (NH), 3035, 2985, 1760
˜
4150
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4147–4154

Difluoro(trimethylsilyl)acetyl Precursors of 3,3-Difluoroazetidinones
FULL PAPER
(C=O), 1603, 1497, 1455, 1442, 1372, 1298, 1266, 1255, 1212, 1194, Syntheses of Chlorodifluoroacetamides 5: Three methods were used.
1107, 1076, 1062, 882, 833, 756, 724, 692 cm–1. MS (EI): m/z (%)
= 305 (3) [M+·], 259 (2), 195 (12), 182 (59), 180 (88), 105 (23), 104
(24), 77 (100), 51 (35). C17H17F2NO2 (305.32): calcd. C 66.87, H
5.61, N 4.59; found C 66.94, H 5.58, N 4.60.
Method A (from the Sodium Salt):[14] PPh3 (3.93 g, 15 mmol,
1 equiv.) and iodine (3.81 g, 15 mmol, 1 equiv.), dissolved in anhy-
drous dichloromethane (150 mL), were introduced into a two-
necked 250 mL flask fitted with a dropping funnel and a condenser.
After homogenization of the mixture, sodium chlorodifluoroacetate
(2.29 g, 15 mmol, 1 equiv.) was added and the mixture was stirred
for 1 h. A mixture of p-anisidine (1.50 mL, 16.5 mmol, 1.1 equiv.),
NEt3 (2.3 mL, 16.5 mmol, 1.1 equiv.), and anhydrous CH2Cl2
(10 mL) was added dropwise. Stirring was maintained for 30 min
and the formed sodium salts were then separated by filtration
through a fritted glass filter (porosity 3). The solid residue was
washed with a saturated aqueous sodium thiosulfate solution
(2×20 mL), a HCl (10%) solution, a saturated aqueous sodium
hydrogencarbonate solution (10 mL), and finally with brine
(20 mL). The final organic phase was separated by decantation and
dried with MgSO4, and the solvent (CH2Cl2) was removed under
vacuum. The crude product was purified by column chromatog-
raphy (light petroleum ether/EtOAc, 2:1). Method B (from the
Acid):[15] The same amides were obtained, albeit in lower yield,
from the corresponding amines by the above procedure. Method C
(via the Acyl Chloride): PCl5 (2.626 g, 12.6 mmol, 1.25 equiv.) was
introduced into a 50 mL Grignard flask, fitted with a distillation
column, a condenser, and a cold trap, and placed in an oil bath
maintained at 35–40 °C. Chlorodifluoroacetic acid (1.313 g,
10.1 mmol, 1 equiv.) was added dropwise by syringe and the tem-
perature of the bath was then raised to 60 °C in order to recover
the final traces of the acyl chloride 6, which appeared as a colorless,
clear liquid (390 mg, 26%); b.p. 26–28 °C. 19F NMR (CDCl3): δ
= –65.0 (s) ppm. Compound 6 was then treated with the appropri-
ate primary amine to give the desired amides 5, albeit in lower
yields.
Cyclization of 2 to Azetidinone 3: A lithium hexamethyldisilylamide
(LiHMDS) solution was prepared under anhydrous conditions in
a 250 mL Grignard flask. Butyllithium (n-BuLi, 1.64 mmol) was
added with stirring to hexamethyldisilazane (264 mg, 1.64 mmol),
diluted in anhydrous THF (10 mL). Separately, amino ester 2
(500 mg, 1.64 mmol) and dried THF (15 mL) were introduced into
a 100 mL Grignard flask. The LiHMDS solution was then slowly
added at –20 °C with stirring. Stirring was continued for 5 min,
and the reaction mixture was quenched with HCl (4 , 25 mL) and
extracted three times with dichloromethane (3×15 mL). The or-
ganic layer was separated and dried with magnesium sulfate. Or-
ganic solvents were evaporated and the residue was subjected to
column chromatography on silica gel (pentane/Et2O, 9:1). Pure 3
was obtained in 65% yield.
