Reductive dehalogenation of polyfluoroarenes by zinc in aqueous ammonia

Article abstract of DOI:10.1016/S0022-1139(99)00082-2

Aqueous ammonia has been found to be a good and versatile medium for the highly selective hydrodehalogenation of polyfluoroarenes by zinc under unprecedentedly mild conditions. The reduction of pentafluorobenzoic acid, 2,3,4,5,6-pentafluorobenzyl alcohol, pentafluorobenzamide, pentafluoropyridine, heptafluoro-2-naphthoic acid, 1,3,4,5,7,8-hexafluoro-2-naphthoic acid, octafluoronaphthalene, octafluorotoluene, decafluorobiphenyl, chloropentafluorobenzene and 4-chlorotetrafluorobenzoic acid give products derived from the removal of one or two halogen atoms. A reduction mechanism, proceeding through electron capture to give a radical anion and then fragmentation of the latter, has been suggested. The observed high selectivity of the process suggests a radical anion formed by direct electron transfer from zinc to substrate. The dehalogenation regioselectivity is basically in accordance with that expected for radical anion fragmentation.

Full text of DOI:10.1016/S0022-1139(99)00082-2

Journal of Fluorine Chemistry 96 (1999) 175±185
Reductive dehalogenation of poly¯uoroarenes by
zinc in aqueous ammonia
*
Sergey S. Laev, Vitalii D. Shteingarts
N.N. Vorozhtsov Institute of Organic Chemistry, Siberian Division of the Russian Academy of Sciences, 9 Ac. Lavrentjev Avenue,
Novosibirsk 630090, Russian Federation
Received 11 January 1999; accepted 25 March 1999
Abstract
Aqueous ammonia has been found to be a good and versatile medium for the highly selective hydrodehalogenation of poly¯uoroarenes
by zinc under unprecedentedly mild conditions. The reduction of penta¯uorobenzoic acid, 2,3,4,5,6-penta¯uorobenzyl alcohol,
penta¯uorobenzamide, penta¯uoropyridine, hepta¯uoro-2-naphthoic acid, 1,3,4,5,7,8-hexa¯uoro-2-naphthoic acid, octa¯uoronaphthalene,
octa¯uorotoluene, deca¯uorobiphenyl, chloropenta¯uorobenzene and 4-chlorotetra¯uorobenzoic acid give products derived from the
removal of one or two halogen atoms. A reduction mechanism, proceeding through electron capture to give a radical anion and then
fragmentation of the latter, has been suggested. The observed high selectivity of the process suggests a radical anion formed by direct
electron transfer from zinc to substrate. The dehalogenation regioselectivity is basically in accordance with that expected for radical anion
fragmentation. # 1999 Elsevier Science S.A. All rights reserved.
Keywords: Reductive de¯uorination; Hydrode¯uorination; Reductive dechlorination; Zinc; Aqueous ammonia; Poly¯uoroarenes; Regioselectivity; Radical
anion; Fragmentation
�
1
. Introduction
from non-halogenated arenes in the system Zn/OH ±DMSO
[13] and from hypericin in the systems Zn/DMF or Zn/DMF±
H O [14]. A sequence offormation and fragmentation of RAs
Per¯uoroarenes and mixed perhalo¯uoroarenes can be
2
easily obtained through the displacement of chlorine or bro-
mine by ¯uorine in the respective perhaloarenes by the action
of a ¯uoride ion [1±3]. The reductive dehalogenation of these
products is, in turn, a promising route to partially ¯uorinated
arenes which are valuable starting compounds for synthesis
but generally less accessible than per¯uoroarenes and perha-
lo¯uoroarenes. However, the synthetic utility of the reduction
asappliedtotheone-stepreplacementof¯uorinebyhydrogen
in arenes is restrained by the scarcity of facile methods [4,5].
was proposed for the de¯uorination of per¯uoroole®ns by Zn
in DMF [15] and per¯uoroarenes by Zn/Cu in DMF±H O [7].
2
Therefore, the regioselectivity of reductive dehalogenation of
poly¯uoroarenes by zinc may be assumed to re¯ect the
regioselectivity of fragmentation of the respective RAs. In
1
this paper we focus on the reductive dehalogenation of
poly¯uoroaromatics (selected per¯uoroarenes, poly¯uoroar-
ene functional derivatives and perchloro¯uoroarenes) by zinc
in aqueous ammonia, i.e., to the best of our knowledge, the
simplest and most versatile among the reductive systems
currently known, yielding partially ¯uorinated arenes highly
selectively. The in¯uence of various factors (ammonia con-
LiAlH is usually employed for hydrode¯uorination of poly-
4
¯
uoroarenes [6]. The reduction by zinc (for example, hydro-
genolysis of some poly¯uoroarenes by Zn(Cu) in aqueous
DMF [7,8], of penta¯uorobenzonitrile [9] and penta¯uoro-
benzoic acid [10] by zinc in aqueous solution and of the latter
compoundinliquidammonia[11])isofparticularinterestdue
to both cheapness of a reducing agent, combined with the
relativesimplicityofexperimentalprocedures, andamechan-
ism likely occurring through the intermediate formation of
radicalanions(RAs)[12].TheformationofRAswasobserved
centration, presence of NH Cl or organic solvent) has been
4
revealed.
2. Results and discussion
2
.1. Reductive defluorination
In the ®rst instance substrates, such as poly¯uorobenzoic
acids and their derivatives, soluble in aqueous ammonia
*
Corresponding author. Fax.: +7-383-235-4752; e-mail:
1
shtein@nioch.nsc.ru
Preliminary results were reported in communications [16,17].
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 0 8 2 - 2

1
76
S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
Table 1
Reaction of compounds 1, 3, 5, 7, 9 and pentafluorobenzonitrile with zinc
Experiment
Starting compound
Time (h)
Arene:Zn mole ratio
NH
3
(%)
Product distribution (mol%)
Starting compound
Defluorination product
1
2
3
4
5
6
7
8
9
1
1
1
1
3
3
5
5
5
4.5
4.5
14
14
29
29
4
1:4
1:4
1:4
1:4
1:6
1:6
1:3.3
1:7
1:7
1:4
1:7
1:3.6
20
15
15
10
20
30
20
20
30
20
20
20
0
17
0
100
83
100
89
11
17
10
0
83
90
99
10
10
5.5
60
5
0
88
0
74
a
1
1
1
0
1
2
C
7
6
F
5
CN
0
ꢀ45
2
97
b
9
0
96 (100)
a
The content of 2,3,5,6-tetrafluorobenzonitrile.
