The reactivity of the hydrofluorocarbon 1,1,1,2-tetrafluoroethane (HFC-134a) and related compounds towards base attack. The generation and stability of the tetrafluoroethyl, trifluorovinyl and related anions

Article abstract of DOI:10.1016/S0022-1139(99)00182-7

Trifluorovinyllithium (CF₂ = CFLi) has been successfully synthesised from 1,1,1,2-tetrafluoroethane using n-butyllithium at -78°C in ether. Other bases (LDA, DBU, MeLi, t-BuLi, t-BuOK) were tried with varied success. The stability of the system was dependent on temperature, concentration and solvent system employed. 1-Chloro-2,2,2-trifluoroethane (133a) led to an analogous polyfluorovinyllithium, but 1,1,1-trifluoro- and 1,1,1,2,2,2-hexafluoro- ethane did not.

Full text of DOI:10.1016/S0022-1139(99)00182-7

Journal of Fluorine Chemistry 99 (1999) 127±131
The reactivity of the hydro¯uorocarbon 1,1,1,2-tetra¯uoroethane
(
HFC-134a) and related compounds towards base attack.
The generation and stability of the tetra¯uoroethyl,
tri¯uorovinyl and related anions
a,*
J. Burdon , P.L. Coe , I.B. Haslock , R.L. Powell
School of Chemistry, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
b
ICI Klea Business, PO Box 8, The Heath, Runcorn, Cheshire WA7 4QD, UK
a
a,1
b,2
a
Received 7 June 1999; accepted 7 June 1999
Abstract
Tri¯uorovinyllithium (CF = CFLi) has been successfully synthesised from 1,1,1,2-tetra¯uoroethane using n-butyllithium at � 788C in
2
ether. Other bases (LDA, DBU, MeLi, t-BuLi, t-BuOK) were tried with varied success. The stability of the system was dependent on
temperature, concentration and solvent system employed. 1-Chloro-2,2,2-tri¯uoroethane (133a) led to an analogous poly¯uorovinyllithium,
but 1,1,1-tri¯uoro- and 1,1,1,2,2,2-hexa¯uoro- ethane did not. # 1999 Elsevier Science S.A. All rights reserved.
Keywords: 1,1,1,2-Tetra¯uoroethane; 134a; 1,2,2,2-Tetra¯uoroethyllithium; Tri¯uorovinyllithium; 1-Chloro-2,2-di¯uorovinyllithium; 133a; 1,1,1-Tri¯uor-
oethane; 1,1,1,2,2,2-Hexa¯uoroethane; Di¯uoromethane
1
. Introduction
Tarrant et al. [6] (starting from bromotri¯uoroethene) and
Seyferth et al. [7] (starting from phenyltris(tri¯uorovinyl)-
tin) were the ®rst to prepare tri¯uorovinyllithium and were
followed by Normant et al. [8] who found that it could be
formed from chlorotri¯uoroethene and n-butyllithium, pro-
vided that THF was used as a solvent: they also con®rmed
previous work, showing that n-butyllithium at � 708C in
ether attacked the double bond and displaced ¯uorine to
form (Z)-1-chloro-1,2-di¯uorohex-1-ene. Normant [1] also
studied the reactivity of various ¯uorinated vinylic species
with alkyllithium reagents in a variety of solvents. However,
Normant and co-workers [9] later discovered that bulky
bases, such as sec-butyllithium or t-butyllithium, did not
lead to nucleophilic attack and could be used in ether as
solvent (poly¯uorovinyllithiums are much less stable in
THF than in ether): for example, only tri¯uorovinyllithium
was formed from chlorotri¯uoroethene and t-butyllithium in
ether.
Fluorinated vinylic organometallics have been generated
[
1,2] from precursors such as bromotri¯uoroethene, chlor-
otri¯uoroethene, 1,1-dichloro-2,2-di¯uoroethene, tri¯uor-
oethene, 1,1-dichloro-2-¯uoro-2-phenylethene and
,2,3,3,3-penta¯uoroethene. The procedure was halogen
or hydrogen)±metal exchange between the poly¯uorovinyl
1
(
halide (hydride) and either an organometallic or the metal
itself.
The ®rst workers in the ®eld were Dixon [3] and Meier
and Bohler [4] who studied the reactivity of chlorotri¯uor-
oethene, and reported that its reaction with organolithium
reagents favoured nucleophilic attack at the di¯uoro end of
the double bond ahead of the chlorine±lithium exchange.