1
3,3-Difluoro-1,4-diphenylazetidin-2-one (3): M.p. 135 °C. H NMR
3
3
(CDCl3): δ = 5.42 (dd, JH,F = 2.1 Hz, JH,F = 7.6 Hz, 1 H, Hc),
7.10–7.50 (m, 10 H, Harom) ppm. 13C NMR (CDCl3): δ = 69.05
2
2
1
(dd, JC,F = 26.9 Hz, JC,F = 24.7 Hz, Cc), 119.48 (dd, JC,F
=
m
1
n
2
284.5 Hz, JC,F = 287.5 Hz, Cb), 118–137 (Carom), 157.87 (t, JC,F
= 31.6 Hz, Ca) ppm. 19F NMR (CDCl3): δ = –114.13 (dd, J
3
2-Chloro-2,2-difluoro-N-(p-methoxyphenyl)acetamide (5a): This
amide was obtained as a yellowish oil, which was purified on a
silica gel column and recovered as white crystals (1.41 g, 60%); m.p.
Hc,Fm
3
= 7.6 Hz, 2JF
= 226.2 Hz, Fm), –119.90 (dd, JH ,F = 2.1 Hz,
m,Fn
2JF ,F = 226.2 Hz, Fn) ppm.[25] IR (KBr): ν = 1765 c(C=O), 1494,
n
˜
1458,m1392, 1316, 1207, 1157, 1108, 1003, 825, 753 cm–1. MS (EI):
n
1
111 °C. TLC: Rf = 0.80. H NMR (CDCl3): δ = 3.81 (s, 3 H, Hg),
6.90 and 7.45 (AAЈBBЈ, 2×2 H, Hd,e), 7.77 (s, NH) ppm. 13C NMR
m/z (%) = 259 (5.6) [M+·], 140 (100) [Ph–CH=CF2+·], 119 (15.7),
104 (8.6), 91 (12.4), 78 (7.2) [O=C=CF2+·], 77 (32.5), 51 (17.3). The
base peak at m/z = 140 is indicative of the preferred fragmentation
pattern of the azetidinone ring, affording isocyanate and olefin
fragments over ketene and imine fragments. This phenomenon has
already been observed.[26] C15H11F2NO (259.25): calcd. C 69.49, H
4.28, N 5.40; found C 69.47, H 4.30, N 5.37.
1
(CDCl3): δ = 55.5 (Cg), 114.4 (Cd), 119.5 (t, JC,F = 300 Hz, Ca),
122.2 (Ce), 128.2 (Cc), 157.0 (t, 2JC,F = 30 Hz, Cb), 157.7 (Cf) ppm.
19F NMR (CDCl ): δ = –64.24 (s) ppm. IR (KBr): ν = 3414 (N–
˜
3
H), 1701 (C=O), 2836, 1512, 1248, 1038, 831, 620 cm–1. MS (EI):
m/z (%) = 237 (24), 235 (80), 200 (13), 149 (16), 122 (100), 95 (14),
52 (8). C9H8ClF2NO2 (235.61): calcd. C 45.88, H 3.42, N 5.94;
found C 45.85, H 3.41, N 5.89.
2-Chloro-2,2-difluoro-N-[(S)-1-phenylethyl]acetamide (5b): This
Attempted Addition of 1 to N-Benzylidenebenzylamine: N-Benzyl-
idenebenzylamine (3.81 g, 20 mmol) was treated with 1 under the
same conditions as had been used for the Schiff base. However,
GC and NMR analysis of the crude product showed that ethyl
compound was obtained from (–)-(S)-1-phenylethylamine ([α]D =
–30 1, c = 10, ethanol; –8 1, c = 10, CHCl3) as a yellowish oil,
which was purified on a silica gel column and recovered in 64%
yield as white crystals (m.p. 70 °C). TLC: Rf = 0.85. [α]D = +55
1
3
difluoroacetate and N-benzylidene-α-(trimethylsilyl)benzylamine
(c = 10, CHCl3). 1H NMR (CDCl3): δ = 1.60 (d, J = 2.1 Hz, 3 H,
2
(4) were formed. Ester: 19F NMR (CDCl3): δ = –127 (d, JH,F
=
Hd), 5.12 (m, 1 H, Hc), 6.42 (s, 1 H, NH), 7.30 (m, 5 H, Hf,g,h)
52.8 Hz) ppm. Imine 4: M.p. 142 °C (ref.[13a] 143–145.3 °C). 1H
NMR (CDCl3): δ = 4.16 (s, 1 H, SiCHPh) ppm. C17H21NSi
(267.44): calcd. C 76.35, H 7.91, N 5.24; found C 76.38, H 7.85, N
5.23.