Before (after) isolation.
b
were investigated. The reduction of penta¯uorobenzoic acid
1), is the reduction of ammonium benzoate, formed upon
dissolving the acid, at room temperature (Table 1) gave
,3,5,6-tetra¯uorobenzoic acid (2) (Scheme 1) as sole pro-
duct. The hydrode¯uorination of 1 at the position para to the
carboxylate group is in line with the previous reductions of
the same compound by zinc in liquid ammonia [11] and
some other systems [7,10,18] but the present experimental
technique is simpler. The complete conversion of 1 has been
achieved after 4.5 h for 20% ammonia concentration
was favorable. This conclusion is illustrated by the fact that
at room temperature compound 2 totally failed to be con-
verted into 4.
(
2
The same regioselectivity was observed for the reduction
of penta¯uorobenzamide (5). The conversion was complete
after 4 h for 20% ammonia to give 2,3,5,6-tetra¯uoroben-
zamide (6) (Scheme 3; Table 1, Exp. 7). Increase in the
reaction time (Table 1, Exps. 8 and 9) resulted in the partial
transformation of 6 into 2,3,5-tri¯uorobenzamide, 3,5-
di¯uorobenzamide and two unidenti®ed products. This
result is caused apparently by the presence of the more
strongly electron withdrawing amido group as compared
with benzoates generated by acids 1 and 3.
(
Table 1, Exp. 1). However, a decrease in ammonia con-
centration distinctly slows down the de¯uorination of 1
Exps. 2±4). Therefore, for the dehalogenation of other
(
poly¯uoroarenes we used 20±30% aqueous ammonia.
The same regioselectivity was observed for the reduction
of 2,3,4,5-tetra¯uorobenzoic acid (3) which produces 2,3,5-
tri¯uorobenzoic acid (4) (Scheme 2), but in considerably
increased reaction time (Table 1, Exps. 5 and 6). However,
besides a number of ¯uorine atoms, their location in the
aromatic ring are signi®cant. Recently, it was shown [19]
that the fragmentation from RAs of diverse poly¯uorinated
benzoates from the position para to the carboxylate group
Penta¯uorobenzonitrile gave a complicated mixture, the
components of which were assigned by 19F NMR analysis
as products of hydrode¯uorination with various residual
numbers of ¯uorine atoms ± 2,3,5,6-tetra¯uorobenzonitrile,
2,3,5-tri¯uorobenzonitrile and 3,5-di¯uorobenzonitrile, of
both monohydrode¯uorination and the conversion of the
cyano group ± 2,3,5,6-tetra¯uorobenzyl alcohol, 2,3,5,6-
tetra¯uorobenzylamine and compound 6, of nucleo-
philic substitution ± 4-amino-2,3,5,6-tetra¯uorobenzonitrile
(
Table 1, Exp. 10). This result is believed to be due to
the electron withdrawing power of the cyano group but,
nevertheless, the preference for primary de¯uorination of
penta¯uorobenzonitrile from para position is obvious.
The reduction by zinc in aqueous ammonia was also
successful for compounds without electron withdrawing
groups other than ¯uorine, such as 2,3,4,5,6-penta¯uoro-
benzyl alcohol (7), which was almost completely converted
Scheme 1.
Scheme 2.
Scheme 3.

S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
177
Scheme 6.
Scheme 4.
The reduction of hepta¯uoro-2-naphthoic acid (11) by
zinc in 20% aqueous ammonia at room temperature took
place readily at position 6 to yield 1,3,4,5,7,8-hexa¯uoro-2-
naphthoic acid (12) (Scheme 6; Table 2, Exp. 1). This
structure was assigned by 19F NMR spectroscopy. The
into a product of para-de¯uorination ± 2,3,5,6-tetra¯uor-
obenzyl alcohol (8) (Scheme 4; Table 1, Exp. 11). However,
in this case the reaction time was much longer as compared
to the conversion of 1 and 5. Besides the substituent
character, the likely reason of this is signi®cantly lower
solubility of 7 in aqueous ammonia as compared with
ammonium benzoates and even compound 5.
The reduction of penta¯uoropyridine (9) is worthy of
particular emphasis due to the presence of a very nucleo-
fugic ¯uorine atom in the position 4 [20,21], so it could be
expected to react rapidly with ammonia to give 4-amino-
presence of two peri-¯uorine couplings with J ꢁ60±
FF
80 Hz and the observed ¯uorine chemical shifts permit
two structures ± 12 and 1,3,4,5,6,8-hexa¯uoro-2-naphthoic
acid with the discrimination being based on the difference in
the JFF values observed in the two down®eld signals belong-
ing to F-1 and F-5, whereas a single value should be
observed for F-1 and F-8 down®eld signals in the spectrum
of 1,3,4,5,6,8-hexa¯uoro-2-naphthoic acid.
2
,3,5,6-tetra¯uoropyridine. Nevertheless, the reaction of 9
An increase in the reaction time was taken for the
de¯uorination of 12 to give complicated mixtures (Scheme
7) with the component distribution not dependent on ammo-
nia concentration but dependent on the reaction time. After
approximately 25 h, the principal product was 3,5,7,8-tetra-
¯uoro-2-naphthoic acid (15). The absence of peri-¯uorine
coupling and chemical shift values in 19F NMR spectrum
could also allow, besides 15, a structure of 1,3,4,7-tetra-
¯uoro-2-naphthoic acid. The discrimination in favor of 15 is
with zinc in 20% aqueous ammonia at room temperature
gave only a minor fraction of the amino derivative (ꢀ5%)
which was easily separated from the principal product ±
2
1
conditions the reduction rate of 9 is faster than that of the
nucleophilic substitution reaction. Moreover, the reaction is
remarkably rapid in spite of low solubility of 9 in water. The
observed regioselectivity of de¯uorination is the same as in
the reactions of this compound with LiAlH4 [20] and
Zn(Cu) [8].
Thus, in all the above cases a ¯uorine atom was lost from
the position para to the non-¯uorine substituent or heteroa-
tom and, as a rule, the primary de¯uorination products could
be obtained as individual compounds. Basically, under the
conditions used the products were substantially unreactive
,3,5,6-tetra¯uoropyridine (10) (Scheme 5; Table 1, Exp.
2) isolated in approximately 70% yield. Thus, under these
1
based on the H NMR spectrum (the presence of single
couplings with a JHF value of 6 or 10 Hz in two of three
hydrogen signals). Compound 15 is derived from the reduc-
tion of the primary de¯uorination products of 12 ± 3,4,5,7,8-
and 1,3,5,7,8-penta¯uoro-2-naphthoic acids (13, 14)
(Scheme 7), their 19F NMR signals were identi®ed in a
ꢀ4:1 ratio in the spectrum of a mixture obtained upon
incomplete conversion of 12. However, after long reaction
time (25 h or more) unidenti®ed compounds were present in
the reaction mixture along with 15, residual 13, 14 and, in
some cases, 12. The content of 15 in these mixtures did not
exceed 70% (Table 2, Exp. 2) but was increased at the
expense of 13, 14 and unidenti®ed products by the addition
(
compounds 2, 4, 8, and 10) or less reactive (compound 6)
than starting materials. In this connection polynuclear poly-
uoroarenes are of interest.