Callander et al. [5], who studied the reactions of penta¯uor-
ophenyllithium with a wide range of polyhaloethenes,
observed nucleophilic attack, elimination (of HHal), halo-
gen±lithium exchange and hydrogen±lithium exchange,
depending on the polyhaloethene used.
2
. Results and discussion
*
Corresponding author.
1
The present paper describes our work in detail [10] on the
preparation of tri¯uorovinyllithium from the now readily
available 1,1,1,2-tetra¯uoroethane (Freon 134a, 1a): in addi-
Present address: Robertson Laboratories, Llanrhos, North Wales, UK.
2
Present address: Department of Chemistry, UMIST, Manchester, M 60
IQD, UK.
0022-1139/99/$ ± see front matter # 1999 Elsevier Science S.A. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 9 9 ) 0 0 1 8 2 - 7

1
28
J. Burdon et al. / Journal of Fluorine Chemistry 99 (1999) 127±131
tion it describes the preparation, or attempted preparation,
of organolithiums from 1-chloro-2,2,2-tri¯uoroethane (i.e.,
needed to be operated, potassium t-butoxide was not suf®-
ciently reactive to achieve deprotonation.
1
(
(
-chloro-2,2-di¯uorovinyllithium),
2,2-di¯uorovinyllithium), 1,1,1,3,3,3-hexa¯uoropropane
2,2-di¯uoro-1-tri¯uoromethyllithium) and di¯uoro-
1,1,1-tri¯uoroethane
When DBU was added to a mixture of 1a and benzalde-
hyde, no ¯uorine-containing products were recovered, only
unreacted benzaldehyde. When trimethylsilyl chloride and
tributyltin chloride were used as electrophiles in the same
way, again no reaction was observed. However, when
tributyltin chloride was added to a mixture of DBU and
methane.
2
.1. Preparation and reactions of trifluorovinyllithium
1
a that had reacted for 2 h at � 788C, a product was observed
Tetra¯uoroethane 1a has two protons and, therefore, it
in very low yield (probably < 5%) in admixture with tribu-
tyltin chloride. The latter compound was all that could be
detected by mass spectroscopy, but the former was clearly
should be possible to obtain two products depending on
whether one or both protons are removed. If one proton were
lost then a tetra¯uoroethyl anion 1b would result and if that
species lost a ¯uoride ion to form tri¯uoroethene 1c, then
this too could be susceptible to base attack so generating the
tri¯uorovinyl anion 1d (Scheme 1). It is known [11] that
tri¯uoroethene can react with n-butyllithium to give tri¯uor-
ovinyllithium.
19
visible in the F NMR spectrum [CF , � 71 (internal CFCl )
3
3
dd (J = 18.3 and 12.2): CFH, � 236 dq (J = 45.8 and 18.3)]
1
and was just detectable in the H [ꢀ = 5.5, dq (J ꢀ 45 and
ꢀ10)], where it was almost lost because of the intensity of
the signals from tributyltin chloride. The NMR evidence
hence strongly suggests that the minor component was
CF CFHSnBu 2, and so DBU is capable of deprotonating
Since the acidity of 1a is not known, a number of bases
were employed to investigate its deprotonation. These were:
methyllithium (MeLi), potassium t-butoxide (t-BuOK), 1,8-
diazabicyclo-[5.4.0]-undec-7-ene (DBU), lithium diisopro-
pylamide (LDA), t-butyllithium (t-BuLi), and n-butyl-
lithium (n-BuLi). Due to the low boiling point of 1a
3
3
1a to the 1,2,2,2-tetra¯uoroethyl anion 1b to a very small
extent or very slowly.
When two molar equivalents of LDA were added to a
mixture of 1a and an electrophile (benzaldehyde, trimethyl-
silyl chloride), the products were only the expected ones
(see Table 1): not surprisingly, there were no products
resulting from the nucleophilic attack of the bulky LDA
on the electrophiles. One molar equivalent of LDA just gave
approximately 50% of the same products and 50% recov-
ered electrophile. When one molar equivalent of LDA and
(
� 26.58C), all the reactions had to be performed at low
temperatures, with � 788C being the most convenient. The
reactions were assayed by treating them with benzaldehyde
when three products could be formed (Scheme 1). The Table
summarises the results.