ppm. 13C NMR (CDCl3): δ = 21.4 (Cd), 48.7 (Cc), 119.5 (t, JC,F
1
= 300 Hz, Ca), 126.5 (Cg), 128.5 (Ch), 129.3 (Cf), 141.4 (Ce), 158.8
2
(t, JC,F = 30.0 Hz, Cb) ppm. 19F NMR (CDCl3); δ = –64.51 (s)
ppm. IR (KBr): ν = 3414 (N–H), 1700 (C=O), 1536, 1138, 984,
˜
Eur. J. Org. Chem. 2006, 4147–4154
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4151

M. Bordeau, F. Frébault, M. Gobet, J.-P. Picard
FULL PAPER
700, 630 cm–1. MS (EI): m/z (%) = 235 (14), 233 (41), 218 (37), 198
3
1H NMR (CDCl3): δ = 0.21 (s, 9 H, Ha), 1. 54 (d, J = 1.7 Hz, 3
(19), 148 (22), 105 (100), 77 (36), 51 (16). C10H10ClF2NO (233.61): H, He), 5.15 (m, 1 H, Hd), 6.44 (s, 1 H, NH), 7.30 (m, 5 H, Hg,h,i
)
calcd. C 51.41, H 4.31, N 5.99; found C 51.44, H 4.38, N 5.91.
ppm. 13C NMR (CDCl3): δ = –4.7 (Ca), 20.7 (Ce), 48.6 (Cd), 122.5
1
(t, JC,F = 300 Hz, Cb), 126.2 (Ch), 127.7 (Ci), 128.8 (Cg), 142.1
2
(Cf), 165.1 (t, JC,F = 30.0 Hz, Cc) ppm. 19F NMR (CDCl3): δ =
–122.82 (s) ppm. IR (KBr): ν = 3414 (N–H), 1669 (C=O), 1539,
˜
1111, 852 cm–1. MS (EI): m/z (%) = 271 (9), 256 (3), 220 (2), 152
(10), 105 (100), 77 (15), 73 (11). C13H19F2NOSi (271.38): calcd. C
57.54, H 7.06, N 5.16; found C 57.50, H 6.97, N 5.21.
2-Chloro-2,2-difluoro-N-isopropylacetamide (5c): This compound
was obtained as a white, crystalline powder (62% yield). TLC: Rf
1
= 0.91. H NMR (CDCl3): δ = 1.24 (d, 3J = 8.1 Hz, 6 H, Hd), 4.10
(sept, 3J = 8.1 Hz, 1 H, Hc), 6.10 (s, NH) ppm. 13C NMR (CDCl3):
1
δ = 22.0 (Cd), 42.9 (Cc), 119.1 (t, JC,F = 300 Hz, Ca), 158.4 (t,
2JC,F = 30 Hz, Cb) ppm. 19F NMR (CDCl3): δ = –64.56 (s) ppm.
IR (KBr): ν = 3413 (N–H), 1700 (C=O), 1135, 990, 623 cm–1. MS
˜
(EI): m/z (%) = 158 (30), 156 (90), 108 (9), 87 (12), 86 (100), 85
(26), 70 (14), 43 (80). C5H8ClF2NO (171.57): calcd. C 35.00, H
4.70, N 8.16; found C 34.94, H 4.78, N 8.11.
2,2-Difluoro-N-isopropyl-2-(trimethylsilyl)acetamide (7c): This com-
pound was obtained as a white, crystalline solid in 41% yield. TLC:
1
3
Rf = 0.48. H NMR (CDCl3): δ = 0.23 (s, 9 H, Ha), 1.20 (d, J =
7.7 Hz, 6 H, He), 4.11 (m, 3J = 7.7 Hz, 1 H, Hd), 6.0 (s, NH) ppm.
13C NMR (CDCl3): δ = –4.8 (Ca), 22.5 (Ce), 42.4 (Cd), 122.5 (t,
1JC,F = 300 Hz, Cb), 165.4 (t, JC,F = 30 Hz, Cc = O) ppm. 19F
2
NMR (CDCl ): δ = –123.0 (s) ppm. IR (KBr): ν = 3414 (N–H),
˜
3
1662 (C=O), 1552, 1256, 1109, 1026, 852, 628 cm–1. MS (EI): m/z
(%) = 209 (4), 194 (40), 158 (41), 116 (14), 81 (20), 77 (38), 73
(100), 43 (56). C8H17F2NOSi (271.38): calcd. C 45.91, H 8.19, N
6.69; found C 45.86, H 8.18, N 6.75.