¯
of NH Cl to aqueous ammonia (Exp. 3). The results reveal
4
that the facility of de¯uorination of ¯uorinated naphthoic
acids by zinc in aqueous ammonia depends drastically on
the number of ¯uorine atoms, as is observed for the benzoic
acids.
Scheme 5.
Table 2
Reaction of naphthoic acids 11 and 12 with zinc
a
Experiment
Starting acid
Time (h)
Acid:Zn:NH
mole ratio
4
Cl
NH
20
3
(%)
Product distribution (mol%)
11
12
13
14
15
1
2
3
11
12
12
2
1:3:0
<1
97
2
<1
0
24±30
24
1:12.5:0
1:12.5:6.5
20±30
25±30
0±8
<1
10±18
4±5
8±12
3±4
59±71
85±86
a
The remainder of the mixture for Exps. 2 and 3 consists of unidentified compounds.

1
78
S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
Scheme 7.
Thus, aqueous ammonia is a medium allowing the highly
selective reductive de¯uorination of poly¯uoroarene func-
tional derivatives by zinc at room temperature. However, the
reduction of basic per¯uoroarenes, such as hexa¯uoroben-
zene, octa¯uorotoluene, octa¯uoronaphthalene and deca-
Scheme 9.
¯uorobiphenyl, seemed to be problematic because of their
very low solubility in aqueous solutions.
In spite of the presence of the strongly electron with-
drawing CF group in octa¯uorotoluene (16), its de¯uor-
ination (Table 3) required considerably increased reaction
time as compared with 1, 5 and 9. The content of the
de¯uorination product, a,a,a,2,3,5,6-hepta¯uorotoluene
considerably increased conversion of 18 (Exp. 6 vs. Exp. 5),
and even 3% of a de¯uorination product of 19 ± 1,2,4,5-
tetra¯uorobenzene ± was observed. Addition of NH Cl to
aqueous ammonia (Exp. 7) only slightly extended the con-
version of 18. Despite incomplete conversion, this result is
signi®cant, taking into account potential separability of the
products by recti®cation.
The reduction of octa¯uoronaphthalene (20) (Table 4)
was slower than for acid 11 to give mainly a product of the
removal of two ¯uorine atoms ± 1,2,4,5,6,8-hexa¯uoro-
naphthalene (21) (Scheme 10). This structure was assigned
by 19F NMR spectroscopy. The presence of peri-coupling in
two of three ¯uorine signals allows only two hexa¯uoro-
naphthalene structures with both hydrogen atoms in b
positions, such as 21 and 1,2,3,4,5,8-hexa¯uoronaphthalene
with the distinction between them being based on the
chemical shift value for the b-¯uorine atom signal. The
3
4
(
17), (Scheme 8) rose with ammonia concentration (Exps.
1 and 2) and remained practically invariant upon the addi-
tion of NH Cl (Exp. 3). The complete conversion of 16 has
been achieved in 20% ammonia (Exp. 4) when compound
4
17 was isolated in 86% yield. Thus, in this case the same
de¯uorination regioselectivity has been observed as for the
reduction of 1, 3, 5, 7, 9 in the present study and of 16 by
LiAlH [22].
4
The reduction of hexa¯uorobenzene (18) (Table 3) to
give penta¯uorobenzene (19) (Scheme 9) was clearly slower
as compared with 16. The use of 30% aqueous ammonia
Scheme 8.
Scheme 10.
Table 3
Reaction of octafluorotoluene (16) and hexafluorobenzene (18) with zinc
a
Experiment
Starting
Time
(h)
Arene:Zn:NH
mole ratio
4
Cl
NH
3
(%)
Product distribution (mol%)
compound
16 or 18
17 or 19
1
2
3
4
5
6
7
16
16
16
16
18
18
18
15
15
15
39
24
24
24
1:4.6:0
1:4.6:0
1:4.6:1.8
1:4.6:0
1:5:0
20
30
30
20
20
30
30
31
11
9
69
89
91
98
24
64
71
2
76
33
25
1:5:0
1:5:2
a
The remainder of the mixture for Exps. 6 and 7 consists of 1,2,4,5-tetrafluorobenzene.

S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
179
Table 4
Reaction of octafluoronaphthalene (20) with zinc
a
Experiment
Time (h)
20:Zn:salt
mole ratio
NH
3
(%)
Salt
Product distribution (mol%)
21
20
1
2
3
4
5
20
20
20
20
20
1:12:0
1:12:0
1:12:6
1:12:6
1:12:6
20
30
30
30
30
±
±
61±64
18±32
<1
30±32
56±67
88±91
35±41
32±40
4
NH Cl
NaCl
KBr
46±51
42±53
a
The remainder of the mixture consists of unidentified defluorination naphthalene.
formation of 21 is in accordance with the reduction of 20 by
(C H ) ZrCl ±Mg±HgCl in THF [23]. With 20% aqueous
5 5 2 2 2
ammonia (Exp. 1) the reaction gave 21 at low conversion
which increased with 30% ammonia (Exp. 2) and still
further with the addition of NH Cl, leading to the high
Scheme 11.
4
content of 21 in a product mixture (Exp. 3). The addition of
other electrolytes, such as NaCl or KBr, to aqueous ammo-
nia however slowed the reduction (Exps. 4 and 5 vs. Exp. 2).
The reduction of deca¯uorobiphenyl (22) (Table 5) was
more complicated as compared with 16, 18 and 20. This
compound was not converted at all in 20% aqueous ammo-
in the presence of NH Cl and THF (Exp. 7), compound 24
4
can be obtained in 27% yield.
The presented experimental data elucidate some regula-
rities. Firstly, de¯uorination drastically depends on the
number of ¯uorine atoms. As a rule, per¯uorinated com-
pounds are de¯uorinated more facilely than partially ¯uori-
nated ones. An exception was noted in the reduction of 20
which led mainly to the product of the removal of two
¯uorine atoms. Secondly, de¯uorination depends on the
location of the ¯uorine atoms (cf. acids 2 and 3, 13 and
14). Thirdly, de¯uorination depends on the aromatic frame-
work. Thus, the poly¯uorinated naphthalene compounds
under study are transformed more rapidly and more com-
pletely than the poly¯uorosubstituted benzene derivatives
(cf. acids 11 and 1, compounds 20 and 18). Finally, de¯uor-
ination depends drastically on the presence and nature of a
non-¯uorine substituent in the aromatic ring.
