New Methods of Free-Radical Perfluoroalkylation of Aromatics and Alkenes. Absolute Rate Constants and Partial Rate Factors for the Homolytic Aromatic Substitution by n-Perfluorobutyl Radical

Article abstract of DOI:10.1021/jo970302s

New methods of free-radical perfluoroalkylation of aromatics and alkenes are reported. n-C₄F₉I has been utilized as source of C₄F₉^. radical through iodine abstraction by phenyl or methyl radical. The

Full text of DOI:10.1021/jo970302s

7128
J . Org. Chem. 1997, 62, 7128-7136
New Meth od s of F r ee-Ra d ica l P er flu or oa lk yla tion of Ar om a tics
a n d Alk en es. Absolu te Ra te Con sta n ts a n d P a r tia l Ra te F a ctor s
for th e Hom olytic Ar om a tic Su bstitu tion by n -P er flu or obu tyl
Ra d ica l
Anna Bravo, Hans-Rene´ Bjørsvik, Francesca Fontana, Lucia Liguori, Andrea Mele, and
Francesco Minisci*
Dipartimento di Chimica del Politecnico, via Mancinelli 7, I-20131 Milano, Italy
Received February 18, 1997X
New methods of free-radical perfluoroalkylation of aromatics and alkenes are reported. n-C4F9I
has been utilized as source of C4F9• radical through iodine abstraction by phenyl or methyl radical.
The reaction with alkenes, carried out in the presence of catalytic amount of Cu(OAc)2, leads to
substitution by a mechanism substantially identical to the aromatic substitution and not to the
usual chain addition of perfluoroalkyl group and iodine atom to the double bond. This has allowed
to measure for the first time the absolute rate constants and the partial rate factors for the homolytic
aromatic perfluoroalkylation by competition kinetics. The C4F9• radical shows a clear-cut electro-
philic character in the aromatic substitution, as already reported for the addition to alkenes, but
the low regio- and chemoselectivities suggest that the polar effect is not the main factor in
determining the high reactivity of perfluoroalkyl radicals toward aromatics (105-106 M-1 s-1, 2-3
orders of magnitude more reactive than alkyl radicals). The enthalpic factor, related to the involved
bond energies, appears to be the major cause of the increased reactivity. The polar effect is
considered as related more to the polarizability than to the polarity of a radical (the σ-perfluoroalkyl
radicals are considered less polarizable and hence less sensitive to polar effects than π-radicals).
Sch em e 1
Perfluoroalkyl iodides (RfI) have been widely utilized
as sources of perfluoroalkyl radicals (Rf•). The free-
radical chain addition to alkenes1 (eqs 1 and 2) and the
homolytic aromatic substitution by thermolysis2 (eq 3)
are well-known processes.
We have, therefore, developed new synthetic strategies
for the free-radical perfluoroalkylation of aromatics and
alkenes under milder conditions, based on eq 4 (i).
Moreover, eq 2 suggested that iodine abstraction from
RfI by a methyl radical (eq 6) should also be effective.
The fact that an effective free-radical chain occurs in
the addition to alkenes suggests that reaction 2 must be
fast enough; probably the polar effect (Scheme 1) plays
a favorable role in the iodine transfer.
Thus we realized that iodine abstraction by phenyl
radical from RfI (eq 4) should be very fast, on the basis
of the stronger Ph-I bond as compared with R-I bonds
and of the very high rate constant of iodine abstraction
from secondary alkyl iodides by phenyl radical3 (eq 5)
We have used the following three simple sources of
methyl radical4 for generating n-perfluorobutyl radical:
(ii) t-BuOOH and Fe(OAc)2OH; (iii) MeCOMe and H2O2;
(iv) DMSO/H2O2/Fe(II).