Synthesis of Silylamides 7. Electrosilylation of Chloroamides 5: Elec-
trodes (aluminium anode and stainless steel cathode) were washed
with a hydrochloric acid solution (10%) for 10 min, then rinsed
with pure water and then acetone, and were finally dried at 120 °C
for 30 min. Tetrabutylammonium bromide (0.25 g, 0.25 mmol,
0.05 equiv.) and a magnetic stirring bar were introduced into the
electrolysis cell fitted with the electrodes. HMPA (10 mL) was intro-
duced under dry nitrogen and the mixture was stirred until the salt
was dissolved. Finally, dried THF (40 mL) and TMSCl (distilled
from magnesium turnings, 3.2 mL, 25 mmol, 5 equiv.) were added.
Preelectrolysis (45 min) was carried out first, and the chloroamide
5 (1.178 g, 5 mmol, 1 equiv., diluted in 5 mL anhydrous THF) was
then added to the mixture at the rate of the electrolysis (i =
100 mA) by automatic syringe over 3 h 45 min (1350 C,
2.5 F·mol–1). The solvents were evaporated under vacuum and the
raw material was purified on a column of silica gel (petroleum
ether/EtOAc, 2:1) to provide the pure silyl amide 7.
Synthesis of β-Hydroxy Amides 8: Dried potassium fluoride
(2.9 mg, 0.05 mmol, 0.05 equiv.), isobutyraldehyde (144.2 mg,
2 mmol, 2 equiv.), and distilled DMF (5 mL) were introduced un-
der anhydrous nitrogen into a two-necked, round-bottomed flask.
(Trimethylsilyl)acetamide (7, 273.4 mg, 1 mmol, 1 equiv.), dissolved
in DMF (3 mL), was added slowly with stirring at –10 °C, the reac-
tion mixture was allowed to return to ambient temperature and
stirred for 2 h, and the mixture was poured into dilute HCl (2 mL,
1% solution in water) and extracted with diethyl ether (3×5 mL).
The organic phases were washed with a cold saturated aqueous
solution of sodium hydrogencarbonate (1 mL) and brine (10 mL)
and dried with magnesium sulfate, and the solvent was evaporated
under vacuum. The raw material was obtained as a colorless, vis-
cous liquid and products 8 were purified by column chromatog-
raphy on silica gel (petroleum ether/EtOAc, 2:1), yielding com-
pounds 8 as white, crystalline solids.
2,2-Difluoro-N-(p-methoxyphenyl)-2-(trimethylsilyl)acetamide (7a):
This compound was obtained as a white, crystalline solid (442 mg,
1
32% yield). TLC: Rf = 0.87. H NMR (CDCl3): δ = 0.29 (s, 9 H,
Ha), 3.80 (s, 3 H, Hh), 6.90 and 7.45 (AAЈBBЈ, 2×2 H, He,f), 7.80
(s, NH) ppm. 13C NMR (CDCl3): δ = –4.7 (Ca), 55.5 (Ch), 114.3
1
(Ce), 122.3 (Cf), 122.6 (t, JC,F = 300 Hz, Cb), 129.2 (Cd), 157.1
2
(Cg), 164.3 (t, JC,F = 30 Hz, Cc) ppm. 19F NMR (CDCl3): δ =
–122.10 (s) ppm. IR (KBr): ν = 3338 (N–H), 1675 (C=O), 1538,
˜
1257, 1031, 855 cm–1. MS (EI): m/z (%) = 273 (47), 258 (13), 222
(4), 180 (12), 122 (51), 77 (23), 73 (100). C12H17F2NO2Si (273.35):
calcd. C 52.73, H 6.27, N 5.12; found C 52.68, H 6.20, N 5.15.