Two systems containing methyllithium were tried, ®rst
with methyllithium alone in ether and secondly with it as a
complex with lithium bromide, also in ether. In both cases a
molar amount of benzaldehyde was added after the methyl-
lithium and 1a had reacted for some time at � 788C and in
both cases the only product isolated was 1-phenylethanol
with no ¯uorine-containing products detected. It follows
that the deprotonation of 1a with methyllithium was slow.
As with methyllithium, two systems containing potas-
sium t-butoxide were tried. First potassium t-butoxide in
THF was added to a solution of 1a in THF with benzalde-
hyde present at � 788C, but no reaction was observed.
Secondly, when the same reaction was carried out in a
mixture of THF and DMF at � 788C the dominant process
was a Cannizzaro reaction of benzaldehyde as benzyl
alcohol was a product: only trace amounts of ¯uorine-
two equivalents of Bu SnCl were used, tetra¯uoroethyl
3
Table 1
Reaction of benzaldehyde with 1,1,1,2-trifluoroethane (1a) and base
a
base
PhCHO‡CF3CFH2ꢁ1a† ! PhCHꢁOH†CF ˆˆ CF2ꢁ3†‡PhCHꢁOH†Rꢁ4†
Base
Relative % 3
Relative% 4
Relative% PhCHO
b
n-BuLi
19
100
49
81
0
0
0
c,d
2 Â n-BuLi
b
LDA
0
51
0
b,c
2
2
2
 LDA
100
40
0
b,c
c,d
 t-BuLi
60
0
0
 t-BuLi
100
Trace
0
0
b,e
f
t-BuOK
d
MeLi
0
100
0
0
b
g
DBU
0
b,h
n-BuLi
100
0
0
19
a
containing materials were detected by F NMR. So at
the low temperatures at which the reactions involving 1a
[R = Me, n-Bu, t-Bu].
Base (10 mmol) added to a mixture of excess of 134a 1a and
b
benzaldehyde (10 mmol) in ether at � 788C. The solution was allowed to
warm to room temperature, then quenched with methanol. Product
mixtures analysed by NMR.
c
2
0 mmol of base to 10 mmol PhCHO.
d
Base (20 mmol) added to an excess of 134a 1a in ether at � 788C
followed by benzaldehyde (10 mmol) ca 1 h later: rest as in footnote b.
e
THF/DMF as solvent and not ether.
f
PhCH
2
OH detected.
Only product detected.
Reaction of CF CF I; n-BuLi added to a mixture of C
CF
g
h
3
2
2 5
F I and PhCHO
Scheme 1.
[12]; product was PhCH(OH)CF
2
3
.

J. Burdon et al. / Journal of Fluorine Chemistry 99 (1999) 127±131
129
compound 2 was detected as a product in about the same low
yield as in the DBU experiment.
t-Butyllithium is also a bulky base and thus would be
expected to be a poor nucleophile like LDA. However, when
two equivalent of t-BuLi were added to a mixture of 1a and
benzaldehyde, a mixture of two products was observed, one
resulting from the attack of tri¯uorovinyllithium (3) and the
other from the attack of t-BuLi (4). But when the aldehyde
added after the t-BuLi and 1a had been allowed to react,
only the tri¯uorovinyl product 3 was formed (Table 1).
Hence t-BuLi must deprotonate at about the same rate as it
reacts with benzaldehyde.
When n-BuLi was added to 1a followed after an hour by
the addition of benzaldehyde, the best yield for obtaining the
adduct 3 was again when two molar amounts of n-BuLi
compared to the aldehyde were used. Again, as with t-BuLi,
when n-BuLi was added to a mixture of the aldehyde and 1a,
the base attacked the electrophile as well as deprotonating
Scheme 2.
showed the same stability. To these ends we have repeated
several reactions that have already been reported, and in all
cases tri¯uorovinyllithium generated by our method
behaved essentially identically with the tri¯uorovinyl-
lithium generated by other methods. We have reacted it
Ð in ether at � 788C Ð with benzaldehyde [6], crotonal-
dehyde [8], carbon dioxide [7], trimethylsilyl chloride [7],
triethylsilyl chloride [14], and zinc chloride and iodide
followed by palladium catalysed couplings with aryl iodides
[15]. New reactions with epoxides [16] and with 2-tri¯uoro-
methylaniline [17] have already been reported by us.