Assuming a mechanism of the reaction under study to
involve the formation of RAs as suggested previously
[7,12], it should proceed through three principal steps ±
dissolving of a compound to be reduced, electron extraction
from metal to give a RA, and decay of the latter to yield
eventually a hydrode¯uorination product. The last step is
believed to be a sequence of the irreversible fragmentation
0
0
nia (Exp. 1), but de¯uorination products, 2,2 ,3,3 ,4,
5,5 ,6,6 -nona¯uorobiphenyl (23) and 2,2,3,3 ,5,5 ,6,6 -
octa¯uorobiphenyl (24), (Scheme 11) formed upon the
0
0
0
0
0
addition of organic cosolvents, such as diethyl ether
(
ꢀ4% conversion for 22 h) or THF (Exp. 2). On reduction
in the presence of THF (Exp. 2), in addition to 23 and 24,
unidenti®ed biphenyls probably containing amino groups,
as indicated by the up®eld shifts of 19F NMR signals and IR
data, and probably formed through the nucleophilic sub-
stitution of ¯uorine were also observed. However, without
zinc, no hydrode¯uorination or aminode¯uorination
occurred. The content of hydrode¯uorination products
increased with ammonia concentration (Exps. 3 and 4) as
well as the amount of unidenti®ed products in the case of
THF as cosolvent (Exp. 4). The addition of NH Cl to 30%
4
aqueous ammonia started hydrode¯uorination of 22 in the
absence of an organic solvent (Exp. 5) and was more
promoted in the presence of Et O or THF (Exps. 6 and
2
7
). By the reaction of 22 with zinc in 30% aqueous ammonia
Table 5
Reaction of decafluorobiphenyl (22) with zinc
a
Experiment
Time (h)
22:Zn:NH
mole ratio
4
Cl
NH
3
(%)
Organic
Product distribution (mol%)
cosolvent
22
23
24
1
2
3
4
5
6
7
22
10
22
22
10
20
22
1:13.3:0
1:13.3:0
1:13.3:0
1:13.3:0
1:13.3:6.7
1:13.3:6.7
1:13.3:6.7
20
20
30
30
30
30
30
±
100
90
83
30
88
40
<1
0
3
0
2
THF
Et
2
O
9
7
THF
±
15
6
11
5
Et
2
O
48
5
11
68
THF
a
The remainder of the mixture for Exps. 2, 4 and 7 consists of unidentified biphenyls.

1
80
S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
with benzene, and the solubility factor are responsible for
the observed polyde¯uorination. The same is probably true
for the more facile de¯uorination of poly¯uoroarene func-
tional derivatives as compared with per¯uoroarenes.
Clearly, the addition of organic solvents to aqueous ammo-
nia considerably diminishes the role of the solubility factor,
thus facilitating the reaction. The ammonia concentration
in¯uence is most likely associated with the electron extrac-
tion rate from the metal as depending both on the activation
of the zinc surface and the solvation of the ensuing zinc
cation, thus affecting the redox potential of the zinc.
In the framework of the mechanistic concept outlined
above, the regioselectivity of this process should be
undoubtedly assigned to the fragmentation of poly¯uoroar-
ene RAs. According to rationalization based on a pseudo-
Scheme 12.
of RA with the removal of ¯uoride ion to produce an aryl
radical [12], its reduction and ®nal protonation of the
ensuing aryl anion (Scheme 12).
The elimination of ¯uoride ion from poly¯uoroarene RAs
generally proceeds quite fast and is unlikely to be the rate
limiting stage. This is con®rmed by the high fragmentation
rate constants of RAs of poly¯uorinated benzonitriles in
DMF [24], penta¯uoroacetophenone, penta¯uorobenzalde-
3
� 1
hyde (ꢂ10 s ), penta¯uorophenol, penta¯uoroaniline
6
� 1
[
H O. However, the lifetime of a RA tightly associated with
25] and poly¯uorinated benzoates [19] (ꢂ10 s ) in
*
Jahn±Teller effect [27], the decrease in the ꢀ±s energy gap,
2
a positive hole on the zinc surface could be much longer
than for a free RA in the solution. On the contrary, the low
solubility of an aromatic compound, which is typical for the
aqueous medium, can considerably slow down the reduc-
tion. Finally, electron extraction from zinc to form a RA is
probably the slow step that is suggested by the relatively
long duration of hydrode¯uorination of 1, 3, 11 and 12 in
this study, of some poly¯uoroaromatics by Zn(Cu) in DMF±
brought about by the multiple ¯uorine substitution, is
favorable for the pseudo-ꢀ-state of poly¯uoroarene RAs
with both the out-of-plane deviation of a C±F bond and the
fragmentation rate through a certain position being con-
trolled to a ®rst approximation by the energy gap between
*
the ꢀ-state, inherent in a planar RA, and the s -state of RA
with an odd electron density located mainly on a C±F bond
to be broken. This criterion is most favorable for a position
with the maximum spin density. The ESR spectra of RAs of
19 [28] and 9 [29] show that the principal electron density is
located just on the carbon atom para to a position with no
¯uorine, in agreement with the regioselectivity observed in
the hydrode¯uorination of these compounds. On the basis of
ESR spectra of poly¯uorinated benzonitrile RAs with ¯uor-
ine in the para position [30], the same should be expected
for penta¯uorobenzonitrile RA. For other poly¯uoroben-
zene derivatives under study (compounds 1, 3, 5, 7, 16, and
22) there is a lack of ESR data, but, nevertheless, there is no
doubt that their RAs should have the electron structure
which provides the observed regioselectivity. The INDO/
UHF calculations predict an odd electron in planar RAs of
19, 9, and penta¯uorobenzonitrile [27,30,31] to occupy the
b1-type (in terms of C2v group) ꢀ-MO located primarily in
the para and ipso positions.
H O [7,8] and of 1 by zinc in aqueous alkali [10] or liquid
2
NH [11], in spite of the absence of a dissolving problem. It
3
seems unlikely in this connection that the electron passes
out from metallic zinc into a solution, as in liquid ammonia
the blue color typical for solvated electrons is not observed
and the reduction process is prolonged to yield selectively a
monode¯uorination product, unlike the fast and deep
de¯uorination with removal of three or more ¯uorine atoms
observed with sodium [11]. In this connection, it is con-
ceivable that electron capture by poly¯uoroarene occurs
directly from the metal surface. Overall, the results with
compounds 1, 2 and 3 con®rm that the electron passing from
zinc into solution is not responsible for their reductive
de¯uorination, otherwise all the substrates and their de¯uor-
�
ination products, trapping e at practically the same rate
aq
constant [26], should undergo ¯uorine removal. The rate of
the substrate reaction with the metal to give a RA is
probably controlled by the substrate redox potential and,
since the removal of a halogen atom renders the redox
potential as a rule more negative [12,24], a monodehalo-
genation product is reduced more slowly than the starting
compound. Evidently, under the conditions used in most
cases the threshold of substrate ability for the electron
extraction from zinc, suf®cient for the reaction to proceed,
lies between a starting poly¯uoroarene and its monode¯uor-
ination product providing the observed monode¯uorination
selectivity. By contrast, the solubility of a monode¯uori-
nated product seems unlikely to be much lower as compared
with the starting compound and to account for this selec-
tivity. Evidently, for poly¯uorosubstituted naphthalene and
biphenyl derivatives both the higher electron af®nities,
inherent in the respective aromatic systems as compared
The ESR data for the RA of 20 [32] testify that the
electron density is greater in the a positions, while the
difference as compared with the b positions is not large.