Aromatic perfluoroalkylation by RfI and equimolar
amounts of peroxides has been reported in a patent,5
which is for several aspects closely analogous to our
procedures i and ii, even if, of course, the mechanism was
not discussed. However, the stoichiometry was correctly
given and supported by product analyses; for example,
with (t-BuO)2, the products CH3I, t-BuOH, and MeCOMe
were isolated and quantified, and it was implicitly
suggested that eq 6 was involved. The main disadvan-
tages in the use of (t-BuO)2 instead of the t-BuOOH/
Fe(III) catalytic system for the aromatic substitution are
the higher temperature required for the generation of
t-BuO• radical (>130 °C), which, for several aromatic
X Abstract published in Advance ACS Abstracts, September 15, 1997.
(1) Yoshida, M.; Kamigata, N.; Sawada, H.; Nakayama, M. J .
Fluorine Chem. 1990, 49, 1 and references therein.
(2) Tiers, G. V. D. J . Am. Chem. Soc. 1960, 82, 5513. Cowell, A. B.;
Tamborski, C. J . Fluorine Chem. 1981, 17, 345.
(4) Fontana, F.; Minisci, F.; Vismara, E. Free Radicals in Synthesis
and Biology; Minisci, F., Ed., Kluwer Acad. Publishers: Dordrecht,
1989; p 53.
(3) Kryger, R. G.; Lorand, J . P.; Stevens, N. R.; Herron, N. R. J .
Am. Chem. Soc. 1977, 99, 7589.
(5) Brace, N. O. U.S. Pat. 3,271,441 (1966). We thank a referee for
this suggestion.
S0022-3263(97)00302-2 CCC: $14.00 © 1997 American Chemical Society

Free-Radical Perfluoroalkylation of Aromatics and Alkenes
J . Org. Chem., Vol. 62, No. 21, 1997 7129
Ta ble 1. Hom olytic Ar om a tic P er flu or o-n -bu tyla tion
substrates, RfI, and solvents, do not allow working at
atmospheric pressure and in general lower effectiveness
in the rearomatization of the intermediate cyclohexadi-
enyl radical adducts, as we will discuss later on.
con-
aromatic
substrate
pro- versiona yieldsb
orientation (%)
cedure
(%)
(%)
Very recently,6 the absolute rate constants for the
addition of perfluoroalkyl radicals to alkenes have been
reported for the first time, emphasizing the great increase
of reactivity of perfluoroalkyl radicals as compared to
benzene
-
-
-
-
i
88.7
58.1
46.8
41.9
87.5
57.7
45.2
42.2
86.6
90.1
86.8
88.1
88.8
91.3
86.9
89.4
94.7
96.1
93.6
95.4
93.3
94.8
96.7
97.6
90.9
93.5
47.1
41.4
42.3
97.2
99.1
97.6
benzened
benzened
benzened
anisole
ii
iii
iv
i
ii
iii
iv
i
i
i
i
i
o (14.7) m (45.5) p (39.6)
o (15.5) m (42.8) p (41.7)
o (15.9) m (41.1) p (43.0)
o (16.1) m (43.4) p (40.5)
•
anisoled
anisoled
anisoled
primary alkyl radicals (e.g. n-C3F7 is four orders of
magnitude more reactive than primary alkyl radicals in
the addition to 1-hexene). It was interesting to verify if
this behavior is also reflected in the addition to aromatic
diphenyl ether o (18.1) m (24.4) p (57.5)
toluene o (46.1) m (30.3) p (23.6)
chlorobenzene o (41.7) m (39.6) p (18.7)
benzonitrile
nitrobenzene
biphenylc
•
rings. The reactivity of n-CmF2m+1 radicals appears to
be the same in all cases when m > 2; thus, results
obtained with n-C4F9• can be extrapolated to the complete
series of n-perfluoroalkyl radicals, with the exception of
o (20.5) m (52.2) p (27.3)
o (31.1) m (36.7) p (32.2)
o (25.1)
R (80.4)
R (86.5)
m + p (74.9)
â (19.6)
i
i
i
naphthalene
thiophene
•
•
CF3 and C2F5 radicals.
â (13.1)
a
Conversion of C4F9I; by procedure i 94-98% of PhI, based on
Resu lts
b
converted C4F9I, were obtained in all cases. Yields based on
converted C4F9I. c The GLC separation of the meta from the
The n-perfluorobutylation of aromatic compounds by
C4F9I has been carried out according to four different
procedures.
prevalent para isomer is not complete enough for a quantitative
d
analysis.