2,2-Difluoro-3-hydroxy-N-(p-methoxyphenyl)-4-methylpentanamide
(8a): This compound was obtained in 33% yield. TLC (petroleum
ether/EtOAc, 2:1): Rf = 0.33. 1H NMR (CDCl3): δ = 1.05 and
1.07 (d, 3J = 7.2 Hz, 2×3 H, Ha and Hb), 2.1 (m, 3J = 7.2 and
3
7.0 Hz, 1 H, Hc), 2.6 (d, J = 6.4 Hz, 1 H, OH), 3.8 (s, 3 H, Hk),
4.00 (sept, 3J = 6.4 and 7.0 Hz, 1 H, Hd), 6.9 and 7.4 (AAЈBBЈ,
2×2 H, Hh,i), 8 (s, NH) ppm. 13C NMR (CDCl3): δ = 17.3 (Ca),
2
19.6 (Cb), 28.8 (Cc), 55.5 (Ck), 75.4 (t, JC,F = 24.9 Hz, Cd), 121.1
1
1
2,2-Difluoro-N-[(S)-1-phenylethyl]-2-(trimethylsilyl)acetamide (7b):
This compound was obtained in 44% yield as a white, crystalline
solid (m.p. 75 °C). TLC: Rf = 0.45. [α]D = –21 1 (c = 10, HCCl3).
(Ci), 122.07 (Ch), 127.5 (Cg), 128.8 (dd, J
= 249 Hz, JC,F =
C,Fm
n
256 Hz, Ce), 157.9 (Cj), 165.1 (t, 2JC,F = 32 Hz, Cf) ppm. 19F NMR
(CDCl3): δ = –110.11 (dd, 3JF = 18.8, 2JF
= 261.8 Hz),
m,H
m,Fn
4152
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4147–4154

Difluoro(trimethylsilyl)acetyl Precursors of 3,3-Difluoroazetidinones
FULL PAPER
an additional 24 h. Evaporation of the solvent under vacuum left
a viscous, orange liquid, which was purified by column chromatog-
raphy on silica gel (petroleum ether/EtOAc, 4:1) to provide 9a as a
white, crystalline solid.
3
2
–120.72 (dd, JF ,H = 7.5, JF ,F = 261.8 Hz) ppm. C13H17F2NO3
(273.28): calcd. C 57.14, H 6.27,mN 5.13; found C 57.19, H 6.30, N
n
n
5.11.
3,3-Difluoro-4-isopropyl-N-(p-methoxyphenyl)azetin-2-one
(9a):
This compound was obtained in 63% yield. TLC (petroleum ether/
1
3
EtOAc): Rf = 0.55. H NMR (CDCl3): δ = 1.02 (d, J = 7.0 Hz, 3
3
H, He), 1.09 (d, J = 7.0 Hz, 3 H, Hf), 2.36 (m, 1 H, Hd), 3.82 (s,
3 H, Hk), 4.30 (m, 1 H, Hc), 6.92 and 7.50 (AAЈBBЈ d, 3J = 9.0 Hz,
2 H, Hi and d, J = 9.0 Hz, 2 H, Hh, respectively) ppm. 13C NMR
3
2,2-Difluoro-3-hydroxy-4-methyl-N-[(S)-1-phenylethyl]pentanamide
(8b): This compound was recovered in 31% yield. TLC (petroleum
ether/EtOAc, 7:1): Rf = 0.17. 13C NMR (CDCl3): δ = 17.0 (Ca and
(CDCl3): δ = 16.9 (Ce), 18.4 (Cf), 27.4 (Cd), 55.5 (Ck), 70.5 (t, 2JC,F
1
= 23 Hz, Cc), 114.6 (Ci), 117.0 (t, JC,F = 300 Hz, Cb), 120.4 (Ch),
120.8 (Cg), 129.0 (Cj), 157.6 (t, JC,F = 30 Hz, Ca) ppm. 19F NMR
2
Cb), 20 (Ch), 28.5 (Cc), 49 (Cg), 75 (Cd), 112.3 (dd, 1J
= 255 Hz,
C,Fm
(CDCl3): δ = –112.18 (dd, 3JF
= 8.8 Hz, 2JF
= 234.4 Hz,
m,H
m,Fn
1JC,F = 256.5 Hz, Ce), 126.1 (Cl), 127.8 (Cj), 128.6 (Ck), 141.5 (Ci),
163 (nt, 2JC,F = 31 Hz, Cf) ppm. Careful examination of the 1H and
19F NMR spectra of the crude material revealed the presence of
two diastereomers in a 1:1 ratio. 1H NMR (CDCl3): one dia-
3
2
Fm), –120.72 (dd, JF ,H = 1.9 Hz, JF ,F = 234,4 Hz, Fn) ppm.