Thus, the observed selective ¯uorine removal from the b
positions in the course of the reaction of 20 with zinc in
aqueous ammonia does not correspond to these data. Per-
*
haps, the involvement of an excited ꢀ -state with the
principal electron density located in the b positions, rather
than of a ground ꢀ-state, in the pseudo-Jahn±Teller inter-
action, takes place upon the greater out-of-plane deviation
and stretching of C±F bonds inherent in a transition state.
Also the presence of two neighboring ¯uorine atoms is
possible to facilitate the elimination of a ¯uorine atom from
the b position of RAs of 20 and 2-H-hepta¯uoronaphtha-
lene. Such an in¯uence of neighboring ¯uorine atoms on the
RA stability was noted for poly¯uorobenzoates [19]. The

S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
181
loss of ¯uorine from position 6 upon the formation of 12 and
1 from 11 and 2-H-hepta¯uoronaphthalene, respectively,
2
allows an analogy with the orientation of the de¯uorination
of 1 and 19 since the mutual location of a ¯uorine atom and a
non-¯uorine group in the positions 2 and 6 of the naphtha-
lene framework is analogous with the para location in the
benzene ring.
Scheme 15.
2
.2. Reductive dechlorination
zene (33), with a minor admixtures of 1,3-dichloro-2,4,6-
tri¯uorobenzene (32) and 1,3,5-tri¯uorobenzene (34)
(Scheme 15) but the reaction time was very long. However,
the reduction of 31 was considerably more rapid in 30%
aqueous ammonia (Exp. 2) and upon the addition of acetone
(Exp. 3) to decrease the content of 32 and increase the yield
of 34. The reaction of 31 with zinc in 30% ammonia at 608C
in an autoclave led to a signi®cantly enhanced yield of 34
(Exps. 4 and 5). Thus, there is no doubt that compounds 33
and 34 can be smoothly prepared by the reaction of 31 with
zinc in aqueous ammonia, being isolated by recti®cation.
The reduction of a- and b-chlorohepta¯uoronaphthalene
(35) (a 1:3 mixture of isomers obtained together with 20
from octachloronaphthalene [2]) in 20% aqueous ammonia
at room temperature for 5 h gave a mixture of poly¯uori-
nated naphthalenes with 21 as the principal component
The reduction of suf®ciently accessible perchloro¯uor-
oarenes is of particular interest since one could expect it to
proceed selectively through the C±Cl bond scission and to
be more facile as compared with hydrode¯uorination [8,33].
The reduction of chloropenta¯uorobenzene (25) by zinc
gave 19 (Scheme 13) with complete conversion after 9 h in
2
0% aqueous ammonia or 5.5 h with 30% ammonia con-
centration. Thus, the dehalogenation of 25 in aqueous
ammonia is much easier as compared with 18 as shown
in both duration (5.5 h vs. 24 h) and conversion (99% vs.
6
7%).
(
Scheme 16). The shorter reaction time resulted in a
decreased content of 21 and appearance of a considerable
amount of both hepta¯uoronaphthalenes. Thus, due to the
more facile removal of chlorine, occurring under milder
conditions as compared with ¯uorine, the conversion of b-
isomer to 21 can be partially suppressed.
Scheme 13.
The reduction of a mixture of dichlorotetra¯uoroben-
zenes (26) (obtained together with 18 and 25 from the
reaction of hexachlorobenzene with potassium ¯uoride
[
1], the content of meta-isomer was 74%) in 20% aqueous
ammonia gave a mixture of 1,2,3,5-tetra¯uorobenzene (27)
73 mol%), 1,2,4,5-tetra¯uorobenzene (28), 1,2,3,4-tetra-
uorobenzene (29), and 1-chloro-2,3,4,5-tetra¯uoroben-
(
¯
Scheme 16.
zene (30) (Scheme 14) for 32 h at room temperature.
Compound 27 can be isolated by recti®cation. Thus, the
dechlorinations of para- and meta-dichlorotetra¯uoroben-
zenes were faster than the ortho-isomer and led to complete
conversion into the respective products of the substitution of
both chlorine atoms by hydrogen. Meanwhile, under the
same conditions the reduction of 18 to compound 28 did not
occur, con®rming dechlorination to be considerably easier
than de¯uorination.
The reduction of 4-chlorotetra¯uorobenzoic acid (36) in
20% aqueous ammonia gave compound 2 (Scheme 17) with
complete conversion after 4 h. The observed regioselectivity
is the same as realized in the electrochemical reduction of
36 [18].
On the whole, in view of the reductions of 25, 31, 36, and
mixtures of isomers 26 and 35 (Schemes 13±17) we can
infer that zinc in aqueous ammonia is a good reagent for the
dechlorination of chloropoly¯uoroarenes on a preparative
scale.
The reduction of 1,3,5-trichloro-2,4,6-tri¯uorobenzene
31) (Table 6) in 20% aqueous ammonia at room tempera-
(
ture (Exp. 1) gave mainly the product of substitution of two
chlorine atoms by hydrogen, 1-chloro-2,4,6-tri¯uoroben-
The selective dechlorination of 25, 26, 31, and 36 is
compatible with much faster elimination of chloride ion
Scheme 14.

1
82
S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
Table 6
Reaction of 1,3,5-trichloro-2,4,6-trifluorobenzne (31) with zinc
Experiment
Temperature
and time
31:Zn mole
ratio
NH
3
(%)
Product distribution (mol%)
32
33
34
1
208C, 60 h
208C, 30 h
208C, 35 h
608C, 35 h
608C, 30 h
1:7.5
1:7.5
1:7.5
1:12
1:10
20
20
30
30
30
7
3
0
0
0
89
84
87
18
0
4
13
a
2
3
13
82
b
4
c
5
100
a
With the addition of acetone.
In a rotating autoclave.
b
c
In an autoclave with a stirrer.
INDO/UHF calculations [31] and spectral data [38], the
same is true for dichlorobenzene RAs.
3
. Experimental
Scheme 17.
Melting points were determined in sealed capillary and
boiling points were measured during distillation, both are
uncorrected. H and 19F NMR spectra were recorded in
CD COCD using a Bruker WP-200 SY spectrometer at
1
from monochloroarene RAs (at least one order of magnitude
difference in the fragmentation rate constant) as compared
with ¯uoride ion removal from mono¯uoro analogs [34±36].