The conversion of C4F9I increases to 70-80% by
doubling the amount of H2O2 or t-BuOOH.
(i) Ben zoyl P er oxid e. The reaction was simply
carried out by refluxing the reagents in acetic acid
solution (116 °C); other solvents, such as 1,2-dichloro-
ethane (DCE) or cyclohexane, appear to be less effective,
probably because hydrogen abstraction from the solvent
by n-C4F9 competes with the aromatic addition. The
reaction is quite clean; PhI and PhCOOH are the main
products from (PhCOO)2.
previously4 developed this procedure for obtaining alkyl
radicals from alkyl iodides. The reaction has been
applied also in this case to benzene and anisole, and the
results are summarized in Table 1.
•
(iv) H2O2, DMSO, a n d F e(II) Sa lt. We have previ-
ously4 developed this procedure as a general method to
obtain alkyl radicals from alkyl iodides; it also works with
Rf-I under very mild conditions (a few minutes at 20-
30 °C). The reaction takes place by Fenton reaction in
DMSO. This procedure has been already applied for
perfluoroalkylation in the pyrrole series,7 and the results
with benzene and anisole are reported in Table 1.
All the procedures failed to induce reaction of pyridine
with C4F9I, whereas the same reactions with alkyl iodides
provided alkylation of the pyridine ring with high yields
and selectivity.
n -P er flu or ob u t yla t ion of Cycloh exen e a n d
1-Octen e. Procedure i, modified by the presence of a
catalytic amount of Cu(OAc)2, was utilized. The allylic
n-perfluorobutylation of the alkene mainly took place.
With cyclohexene the main reaction product (85%) was
3-(perfluorobutyl)cyclohexene (1). n-Perfluorobutylcyclo-
hexane (2) was the only significant byproduct (15%).
Compound 2 was the main reaction product (73%)
obtained by procedure ii in the absence of Cu(OAc)2.
However, in the presence of a catalytic amount of
Cu(OAc)2, procedure ii also led to 1 (83%). Similarly,
with 1-octene procedure i gave trans-1-(n-perfluoro-
butyl)-2-octene (3) as the main reaction product (83%)
in the presence of Cu(OAc)2. The cis isomer (4, 16%) and
1-(n-perfluorobutyl)-1-octene (5, 2%) were the only sig-
nificant byproducts. With procedure ii 1,1,1,2,2,3,3,4,4-
nonafluorododecane, n-C4F9C8H17 (6, 73%), was obtained
in the absence of Cu(OAc)2, whereas 3, 4, and 5 were the
reaction products in the presence of Cu(OAc)2. In all the
experiments performed with benzoyl peroxide, PhI was
obtained in 92-94% yield based on converted C4F9I and
alkenes.
Neither significant hydrogen abstraction from the
solvent nor aromatic phenylation by phenyl radical were
observed, which supports our expectations about the high
rate of eq 4. The presence of catalytic amounts of Cu(II)
or Fe(III) salts, which generally make the rearomatiza-
tion of the intermediate cyclohexadienyl radical more
selective, in this case led to only little improvement,
indicating that the rearomatization step by (PhCOO)2 is
effective. This method is particularly useful because the
amount of PhI, 94-98% based on the converted C4F9I,
•
reflects the amount of C4F9 radical generated and
therefore provides the possibility to verify the efficiency
of the aromatic perfluoroalkylation. The procedure was
used with all the investigated aromatic compounds, with
the exception of pyrrole, which is sensitive to benzoyl
peroxide under the reaction conditions. The results are
reported in Table 1.
(ii) t-Bu OOH a n d F e(III) Sa lt. The reaction was
carried out by refluxing (116 °C) an acetic solution of the
aromatic compound, C4F9I, t-BuOOH, and a catalytic
amount of Fe(OAc)3. The procedure is less effective than
i, in that it needs more t-BuOOH, as compared to
(PhCOO)2, in order to obtain the same conversion of
C4F9I, but the yields based on the converted C4F9I are
as much high. It has been applied to benzene and
anisole, and the results are reported in Table 1.