n
n
m
Coupling constants were obtained through calculations carried out
with a simulation program.[27] Owing to the asymmetric nature of
Cc, the two fluorine atoms are magnetically inequivalent and have
3
3
stereomer: δ = 1.0 (d, J = 8.1 Hz, 3 H, Ha), 1.015 (d, J = 8.6 Hz,
3
different chemical shifts and coupling constants JH,F. The larger
3
3 H, Hb), 1.50 (d, J = 6.8 Hz, 3 H, Hh), 1.91 (m, 1 H, Hc), 2.60
3JH,F value (8.8 Hz) was attributed to Fm cis to Hc and the lower
(1.9 Hz) to Fn trans to Hc.[25] Apparently, protons of one methyl
3
3
3
(s, OH), 3.85 (ddd, JH,F = 18.1 Hz, JH,F = 7.2 Hz, JH,H
=
m
n
3
5.3 Hz, 1 H, Hd), 5.07 (quint, J = 7.1 Hz, 1 H, Hg), 6.60 (s, NH),
7.30–7.40 (m, 5 H, Hj,k,l) ppm; other diastereomer: δ = 0.99 (d, J
group (Mee, for instance) are coupled with Hc (4J
= 0.5 Hz)
Hc,He
3
whilst those of Mef are not. This would result from the dia-
3
3
= 8.5 Hz, 3 H, Ha), 1.02 (d, J = 8.4 Hz, 3 H, Hb), 1.56 (d, J =
stereotopy of these methyl groups. A COSY H–H experiment
3
7.3 Hz, 3 H, Hh), 2.01 (m, 1 H, Hc), 2.65 (s, OH), 3.88 (td, J
H,Fm
4
showed Mee and Mef to be coupled and J
relating to Mef to
Hc,He
= 22.7 Hz, 3JH,F = 7.35 Hz, 3JH,H = 7.35 Hz, 1 H, Hd), 5.15 (quint,
n
be less than 0.5 Hz, a value too small to show up in a 400 MHz
spectrum. Similarly, Fn,Hc coupling is not visible, although simula-
tion indicates a 1.9 Hz value. MS (EI): m/z (%) = 255 (42), 240 (7),
149 (100), 134 (51). The presence of the ion m/z = 149 as the base
peak is indicative of a pattern of a preferred fragmentation of the
azetidinone ring to afford isocyanate and olefin fragments over a
fragmentation giving ketene and imine fragments. This phenome-
non had already been observed previously.[26] C13H15F2NO2
(255.26): calcd. C 61.17, H 5.92, N 5.49; found C61.19, H 5.90, N
5.53.
3J = 8.4 Hz, 1 H, Hg), 6.80 (s, NH), 7.35–7.45 (m, 5 H, Harom
)
ppm. 19F NMR (CDCl3): one diastereomer: δ = –110.7 (dd, 3J
Fm,H
= 18.8 Hz, 2JF
= 263.7 Hz), –120.8 (dd, 3JF ,H = 7.5 Hz, JF
m,Fn
n
2
m,Fn
= 263.7 Hz) ppm; other diastereomer: δ = –110.1 (dd, 3JF
=
=
m,H
18.8 Hz, 2JF
= 263.7 Hz), –121.9 (dd, 3JF ,H = 7.5 Hz, 2JF
m,Fn
m,Fn
267.1 Hz) ppm. C14H19F2NO2 (271.30): calcnd. C 61.98, H 7.06, N
5.16; found C 62.08, H 7.01, N 5.18.