In the framework of a concept of the rate limiting electron
extraction from zinc by a substrate (vide supra), this is
consistent with the more positive redox potentials found for
C H Cl and C HCl as compared with C H F and C HF
5
3
3
200.1 and 188.3 MHz, respectively. Chemical shifts are
reported with respect to TMS and CFCl . IR spectra were
3
measured on a Specord M80 spectrometer. HRMS data were
obtained on a Finnigan MAT-8200 high resolution mass
spectrometer. GLC analysis was performed on a HP 5890
instrument using the HP G1800A GCD system (capillary
GC column 0.26 mm/30 m, 0.25 mm ®lm HP-5 phase).
Aqueous ammonia was ``Pure'' grade, zinc powder was
``ZP-2''. Ammonia concentrations were 20±22% and 30±
32%. Organic solvents were reagent quality. Compounds 1,
7, 9, 16, 18, 20, 22, 25, 31 and penta¯uorobenzonitrile were
commercially available.
6
5
6
5
6
5
6
from electrochemical reduction [37]. The regularities ana-
logous to those for de¯uorination can be noted for dechlor-
ination. Firstly, it drastically depends on the number of
halogen atoms: at room temperature the dechlorination of
2
5 is relatively rapid, is longer for chlorotetra¯uoroben-
zenes and 32 and very slow for 33. Secondly, dechlorination
depends, to some extent, on the location of the halogen
atoms (the more rapid conversion of para- and meta-isomers
of 26 than of ortho-isomer). Also at room temperature the
removal of two chlorine atoms from 31 is more dif®cult than
from para- and meta-dichlorotetra¯uorobenzenes (the com-
plete conversion requires >60 h vs. 32 h for 20% ammonia
concentration). Thirdly, dehalogenation depends on the
aromatic framework. Thus, a- and b-chlorohepta¯uoro-
naphthalenes exhibited more rapid conversion than perha-
logenated benzenes. The rationalization of these regularities
is believed to be the same as the above one for hydrode-
3.1. Reaction of pentafluorobenzoic acid (1) with zinc
Compound 1 (2.12 g, 10 mmol) was added to zinc powder
(2.60 g, 40 mmol) in 20% aqueous ammonia (22 ml) and the
mixture was stirred at room temperature for 4.5 h (Table 1,
Exp. 1). Unreacted zinc was separated and washed with
water. The aqueous solution was acidi®ed with hydrochloric
acid and extracted with diethyl ether (3Â30 ml). The com-
bined ether extract was dried over MgSO . The solvent was
4
removed by distillation to yield 1.81 g of compound 2 (yield
93%), m.p. 150±1528C (lit. 152±1548C [39]).
1H; F NMR and IR spectra were identical to those of
an authentic specimen.
¯
uorination.
Unlike de¯uorination, the observed regioselectivity of
19
poly¯uorochloroarene reduction displays no relation to the
number and location of halogen atoms: the reaction pro-
ceeds selectively as dechlorination. The reason for this lies
evidently in the electron structures of their RAs. ESR data
3.2. Reaction of 2,3,4,5-tetrafluorobenzoic acid (3) with
zinc
[
38] revealed the RA of 25 to be a s-type radical with an
*
unpaired electron occupying the antibonding s -orbital
localized predominantly on the C±Cl bond. According to
Compound 3 [40] (0.97 g, 5 mmol) and zinc powder
(1.95 g, 30 mmol) in aqueous ammonia (20 ml) were

S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
183
stirred at room temperature for 29 h (Table 1, Exps. 5
3.7. Reaction of 1,3,4,5,7,8-hexafluoro-2-naphthoic acid
(12) with zinc
and 6). The isolation described for 1 was used. The
organic mixture (yield 92±93%) was analyzed by
19F NMR.
Compound 12 (1.12 g, 4 mmol), zinc powder (3.25 g,
0 mmol) and NH Cl (1.40 g, 26 mmol) were stirred in 25±
0% aqueous ammonia (25 ml) at room temperature for
4 h (Table 2, Exp. 3). 0.89 g of a crude product (see
5
3
2
4
3
.3. Reaction of pentafluorobenzamide (5) with zinc
Compound 5 [41] (0.63 g, 3 mmol) was added to zinc
isolation for 1) was twice crystallized from benzene to give
1
compound 15, m.p. 202.5±204.58C. H NMR, ꢁ: 7.61 (dt,
powder (0.65 g, 10 mmol) in 20% aqueous ammonia
18 ml) and the mixture was stirred at room temperature
H-6, Jˆ10 and 6 Hz), 7.80 (dd, H-4, Jˆ10 and 1 Hz), 8.66
(
1
9
(
1
1
dd, H-1, Jˆ6 and 1 Hz), 9.85 (s, CO2H). F NMR, ꢁ:
for 4 h (Table 1, Exp. 7). Unreacted zinc was separated and
washed with water and diethyl ether. The aqueous solution
was extracted with the ether (3Â20 ml). The combined ether
12.0 (F-3), 121.9 (F-5, Jˆ20 Hz), 138.3 (F-7, Jˆ18 Hz),
‡
51.5 (F-8, Jˆ18 and 20 Hz). HRMS m/z: 244.0148 (M ),
calculated C H F O ˆ244.0147. Anal. calc. for
11 4 4 2
extract was dried over MgSO . The solvent was removed by
4
C H F O : C, 54.12; H, 1.65; F, 31.13. Found: C, 53.66;
11 4 4 2
distillation to obtain 0.53 g of compound 6 (yield 92%),
m.p. 131.5±1338C (lit. 130.5±1328C [42]). 19F NMR , ꢁ:
H, 1.65; F, 30.70%.
138.6 (2F, F-3,5), 142.5 (2F, F-2,6), The IR spectrum was in
accord with literature data [42].
3
.8. Reaction of octafluorotoluene (16) with zinc
Compound 16 (5.9 g, 25 mmol) and zinc powder (7.5 g,
15 mmol) in 20% aqueous ammonia (40 ml) were stirred at
3
.4. Reaction of 2,3,4,5,6-pentafluorobenzyl alcohol (7)
with zinc
1
room temperature for 39 h (Table 3, Exp. 4) to obtain
after isolation described for 9 4.7 g of compound 17 (yield
8
According to the procedure for 5, compound 7 (0.99 g,
mmol) was reduced by zinc powder (2.28 g, 35 mmol) for
0 h (Table 1, Exp. 11) to obtain 0.65 g of compound 8
6%), b.p. 110.5±111.58C (lit. 111±1128C [22]).
H; F NMR spectra were identical to those of an authen-
tic specimen.
5
6
1
19
(
1
yield 72%), m.p. 31±32.58C (lit. 32±348C [18]).
H; 19 F NMR and IR spectra were in accord with literature
NMR [18] and IR [43] data.