(iii) H2O2 a n d Aceton e. The aromatic substitution
simply occurred by refluxing (56 °C) a solution of n-C4F9I,
H2O2, and the aromatic substrate in acetone in the
presence of CF3COOH as catalyst. Traces of transition
metal salts are harmful in this procedure because they
Com p etition Kin etics. (A) An excess of 1-octene and
aromatic substrate (ArH), in ratios depending on their
relative reactivities, were refluxed in acetic solution in
•
decompose H2O2 without generating C4F9 . We have
(6) (a) Avila, D. V.; Ingold, K. U.; Lusztyk, J .; Dolbier, W. R., J r.;
Pan, H. Q.; Muir, M. J . Am. Chem. Soc. 1994, 116, 99. J . Org. Chem.
1996, 61, 2027. (b) Dolbier, W. R., J r. Chem. Rev. 1996, 96, 1557.
(7) Baciocchi, E.; Muraglia, E. Tetrahedron Lett. 1993, 34, 3799.

7130 J . Org. Chem., Vol. 62, No. 21, 1997
Bravo et al.
Sch em e 2
Sch em e 3
Ta ble 2. Absolu te Ra te Con sta n ts a n d Rela tive Ra tes
•
for th e Rea ction s of C4F 9 Ra d ica l w ith Ar om a tic
Su bstr a tes
k/105 M-1 s-1
relative rates
B
aromatic
substrate
Aa
Ba
A
Ca
chlorobenzeneb
nitrobenzeneb
benzonitrileb
benzene
toluene
anisole
diphenyl ether
thiophene
biphenyl
0.20
0.21
0.23
1.0
1.4
3.4
3.6
4.8
5.8
10.8
Sch em e 4
2.4
3.8
7.6
2.1
3.6
7.7
1.0
1.6
3.2
3.7
4.2
5.3
10.7
-
1.0
1.7
3.7
4.0
5.0
6.3
8.9
8.4
10.1
12.8
25.7
10.6
13.3
26.6
-
naphthalene
12.7
-
pyrrolec
>100
-
a
A: Competition with 1-octene, assuming the value of 6.2 ×
106 M-1 s-1 for the addition of C4F9 to 1-octene. B: Competition
•
with cyclohexene, assuming the value of 1.2 × 106 M-1 s-1 for the
addition of C4F9• to cyclohexene. C: Competition between aromatic
b
substrates.
The absolute rate constants, evaluated by the
procedure C, are 4.4, 4.6, and 5.1 × 104, respectively, for chloro-
benzene, nitrobenzene, and benzonitrile. c Competition between
pyrrole and benzene by procedure iv (H2O2/DMSO/Fe(III)).
Discu ssion
the presence of C4F9I, (PhCOO)2, and Cu(OAc)2. The
reaction was clean: 3, 4, 5, and n-C4F9-Ar were the only
significant reaction products in addition to PhI and
PhCOOH. The rate constants were evaluated by Scheme
2.
Mech a n ism of Ar om a tic Ra d ica l P er flu or oa lk yl-
a tion . The reaction mechanism according to procedure
i clearly involves homolytic decomposition of benzoyl
peroxide with the formation of phenyl radical; the reac-
tion of this latter according to eq 4 appears to be much
faster (>108 M-1 s-1) than hydrogen abstraction from the
solvent (∼105 M-1 s-1)3 or than addition to the aromatic
ring (∼105 M-1 s-1),3 because no significant aromatic
phenylation was observed. The radical C4F9• adds to the
aromatic ring, and the radical adduct is then oxidized in
a chain process. The presence of a catalytic amount of
Cu(II) salt increases the reaction rate because a redox
chain is superimposed to the radical chain (Scheme 3).
The reaction mechanism operating in method ii is
based on the catalytic decomposition of t-BuOOH, initi-
ated by the slow reaction of t-BuOOH with Fe(III), and
a faster propagation chain (Scheme 4).