2,2-Difluoro-3-hydroxy-N-isopropyl-4-methylpentanamide (8c): This
1
compound was obtained in 28% yield. TLC: Rf = 0.86. H NMR
3
3
(CDCl3): δ = 1.02 (d, J = 7.5 Hz, 3 H, Ha), 1.04 (d, J = 7.1 Hz,
3
3 H, Hb), 1.22 (d, J = 6.4 Hz, 6 H, Hh), 2.00 (m, 1 H, Hc), 2.72
3
(d, J = 6.4 Hz, 1 H, OH), 3.91 (m, 1 H, Hd), 4.11 (m, 1 H, Hg),
6.25 (s, 1 H, NH) ppm. 13C NMR (CDCl3): δ = 17.1 (Ca), 19.3
2
(Cb), 22.4 (Ch), 28.6 (Cc), 42.0 (Cg), 75.5 (t, JC,F = 25 Hz, Cd),
1
1
2
116.0 (dd, JC,F = 255 Hz, JC,F = 257 Hz, Ce), 165.0 (t, JC,F
=
=
=
m
n
31 Hz, Cf) ppm. 19F NMR (CDCl3): δ = –110.7 (dd, 3JF
Acknowledgments
m,H
18.7 Hz, 2JF
= 267.5 Hz), –120.9 (dd, 3JF ,H = 7.5 Hz, 2JF
m,Fn
n
m,Fn
267.5 Hz) ppm. C9H17F2NO2 (271.30): calcd. C 51.66, H 8.19, N
6.69; found C 51.61, H 8.25, N 6.65.
The authors wish to thank Aquitaine Regional Government for
funding part of this work through an Aquitaine-Euskadi Coopera-
tion Program, P. B. Gonzalez Perez (Aquitaine-Euskadi fellow) for
some preliminary experiments, and D. Nuñez Villanueva (Erasmus
fellow from the University of Malaga, Spain) for complementary
work. They are also indebted to M. Petraud and J. C. Lartigue
(CESAMO, University Bordeaux 1) for their very kind and ef-
ficient assistance in NMR studies. They also acknowledge the as-
sistance of M.-H. Lescure for the recording of GC/MS data.
Cyclization of 8a Into Azetidinone 9a: Triphenylphosphane
(104.3 mg, 0.4 mmol, 2 equiv.) and 8a (54.8 mg, 0.2 mmol,
1 equiv.), dissolved in dry THF (3 mL), were introduced under dry
nitrogen into a 10 mL round-bottomed flask . The resulting solu-
tion was stirred at 20 °C for 15 min and DEAD (61 µL, 0.4 mmol,
2 equiv.) was added dropwise at 0 °C. The mixture was stirred for
[1] S. Thaisrivongs, H. J. Schostarez, D. T. Pals, S. R. Turner, J.
Med. Chem. 1987, 30, 1837.
[2] J.-L. Maillard, C. Favreau, M. Reboud-Ravaux, R. Kobaiter,
R. Joyeau, M. Wakselman, Eur. J. Cell Biol. 1990, 52, 213.
Eur. J. Org. Chem. 2006, 4147–4154
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4153

M. Bordeau, F. Frébault, M. Gobet, J.-P. Picard
FULL PAPER
[3] R. Joyeau, H. Molines, R. Labia, M. Wakselman, J. Med.
Chem. 1988, 31, 370.
[19] E. J. Thomas, C. T. Brain, A. Chen, A. Nelson, N. Tanikkul,
Tetrahedron Lett. 2001, 42, 1247.
[4] See also the dehydrohalogenation route from β-bromo-3,3-di-
[20] C. A. Townsen, L. T. Nguyen, J. Am. Chem. Soc. 1981, 103,
fluoroamides, ref.[3]
4582.
[5] T. Taguchi, O. Kitagawa, Y. Suda, S. Ohkawa, A. Hashimoto,
Y. Iitaka, Y. Kobayashi, Tetrahedron Lett. 1988, 29, 5291; O.
Kitagawa, T. Taguchi, Y. Kobayashi, Tetrahedron Lett. 1988,
29, 1803.
[6] S. Marcotte, X. Pannecoucke, C. Feasson, J.-C. Quirion, J. Org.
Chem. 1999, 64, 8461.
[21] C. A. Townsen, L. T. Nguyen, Tetrahedron Lett. 1982, 23, 4859.