3
.9. Reaction of hexafluorobenzene (18) with zinc
Compound 18 (5.2 g, 28 mmol), zinc powder (9.1 g,
40 mmol) and in Exp. 7 NH Cl (3.0 g, 56 mmol) in aqu-
3
.5. Reaction of pentafluoropyridine (9) with zinc
1
4
eous ammonia (45 ml) were stirred at room temperature for
4 h (Table 3, Exps. 5±7). The isolation described for 9 was
Compound 9 (4.22 g, 25 mmol) and zinc powder (5.85 g,
0 mmol) in 20% aqueous ammonia (30 ml) were stirred at
2
9
used. The organic mixture (yield ꢀ50±60%) was analyzed
room temperature for 5 h (Table 1, Exp. 12). The mixture
was diluted with water and gradually heated to 1008C.
Compound 10 was collected in a Dean±Stark trap, separated
from water and dried (yield ꢀ70%), b.p. 1028C (lit. 1028C
19
by F NMR.
3
.10. Reaction of octafluoronaphthalene (20) with zinc
1
19
[
[
20]). H; F NMR spectra were identical to literature
Compound 20 (0.90 g, 3.3 mmol), zinc powder (2.6 g,
0 mmol) and in Exps. 3±5 the salt (20 mmol) in aqueous
44].
4
ammonia (22±25 ml) were stirred at room temperature for
20 h (Table 4). The isolation described for 5 was used. The
organic mixture (yield ꢀ70±80%) was analyzed by
3
.6. Reaction of heptafluoro-2-naphthoic acid (11) with
zinc
1
9
F NMR and GLC. A crude product (0.56 g) from Exp.
Compound 11 [45] (1.19 g, 4 mmol) was stirred with zinc
3 was crystallized from ethanol to afford 0.35 g of com-
pound 21 (yield 45%). 19F NMR, ꢁ: 117.3 (F-4,8, Jˆ64 Hz),
136.4 (F-2,6), 149.4 (F-1,5, Jˆ64 Hz). Anal. calc. for
C H F : C, 50.87; H, 0.85; F, 48.28. Found: C, 50.64;
powder (0.78 g, 12 mmol) in 20% aqueous ammonia
25 ml) at room temperature for 2 h (Table 2, Exp. 1). A
crude product (see isolation for 1) was crystallized from
(
10 2 6
benzene to afford 0.96 g of compound 12 (yield 86%), m.p.
H, 0.78; F, 48.11%.
1
1
7
87±1898C. H NMR, ꢁ: 7.78 (dt, H-6, Jˆ11 and 6 Hz),
19
.99 (s, CO H). F NMR, ꢁ: 116.8 (F-5, Jˆ63.5 Hz), 117.6
3.11. Reaction of decafluorobiphenyl (22) with zinc
Compound 22 (1.00 g, 3 mmol), zinc powder (2.6 g,
40 mmol) and in Exps. 5±7 NH Cl (1.07 g, 20 mmol) in
aqueous ammonia (25 ml) with addition of organic solvent
(8 ml) in Exps. 2±4, 6 and 7 were stirred at room tempera-
2
(
F-1, Jˆ70 Hz), 134.9 (F-7), 139.9 (F-3), 147.5±148.3 (F-4
‡
and F-8, Jˆ63.5 and 70 Hz). HRMS m/z: 279.9963 (M ),
calculated C H F O ˆ279.9959. Anal. calc. for
1
1
2
6
2
4
C H F O : C, 47.16; H, 0.72; F, 40.69. Found: C, 47.43;
1
1 2 6 2
H, 0.82; F, 40.56%.

1
84
S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
ture (Table 5). The isolation described for 5 was used. The
organic mixture (yield ꢀ80±90%) was analyzed by
3.16. Reaction of 4-chlorotetrafluorobenzoic acid (36)
with zinc
19
F NMR and GLC. A crude product (0.79 g) from Exp.
was crystallized from ethanol to afford 0.25 g of solid
7
According to the procedure for 1, compound 36 [48]
(1.14 g, 5 mmol) was reduced by zinc powder (1.30 g,
20 mmol) for 4 h to give 0.91 g of compound 2 (yield
94%), m.p. 150±1528C.
containing 95% of compound 24 (yield 27%), m.p. 80.5±
2.58C (lit. 82±83.58C [46]). The IR spectrum was in accord
with literature data [46].
8
3
.12. Reaction of chloropentafluorobenzene (25) with zinc
Acknowledgements
According to the procedure for 9, compound 25 (10.1 g,
0 mmol) was reduced by zinc powder (10.4 g, 160 mmol)
5
in 20% or 30% aqueous ammonia (50 ml) for 9 or 5.5 h to
This work was supported by the Russian Foundation of
Basic Research under Grant 96-03-32984a.
obtain compound 19 (yield ꢀ60±65%), b.p. 85.58C (lit.
1
19
8
58C [47]). H; F NMR spectra were identical to those of
an authentic specimen.
References
3
.13. Reaction of dichlorotetrafluorobenzene mixture (26)
with zinc
[1] N.N. Vorozhtsov, V.E. Platonov, G.G. Yakobson, Izv. Akad. Nauk
SSSR, Ser. Khim. (1963) 1524.
[
2] G.G. Yakobson, V.D. Shteingarts, N.N. Vorozhtsov, Izv. Akad. Nauk
SSSR, Ser. Khim. (1964) 1551.
According to the procedure for 9, mixture 26 [1] (13.1 g,
0 mmol; the content of meta-isomer was 74%) was reduced
[
[
3] G. Filler, J. Chem. Soc. (1965) 6264.
6
4] M. Hudlicky, Reduction in Organic Chemistry, Ellis Horwood,
Chichester, 1984.
by zinc powder (22.8 g, 350 mmol) in 20% aqueous ammo-
nia (60 ml) for 32 h to give a mixture (yield 77%) containing
73% of 27, 7% of 28, 9% of 29 and 11% of 30 ( F NMR
data).
[
[
5] J.L. Kiplinger, T.G. Richmond, C.E. Osterberg, Chem. Rev. 94
(1994) 373.
19
6] M. Hudlicky, Chemistry of Organic Fluorine Compounds, Ellis
Horwood, Chichester, 1976.
[
[
[
7] V.I. Krasnov, V.E. Platonov, Zh. Org. Khim. 29 (1993) 1078.
8] V.I. Krasnov, V.E. Platonov, Zh. Org. Khim. 30 (1994) 1271.
9] Y. Masahiko, S. Masanori, N. Shusuke, JP02115156/1990, Chem.
Abstr. 113 (1990) 97208.
3
.14. Reaction of 1,3,5-trichloro-2,4,6-trifluorobenzene
31) with zinc
(
[
10] T. Nobuo, K. Osamu, N. Tomoaki, JP60258143/1986, Chem. Abstr.
104 (1986) 148523.
Compound 31 (4.7 g, 20 mmol) and zinc powder (9.8 g,
50 mmol) in aqueous ammonia (40 ml) with addition of
acetone (7 ml) in Exp. 2 were stirred at room temperature
Table 6, Exps. 1±3). The isolation described for 9 was used.