(B) The same procedure as in A was utilized with
cyclohexene: n-C4F9-Ar, 1, and 2 were the only reaction
products. A competition between 1-octene and cyclohex-
ene under these conditions has shown that 1-octene is
4.92 times more reactive than cyclohexene, in good
agreement with the value of 5.17 evaluated from the
absolute rate constants.6 A complete kinetic analysis by
measurement of the product ratios at various concentra-
tions of the two substrates was not carried out: thus,
the error limit is not defined. Nevertheless we consider
the rate constants reasonably accurate on the basis of
the high yields and the consistent results found for
1-octene and cyclohexene.
(C) The relative rates, kArH/kAr′H, were determined by
competition kinetics of two aromatic substrates with the
same procedure, the only exception being pyrrole, which
is sensitive to oxidation by procedures i-iii. In this case,
the competition was carried out by procedure iv, which
gave clean products. The relative rates are consistent
with those obtained by procedures A and B.
The C4F9• radical is formed, in this case, through eq 6,
the methyl radical being generated by â-scission of
t-BuO•, favored by temperature and by the use of AcOH
as solvent.
The stoichiometry of procedure iii is shown by eq 7: a
methyl radical is generated from acetone peroxide, acid
catalysis being necessary for achieving the equilibrium
between H2O2 and acetone peroxide (eq 8). Traces of
transition metal salts are, in this case, harmful because
they decompose H2O2, shifting equilibrium 8 at left and
preventing the formation of methyl radical by thermal
homolysis of acetone peroxide (eq 9). The rearomatiza-
All the absolute and relative rates are reported in
Table 2. In first approximation we assume the temper-
ature effect to be similar for the competing reactions
characterized by similar mechanism and on the ground
of consistent results by the procedures A-C.

Free-Radical Perfluoroalkylation of Aromatics and Alkenes
J . Org. Chem., Vol. 62, No. 21, 1997 7131
tion of the intermediate cyclohexadienyl adduct occurs
by induced homolysis of the acetone peroxide4 (eq 10).
pound is significantly lower than the yield, always high,
of PhI (Table 1); in these cases it would appear that
hydrogen abstraction from the solvent by n-C4F9 com-
•
petes with the addition to the deactivated aromatic ring,
even if an analysis of the volatile reaction products was
not carried out.
Aromatic substitution by procedures ii-iv is generally
cleaner than the previously reported substitutions by
thermolysis2 of RfI or by homolysis5 of (t-BuO)2. The
main reason is related, in our opinion, to the lower
efficiency of the rearomatization of the intermediate
cyclohexadienyl radical, which, under these conditions,
can also undergo disproportionation and dimerization (i.e.
eq 14), leading to high-boiling derivatives.
Reaction 11 represents the stoichiometry of procedure
iv; it would formally appear to be quite similar to reaction
7, but the mechanism for the generation of methyl radical
(eqs 12 and 13) is rather different. The substitution is
explained by a redox chain similar to the one described
by Scheme 2. The very fast reactions,8 eqs 12 and 13,
allow the use of mild conditions (a few minutes at room
temperature), suitable also for substrates, such as pyr-
role, which are very sensitive to oxidation; methods i-iii,
on the opposite, cannot be used with pyrrole.
(t-BuO)2 does not undergo induced homolysis, as do
(PhCOO)2 or acetone peroxide, and the rearomatization
by hydrogen abstraction (eq 15) is generally not selective
because the involved radicals are transient; therefore the
Ingold-Fischer effect is not fulfilled and the self-bimo-
lecular reactions of the radicals, such as eq 14, compete
with the cross-reaction (eq 15). However, when the
disproportionation prevails on the homo-coupling in eq
14, the aromatic substitution can become selective also
in these cases through further oxidation of the dihy-
drobenzene derivatives. The oxidation of cyclohexadienyl
radicals by Cu(II) or Fe(III) (Schemes 1 and 2) are, on
the contrary, very fast and selective in all cases, and they
are always preferable when a catalytic redox chain can
be achieved.