We have also reacted it with camphor and 4-t-butylcy-
clohexanone and since these reactions have not been
reported before, we describe them here. Camphor gave
an alcohol (5) which would be expected to be the one from
endo-attack (Scheme 2) : camphor is invariably [18,19]
attacked endo-by nucleophiles for steric reasons. 4-t-butyl-
cyclohexanone gave (Scheme 2) a mixture of two alcohols
in the ratio 64 : 36 (6a and 6b). Literature results [18,20]
would suggest that the major one (6a) arises from equatorial
attack, and if this is so, then it also suggests that tri¯uoro-
vinyllithium is not very large Ð somewhere between
methyl- and ethyl-Grignard reagents.
1
a: the extent of this attack (81%) was not unexpectedly
greater in the case of n-BuLi than with the bulkier t-BuLi
60% even with two equivalents of t-BuLi). No products
(
containing the CF CFH-group were detected in the benzal-
3
dehyde reaction, or in reactions with any other electrophile.
Gassman [12,13] studied the reactivity of penta¯uor-
oethyllithium with a wide variety of electrophiles, for
example aldehydes, ketones and esters. Quite remarkably
he noted that the best yields were obtained when the
electrophile was stirring in situ with the penta¯uoroethyl
iodide while the methyllithium was added: no evidence of
methyllithium attack on the electrophiles was observed (we
have con®rmed this observation). This suggested that the
halogen±metal exchange between methyllithium and penta-
¯
uoroethyl iodide was much faster than the nucleophilic
attack of the potentially highly nucleophilic methyllithium
on the electrophiles. This is quite clearly different from our
experience when the addition of the base to a mixture of the
electrophile and 1a led to the base attacking the electrophile
to some extent. It follows that the abstraction of a proton
from 1a is slower than iodine±lithium exchange in penta-
¯
uoroiodoethane.
Our results show that the best method for generating and
The stability of tri¯uorovinyllithium prepared by our
method is much the same as that of the material prepared
in other ways [6±9]. Decomposition in THF is much more
rapid than in ether; the higher the concentration, the less
stable the compound is; and, in ether, the upper temperature
limit for practicable use is about � 308C.
using tri¯uorovinyllithium is to allow an excess of 1a to
react with two equivalents of a base (two with respect to the
electrophile) and then to add one equivalent of the electro-
phile later. LDA, t-BuLi and n-BuLi are all suitable, but
since the ®rst is prepared by using the last in a separate step,
there seems no point in using it. We have chosen to use n-
BuLi in all our subsequent experiments.
We also conclude that tetra¯uoroethyllithium 1b is not
available in experimentally useful amounts from 1a and any
base: it was only trapped using the two bases LDA and DBU
and only when the electrophile tributyltin chloride was the
trapping agent, and then only in extremely low yield.
Because our method of preparation of tri¯uorovinyl-
lithium differed from those already reported, we deemed
it necessary to check whether it reacted in the same way and
2.2. Preparation and reactions of 1-chloro-2,2-
difluorovinyllithium
In our original note [10], we outlined the preparation of
this vinylic compound from 1-chloro-2,2,2-tri¯uoroethane
(133a) and n-butyllithium (Scheme 3): the reaction is
obviously analogous to the formation of tri¯uorovinyl-
lithium. We also mentioned a few of its reactions, and have
since published [17±19] those with 2-aminobenzotri¯uoride

1
30
J. Burdon et al. / Journal of Fluorine Chemistry 99 (1999) 127±131
a syringe in a syringe pump and the reaction was left to stir
for another hour with the temperature staying between
788C and � 708C. This mixture was now ready to use
in other experiments which all had the electrophilic reagents
10 mmol for the quantities given here) added directly to
�
(
this system, followed by slow warming to room temperature
and then quenching.
Reactions under other conditions and with other bases
were performed very similarly (see Table 1), except that the
t-BuOK reactions were carried out in THF or in a mixture of
it and DMF.
Scheme 3.
and with epoxides: colleagues at Birmingham have also
published [21] a more extensive set of its reactions.