1
[11] S.S. Laev, V.D. Shteingarts, I.I. Bilkis, Tetrahedron Lett. 36 (1995)
655.
4
[
12] V.D. Shteingarts, L.S. Kobrina, I.I. Bilkis, V.F. Starichenko,
Chemistry of Polyfluoroarenes: Reaction Mechanisms and Inter-
mediates, Nauka, Novosibirsk, 1991 (in Russian).
(
The organic mixture (yield ꢀ65±80%) was analyzed by
1
9
F NMR.
Compound 31 (3.5 g, 15 mmol) and zinc powder in
[13] K.L. Handoo, K. Gadru, Tetrahedron Lett. 27 (1986) 1371.
[14] F. Gerson, G. Gescheidt, P. Haring, Y. Mazur, D. Freeman, H.
Spreitzer, J. Daub, J. Am. Chem. Soc. 117 (1995) 11861.
aqueous ammonia (35 ml) were heated in a rotating auto-
clave or autoclave with stirrer (50 ml) at 608C (Table 6,
Exps. 4 and 5). After cooling diethyl ether was added
to a mixture. Unreacted zinc was separated and washed
with the ether. The aqueous solution was extracted with
diethyl ether. The combined ether extract (ꢀ40 ml, yield
[
[
[
15] Ch.-M. Hu, F. Long, Z.-Q. Xu, J. Fluorine Chem. 48 (1990) 29.
16] S.S. Laev, V.D. Shteingarts, Tetrahedron Lett. 38 (1997) 3765.
17] S.S. Laev, V.D. Shteingarts, J. Fluorine Chem. 91 (1998) 21.
[18] F.G. Drakesmith, J. Chem. Soc., Perkin Trans. 1 (1972) 184.
[19] V.V. Konovalov, Yu.D. Tsvetkov, I.I. Bilkis, S.S. Laev, V.D.
Shteingarts, Mendeleev Commun. (1993) 51.
[
20] R.E. Banks, J.E. Burgess, W.M. Cheng, R.N. Haszeldine, J. Chem.
Soc. (1965) 575.
ꢀ
60%) was dried over MgSO and quantitatively analyzed
4
by 19F NMR.
[
21] G.A. Selivanova, T.V. Chuikova, A.A. Shtark, V.D. Shteingarts, Zh.
Org. Khim. 24 (1988) 2513.
[
[
[
22] D.J. Alsop, J. Burdon, J.C. Tatlow, J. Chem. Soc. (1962) 1801.
23] J.L. Kiplinger, T.G. Richmond, Chem. Commun. (1996) 1115.
24] N.V. Efremova, V.F. Starichenko, V.D. Shteingarts, Zh. Org. Khim.
3
.15. Reaction of a-, b-chloroheptafluoronaphthalene
mixture (35) with zinc
24 (1988) 57.
Mixture 35 [2] (1.15 g, 4 mmol; the ratio of a-, b-isomers
was 1:3) and zinc powder (1.24 g, 19 mmol) in 20% aqu-
[25] L.C.T. Shoute, J.P. Mittal, J. Phys. Chem. 100 (1996) 14022.
[
[
26] V.V. Konovalov, I.I. Bilkis, S.S. Laev, A.M. Raitsimring, Proceeding
of 5th Working Meeting on Radiation Interaction, Leipzig, 1990,
p. 204.
eous ammonia (20 ml) were stirred at room temperature for
5
h to give after isolation described for 5 a mixture (yield
75±85%) containing ꢀ75% of compound 21 ( F NMR
27] L.N. Shchegoleva, I.I. Bilkis, P.V. Schastnev, Chem. Phys. 82 (1983)
19
ꢀ
343.
and GLC data).
[28] M.B. Yim, D.E. Wood, J. Am. Chem. Soc. 98 (1976) 2053.

S.S. Laev, V.D. Shteingarts / Journal of Fluorine Chemistry 96 (1999) 175±185
185
[
[
[
[
29] M.B. Yim, S. Di Gregorio, D.E. Wood, J. Am. Chem. Soc. 99 (1977)
260.
[38] M.C.R. Symons, J. Chem. Soc., Faraday Trans. 1 77 (1981) 783.
4
[39] C. Tamborski, E.J. Soloski, J. Org. Chem. 31 (1966) 743.
[40] G.G. Yakobson, V.H. Odinokov, N.N. Vorozhtsov, Zh. Obshchei
Khim. 36 (1966) 139.
30] V.F. Starichenko, L.N. Shchegoleva, N.V. Efremova, V.D. Shtein-
garts, V.O. Saik, P.V. Schastnev, Chem. Phys. 100 (1985) 79.
31] L.N. Shchegoleva, P.V. Schastnev, Molecular Distortions in Ionic and
Excited States, CRC Press, Boca Raton, 1995.
[41] A.K. Barbour, M.W. Buxton, P.L. Coe, R. Stephens, J.C. Tatlow, J.
Chem. Soc. (1961) 808.
32] V.M. Grigoryants, O.A. Anisimov, I.I. Bilkis, Conference on
Chemically Induced Spin Polarization and Magnetic Effects in
Chemical Reactions, Abstracts, Novosibirsk, 1981, p. 144.
33] G.G. Yakobson, V.D. Shteingarts, N.E. Mironova, N.N. Vorozhtsov,
Zh. Obshchei Khim. 36 (1966) 145.
[42] T.N. Vasilevskaya, A.G. Badashkeeva, T.N. Gerasimova, V.A.
Barkhash, N.N. Voorozhtsov, Zh. Org. Khim. 6 (1970) 126.
[43] E.V. Aroskar, P.J.N. Brown, R.G. Plevey, R. Stephens, J. Chem. Soc.
C (1968) 1569.
[
[44] J. Lee, K.G. Orrell, J. Chem. Soc. (1965) 582.
[45] O.I. Osina, V.D. Shteingarts, Zh. Org. Khim. 10 (1974) 329.
[46] G.G. Yakobson, V.D. Shteingarts, A.I. Miroshnikov, N.N. Vorozht-
sov, Dokl. Akad. Nauk SSSR 159 (1964) 1109.
[
[
[
34] P. Neta, D. Behar, J. Am. Chem. Soc. 103 (1981) 103.
35] D. Behar, P. Neta, J. Am. Chem. Soc. 103 (1981) 2280.
36] V.V. Konovalov, A.M. Raitsimring, Yu.D. Tsvetkov, I.I. Bilkis,
Chem. Phys. Lett. 157 (1989) 257.
[47] E. Nield, R. Stephens, J.C. Tatlow, J. Chem. Soc. (1959) 166.
[48] K.V. Dvornikova, V.E. Platonov, G.G. Yakobson, Zh. Org. Khim. 27
(1991) 609.
[
37] M.D. Hawley, in: A.J. Bard (Ed.), Encyclopedia of Electrochemistry
of the Elements, Organic Section, vol. 14, part XIV-3, Marcel
Dekker, New York, 1980.