The solvent is important for the efficiency of the
aromatic perfluoroalkylation, because perfluoroalkyl radi-
cals can effectively abstract hydrogen from C-H bonds
of the solvent (the rates of abstraction by Rf• are >103
larger than those of the analogue hydrocarbon radicals).6b
•
Thus, the rate of hydrogen abstraction by n-C7F15 from
THF, whose R-position is particularly activated by both
enthalpic and polar factors toward electrophilic radicals,
has been evaluated at 6.1 × 105 M-1 s-1
. We expect
significantly lower rate constants for hydrogen abstrac-
tion from AcOH, DMSO, and MeCOMe by Rf•, always for
enthalpic and polar reasons. We have verified9 these
•
effects in hydrogen abstraction by C4F9 from 1-chloro-
hexane (the rates gradually increase from the methylene
group in position 1 to the one in position 5, and the
methyl group is less reactive than methylene). The
competition of hydrogen abstraction from the used sol-
vents (AcOH, DMSO, and MeCOMe) appears to be
negligible with activated aromatic substrates and the
efficiency of aromatic n-perfluoroalkylation is high, as
shown by the comparable yields of PhI and n-C4F9-Ar.
However, with deactivated aromatic substrates, such as
PhNO2 or PhCN, the conversion of the aromatic com-
No substitution takes place with pyridine in AcOH
solution, whereas nitrobenzene and benzonitrile are
attacked by C4F9 radical.
•
The fact that alkyl radicals, generated from alkyl
iodides by the same procedures, do not react with homo-
cyclic aromatic substrates suggested that perfluoroalkyl
radicals must be much more reactive than alkyl radicals.
On the other hand, protonated pyridine gives good results
(8) Minisci, F. Sulfur Centered Reactive Intermediates in Chemistry
and Biology; Chatgilialoglu, C., Asmus, K. D., Eds.; Plenum Press:
New York, 1990; p 303.
(9) Bravo, A.; Bjørsvik, H. R.; Fontana, F.; Liguori, L.; Minisci, F.
Chem. Commun., 1997, 1501.
•
with alkyl radicals, but is quite unreactive toward C4F9
under the same conditions.4 This would indicate that
alkyl and perfluoroalkyl radicals have opposite behavior,

7132 J . Org. Chem., Vol. 62, No. 21, 1997
Bravo et al.
Sch em e 5
Sch em e 6
due to their opposite polar character (electrophilic and
nucleophilic respectively for perfluoroalkyl and alkyl
radicals).
The electrophilic character of C4F9• radical is, however,
scarcely reflected on the orientation of the substitution.
The ortho, meta, and para positions of toluene, chloro-
benzene, anisole, diphenyl ether, benzonitrile, and ni-
trobenzene are substituted with low selectivity, without
a macroscopic influence of the polar character of the
substituent, the ortho position being, obviously, affected
by steric effects (e.g. the isomer distribution in chloroben-
zene2 is characterized by 52% in the ortho position by
Fe(III) salts, whereas the alkyl radical adduct is oxidized
by Cu(II), but it is reduced to C4F9CH2CH2CH2R (6) by
hydrogen abstraction from the large excess of solvent in
the presence of Fe(III) salt. If procedure ii is applied in
the presence of Cu(II) salt, eq 17 is prevailing. On the
other hand, it is well-known that alkyl radicals are
rapidly and selectively oxidized to alkenes by Cu(II) salts,
but that they are not oxidized by Fe(III) salts. With
cyclohexene, the product of reductive alkylation 2 is a
byproduct also in the presence of Cu(OAc)2, probably
because the C4F9 group in the â position diminishes to
some extent the oxidizability of the cyclohexyl radical.
The reaction is particularly successful because the higher
oxidation potential of perfluoroalkyl radicals prevents
their oxidation by Cu(II) salt. The same reaction cannot,
in fact, be applied to alkyl radicals, whose oxidation by
Cu(II) salts is much faster than their addition to alkenes.
•
CF3 , but only 14% by n-C10F21• radical). Substitution of
naphthalene, biphenyl, thiophene, and pyrrole is, on the
opposite, more selective. It would appear that the
orientation is related more to the stabilization of the
intermediate radical adducts (Scheme 5) than to the
electron availability of the ring position.