3.3. Reaction of camphor with trifluorovinyllithium
2
.3. Attempted preparation of other
polyfluorovinyllithiums
Camphor (1.44 g, 9.47 mmol) was added to a solution of
tri¯uorovinyllithium (10 mmol) at � 788C and the reaction
left to stir for 2 h with the system slowly warming to room
temperature. The reaction was quenched with saturated
Reactions of 1,1,1-tri¯uoroethane, 1,1,1,3,3,3-hexa¯uoro-
propane and di¯uoromethane with n-butyllithium in ether at
NH Cl solution and the crude product isolated by ether
4
�
788C, followed by quenching with benzaldehyde were all
extraction. It was puri®ed by column chromatography
(silica gel; hexane/ethyl acetate solvent gradient) to give
1,7,7-trimethyl-2-endo-tri¯uorovinylbicyclo[2.2.1]heptan-
unsuccessful (Scheme 3). In the ®rst and last cases, 1-
phenylpentan-1-ol, formed from reaction between n-butyl-
lithium and benzaldehyde, was the only product detected.
Hence we conclude that the protons on the two ¯uorohy-
drocarbons are not acidic enough to react with n-butyl-
lithium. In the second case, however, the isolated product
was benzaldehyde, and so the n-butyllithium must have
been consumed by the ¯uorohydrocarbon, conceivably as
shown in Scheme 3: if this is so, then the activating effect of
19
2-ol (5, 1.3 g, 59%): F NMR: ꢀ = � 102 (1F, dd, J = 82.4,
33.6), � 111 (1F, dd, J = 82.4, 109.8), � 166 (1F, dd, J = 33.6,
13
109.8); CNMR:ꢀ = (allsunlessstatedotherwise)11,21,21,
27, 31, 42, 45, 49, 54, 80 (ddd, J = 20), 134 (ddd, J = 250, 40,
1
40),155(dt,J = 298,45); HNMR:ꢀ = 0.9(3H,s),1.0(2H,s),
1.1 (2H, m), 1.2 (3H, s), 1.5 (1H, m), 1.7 (2H, m), 2.0 (2H, m),
+
4.4 (1H, s, exchangeable with D O): m/z (EI) 234 (M ), 219
2
+
the two CF groups must be suf®cient to make the CH2
3
(M � Me),201(219� H O),191,173,151,124,123,110,109,
2
hydogens acidic enough for them to be abstracted by n-
butyllithium Ð but since 1,1,1-tri¯uoroethane did not react,
one tri¯uoromethyl is not.
108, 95, 83, 81, 69, 67, 57, 55, 53, 43, 41.
3.4. Reaction of trifluorovinyllithium with 4-t-
butylcyclohexanone
3
. Experimental
4-t-Butylcyclohexanone (1.48 g, 9.61 mmol) was added
to tri¯uorovinyllithium (10 mmol) at � 788C and the mix-
ture stirred for 3 h as it warmed to room temperature. The
3
.1. NMR and mass spectrometry
3
reaction was quenched with methanol (20 cm ) and all
1
13
The H and C NMR spectra were measured on a Bruker
C300 spectrometer at 300 and 75 MHz, respectively, and
solvents removed by evaporation. Ether was added to the
residue, insoluble were ®ltered off and evaporation of the
ether left a mixture (1.7 g, 75%) of the two isomers (64 : 36
8
1
9
the F on a Jeol FX90Q spectrometer at 84.6 MHz with
CFCl as internal standard. CDCl was used as solvent.
Mass spectra were measured on a Kratos Pro®le instru-
ment.
19
ratio by F NMR) of 4-t-butyl-1-tri¯uorovinylcyclohexan-
1-ol (6a, 6b). The major isomer's (6a) NMR spectral
3
3
19
parameters were: F: ꢀ = � 101 (1F, dd, J = 36.6, 85.4),
�
114 (1F, dd, J = 109.8, 85.4), � 179 (1F, dd, J = 109.8,
1
3
3
.2. Preparation of trifluorovinyllithium
36.3); C: ꢀ = (all s unless stated otherwise) 25, 28, 35, 38,
8, 72 (d, J = 19.9), 131 (ddd, J = 236.9, 44.3, 13.2), 151
4
(1F, ddd, J = 287.2, 279.4, 51.0); H: ꢀ = 0.86 (9H, s), 1.6
1
For preparative purposes, we have found it best to pro-
3
ceed as follows. Under a dry nitrogen atmosphere a 100 cm
three-necked round-bottomed ¯ask ®tted with an acetone/
solid CO condenser and cooled in an acetone/solid CO
(4H, m), 1.78 (1H, m), 2.26 (4H, m), 4.59 (1H, s, exchange-
able with D O).