[22] C. T. Brain, A. Chen, A. Nelson, N. Tanikkul, E. J. Thomas,
Tetrahedron Lett. 2001, 42, 1247.
[23] A. Otaka, J. Watanaba, A. Yukimasa, Y. Sasaki, H. Watanabe,
T. Kinoshita, S. Oishi, H. Tamamura, N. Fujii, J. Org. Chem.
2004, 69, 1634.
[7] K. Sato, A. Tarui, S. Matsuda, M. Omote, A. Ando, I. Kuma-
daki, Tetrahedron Lett. 2005, 46, 7679.
[8] A. Otaka, H. Watanabe, E. Mitsuyama, A. Yukimasa, H. Tam-
amura, N. Fujii, Tetrahedron Lett. 2001, 42, 285.
[9] S. Lacroix, A. Cheguillaume, S. Gérard, J. Marchand-Brynaert,
Synthesis 2003, 2483.
[24] See for instance: D. L. Hughes, R. A. Reamer, J. J. Bergan,
E. J. J. Grabowski, J. Am. Chem. Soc. 1988, 110, 6487; J. P.
Coleman, F. X. Felpin, W. Chen, J. Org. Chem. 2004, 69, 7309.
[25] It has been reported that JH,H (cis) is larger than JH,H (trans)
in small rings such as cyclobutane, oxetane, and azetidine de-
rivatives. See, for example: R. M. Silverstein, G. C. Bassler,
T. C. Morill, Spectrometric Identification of Organic Com-
pounds, 5th ed., J. Willey, New York, 1991, p. 221; D. L. Pavia,
G. M. Lampman, G. S. Kriz Jr, Introduction to Spectroscopy,
Saunders College Publishing, Orlando, Florida, USA, 1979, p.
117. The same phenomenon was also observed for JH,F in mon-
ofluorinated oxirane; see, for example: E. Pretsch, P.
Bühlmann, C. Affolter, Structure Determination of Organic
Compounds, 3rd ed., Springer, Berlin, 2000, p. 196.
[26] B. E. Autier, M. Fétizon, H. B. Kagan, J. L. Luche, Bull. Soc.
Chim. Fr. 1967, 2297; V. S. Georgiev, D. C. Coonber, G. B.
Mullen, Org. Mass Spectrom. 1988, 23, 283; G. Bourgeois, J. P.
Picard, F. P. Cossio, C. Palomo, Adv. Mass Spectrom. 1989, 11,
876.
[10] P. Clavel, C. Biran, M. Bordeau, N. Roques, S. Trevin, Tetrahe-
dron Lett. 2000, 41, 8763.
[11] For alternative syntheses of 1 starting from trifluoroacetates
see: a) K. Uneyama, G. Mizutani, K. Maeda, T. Kato, J. Org.
Chem. 1999, 64, 6717; b) H. Amii, T. Kobayashi, K. Uneyama,
Synthesis 2000, 2001.
[12] For a previously reported example, see ref.[11a]
[13] a) E. Popowsky, Z. Chem. 1974, 14, 289; b) E. Popowsky, Z.
Chem. 1975, 15, 275; c) M. Komatsu, M. Ohno, S. Tsuno, Y.
Ohshiro, Chem. Lett. 1990, 575.
[14] Q. Z. Zhou, Z. C. Chen, Synth. Commun. 2000, 30, 3189.
[15] L. E. Barstow, V. J. Hruby, J. Org. Chem. 1971, 36, 1305.
[16] R. S. Corley, S. G. Cohen, M. S. Simon, H. T. Wolosinski, J.
Am. Chem. Soc. 1956, 78, 2608.
[17] J. H. Saunders, R. I. Slocombe, I. I. Hardy, J. Am. Chem. Soc.
1949, 71, 752.
[27] Simulation performed with the MestreC 23 program. Fluorine,
having a spin of 1/2, like the proton, was integrated into the
calculations without difficulty.
[18] H. Oberhammer, K. I. Gobbattto, C. Leibold, S. Centeno,
C. O. Delle Vedona, H. G. Mack, J. Mol. Struct. 1996, 380, 55.
Received: February 20, 2006
Published Online: July 12, 2006
4154
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4147–4154