This behavior conflicts with that of electrophilic nitro-
gen-centered radicals, R2NH•+, in which a very high
activation is associated to a high selectivity in the
substitution of anisole and diphenyl ether (only the ortho
and para positions are substituted).10 It will be discussed
later on, on the basis of absolute rate constants and the
partial rate factors, how the polar effect is, in our opinion,
significant but it is not the major factor affecting the
reactivity and selectivity of the homolytic aromatic per-
fluoroalkylation, the main factor being the enthalpic
effect. The lack of reactivity of protonated pyridine must
be ascribed to the exceptional sensitivity to polar effects
of protonated heteroaromatic bases toward free radicals
(small variations of polarity are reflected in large changes
in reactivity).10
Mech a n ism of P er flu or oa lk yla tion of Alk en es.
The free-radical chain addition of perfluoroalkyl iodides
to alkenes (eqs 1 and 2) is well-known.1 We have
developed a new synthetic methodology leading to per-
fluoroalkyl derivatives of alkenes, based on eqs 13 and
17. The reaction takes place through addition of the
perfluoroalkyl radical to the double bond and oxidation
of the radical adduct by Cu(II) salt according to Scheme
6.
Absolu te Ra te Con sta n ts a n d P a r tia l Ra te F a c-
t or s in t h e H om olyt ic Ar om a t ic P er flu or oa lk yl-
a tion . It was recently shown6 that perfluoroalkyl radi-
cals are much more reactive than alkyl radicals toward
alkenes. It was suggested that6b “the dominant factor
that gives rise to the observed high reactivities of
perfluoroalkyl radicals, particularly in their addition to
electron-rich alkenes, would appear to be the high
electrophilicities of these very electron-deficient radicals”.
A dominant SOMO-HOMO interaction with large charge
separation in the transition state has been assumed for
these reactions. The prevalence of the polar effect over
the other effects would not, however, be plainly justified
by the relatively low value of F (-0.53) in the Hammett
correlation of the rates of addition of the n-C8F17• radical
to a series of para-substituted styrenes and by the fact
that n-perfluoroalkyl radicals react with acrylonitrile
even faster than primary alkyl radicals (2.2 × 106 and
5.3 × 105 M-1 s-1, respectively, for n-C8F17•6 and primary
alkyl11 radical at 25 °C); this means that the deactivating
polar effect of the cyano group is more than balanced by
the activating enthalpic effect.
The Cu(I) salt catalyzes the decomposition of benzoyl
peroxide, generating a redox chain. Thus, the mecha-
nism of alkene substitution is substantially similar to
aromatic substitution (Scheme 6 for alkenes and Scheme
3 for aromatics), the main difference being the higher
oxidizability of the cyclohexadienyl radical as compared
to the alkyl radical. The cyclohexadienyl radical is, in
fact, fast and selectively oxidized by both Cu(II) and
(10) Minisci, F. Substituent Effects in Free Radical Chemistry; Viehe,
H. G., Ed.; Reidel Publishing Co.: Dordrecht, 1986; p 391.
(11) Citterio, A.; Arnoldi, A.; Minisci, F. J . Org. Chem. 1979, 44,
2674.

Free-Radical Perfluoroalkylation of Aromatics and Alkenes
J . Org. Chem., Vol. 62, No. 21, 1997 7133
•
Ta ble 3. P a r tia l Ra te F a ctor s for th e Rea ction s of C4F 9
Sch em e 7
Ra d ica l w ith Ar om a tic Su bstr a tes
partial rate factors
aromatic
substrate
relative
ratesa
fo
fm
fp
chlorobenzene
nitrobenzene
benzonitrile
toluene
anisole
diphenyl ether
thiophene
0.20
0.21
0.23
1.57
3.43
3.77
4.67
11.4
0.25
0.18
0.12
2.04
1.51
2.05
12.17 (fR)
11.71 (fR)
0.24
0.22
0.31
1.86
4.70
4.70
1.84 (fâ)
2.85 (fâ)
0.22
0.38
0.32
1.76
8.15
11.8
The partial rate factors, however, would suggest that
the electrophilic character is a significant, but not the
major factor, in determining the increased aromatic
reactivity of perfluoroalkyl radicals compared to alkyl
radicals. The influence of strongly polar substituents,
such as MeO, CN, NO2, is not particularly relevant and
the selectivity in the meta/para orientation is always low
with both electron-withdrawing and electron-releasing
substituents, suggesting that the orientation is the result
of the superimposition of polar and radical stabilizing
effects of the substituents. Pyrrole represents the only
macroscopic exception to this behavior: the particularly
high reactivity (>107 M-1 s-1) and selectivity in the
R-position must be ascribed, in our opinion, to the low
aromaticity and to the conjugated diene character, which
makes this substrate as much reactive as styrenes6 by
combination of high electron availability with radical
stabilizing effects. The same effects (Scheme 5) increase
the reactivity and selectivity with thiophene, biphenyl,
and naphthalene to a minor extent.