2
19
2
2
The minor isomer's (6b) were: F: ꢀ = � 104 (1F, dd,
J = 33.6, 88.5), � 115 (1F, dd, J = 106.8, 91.5), � 181 (1F,
3
bath was charged with anhydrous ether (40 cm ) and an
3
excess of 1,1,1,2-tetra¯uoroethane 1a (2±3 cm , 20±
3
2
13
dd, J = 106.8, 33.6); C: ꢀ = (all s unless stated otherwise)
22, 28, 33, 35, 48, 69 (d, J = 21.3 Hz), about 131 and 150
3
0 mmol). n-Butyllithium (2.5 M in hexanes, 8 cm ,
0 mmol) was added, with stirring, over half an hour using
1
(both partially obscured by signals from major isomer); H:

J. Burdon et al. / Journal of Fluorine Chemistry 99 (1999) 127±131
131
[
[
[
3] S. Dixon, J. Org. Chem. 21 (1956) 400.
4] R. Meier, F. Bohler, Chem. Ber. 90 (1957) 2344.
5] D.D. Callander, P.L. Coe, J.C. Tatlow, R.C. Terrell, J. Chem. Soc. (C)
1971) 1542.
6] P. Tarrant, P. Johncock, J. Savory, J. Org. Chem. 28 (1963) 839.
ꢀ
4
= 0.88 (9H, s), 1.0 (4H, m), 1.74 (1H, m), 1.96 (4H, m),
.27 (1H, s, exchangeable with D O).
2
(
3
.5. Reactions of other fluorohydrocarbons with n-
butyllithium
[
[7] D. Seyferth, D.E. Welch, G. Raab, J. Am. Chem. Soc. 84 (1962)
266.
4
[
8] J.F. Normant, J.P. Foulon, D. Masure, R. Sauvetre, J. Villieras,
Synthesis (1975) 122.
Reactions of 1,1,1-tri¯uoromethane, di¯uoromethane and
,1,1,2,2,2-hexa¯uoroethane with base (n-BuLi, t-BuLi,
1
[
9] J.P. Gillet, R. Sauvetre, J.F. Normant, Synthesis (1986) 355.
LDA, and t-BuOK were all tried with the ®rst two com-
pounds, and n-BuLi and LDAwith the last) were carried out
in the same way as for reactions with 1,1,1,2-tetra¯uor-
oethane 1a. Benzaldehyde (10 mmol) was added to a mix-
ture, usually in ether, of the substrate and the base (20 mmol)
that had been stirring at � 788C for about an hour. Work-up
was as with 1a and product analysis was by NMR.
[
[
10] I.B. Haslock, J. Burdon, P.L. Coe, R.L. Powell, J. Chem. Soc., Chem.
Commun. (1996) 49.
11] F.G. Drakesmith, R.D. Richardson, O.J. Stewart, P. Tarrant, J. Org.
Chem. 33 (1966) 286.
[
[
[
12] P.G. Gassman, N.J. O'Reilly, Tetrahedron Lett. 26 (1985) 5243.
13] P.G. Gassman, N.J. O'Reilly, J. Org. Chem. 52 (1987) 2481.
14] F.G. Drakesmith, O.J. Stewart, P. Tarrant, J. Org. Chem. 33 (1966)
472.
[
[
[
15] F. Tellier, R. Sauvetre, J.F. Normant, Y. Dromzee, Y. Jeannin, J.
Organomet. Chem. 331 (1987) 281.
16] I.B. Haslock, J. Burdon, P.L. Coe, R.L. Powell, accepted for
publication.
Acknowledgements
17] I.B. Haslock, J. Burdon, P.L. Coe, R.L. Powell, J. Fluorine Chem. 85
(1997) 151.
We thank ICI Klea Business for ®nancial support (for
IBH).
[
[
18] E.C. Ashby, J.T. Laemmle, Chem. Rev. 75 (1975) 521.
19] M.L. Capmau, W. Chodkiewicz, P. Cadiot, Tetrahedron Lett. (1965)
1
619.
20] G.D. Meakins, R.K. Percy, E.E. Richards, R.N. Young, J. Chem. Soc.
C) (1968) 1106.
[
[
References
(
21] J.M. Bainbridge, S.J. Brown, P.N. Ewing, R.G. Gibson, J.M. Percy, J.
Chem. Soc., Perkin Trans. 1 (1998) 2541.
[
[
1] J.F. Normant, J. Organomet. Chem. 400 (1990) 19.
2] D.J. Burton, Z.Y. Yang, P.A. Morken, Tetrahedron 50 (1994) 2993.