naphthalene
a
The relative rates are the average of the values reported in
Table 2.
A strongly electrophilic radical, such as R2NH•+, is, for
example, highly reactive toward 1-hexene (>106 M-1 s-1),
but completely unreactive toward acrylonitrile.10 More-
over, n-C3F7 , n-C7F15•, and n-C8F17 radicals appear to
have identical reactivities, and, as such, the rates of these
radicals can be considered as generic n-perfluoroalkyl
•
•
•
radicals. Thus the results reported above for the n-C4F9
radical can be considered general for n-perfluoroalkyl
radical reactions with alkenes and aromatics.
On the basis of the above considerations, the knowl-
edge of the absolute rate constants for the addition of
perfluoroalkyl radicals to aromatic substrates was of
great interest for two reasons: (i) to verify if the large
increase in perfluoroalkyl reactivity with alkenes is also
reflected on aromatic substrates, (ii) the evaluation of the
partial rate factors with aromatic compounds should give
significant indications about the importance of the polar
effect. As far as we know, no absolute rate constant has
been reported for reactions of perfluoroalkyl radicals with
aromatic compounds. The recent evaluation of the
absolute rate constants for the addition of perfluoroalkyl
radicals to alkenes6 and the above-reported substitution
of alkenes with perfluoroalkyl radicals by a mechanism
substantially identical to the one of homolytic aromatic
substitution (Scheme 3 for the aromatic substitution and
Scheme 6 for the alkene substitution) made it easy to
measure the absolute rate constant for the homolytic
aromatic perfluoroalkylation by competition kinetics.
We have utilized cyclohexene and 1-octene in competi-
tion with a variety of organic compounds, n-C4F9I, and
benzoyl peroxide in acetic acid solution in the presence
of catalytic amount of Cu(OAc)2. It is reasonable to
assume the same rate constants for 1-octene and 1-hex-
ene; the values of 1.2 × 106 and 6.2 × 106 M-1 s-1 at 25
°C, respectively, for cyclohexene and 1-octene can be
considered reliable reference rate constants for measur-
Also, the similar partial rate factors for chloro-, cyano-,
and nitrobenzene, in spite of the differences between the
Hammett σ constants, are a further evidence that the
polar effect is not dominant, but that the higher polar
deactivating effect of the cyano and nitro groups com-
pared to chlorine is balanced by the higher stabilizing
effects of the intermediate cyclohexadienyl radical ad-
ducts (Scheme 7).
•
The fact that the C4F9 radical is much more reactive
with nitrobenzene (∼104 M-1 s-1) than primary alkyl
radicals are with benzene (∼102 M-1 s-1) is a significant
evidence that the main factor, which determines the
increased reactivity in aromatic perfluoroalkylation, is
not the polar factor. On the other hand, the few available
data13 for relative rates of alkyl radical reactions with
substituted benzenes would indicate that also with cyano-
and nitrobenzene perfluoroalkyl radicals should be more
reactive than alkyl radicals. This is also supported by
the fact that alkyl radicals, generated with the same
•
procedure and under the same conditions of n-C4F9
radical (without a large excess of aromatic substrates)
react only in traces with cyano- and nitrobenzene.
If we compare the partial rate factors of aromatic
perfluoroalkylation with those of homolytic amination by
Me2NH+• radical, in which the polar effect is certainly
dominant, we observe a much higher substituent effect
in this latter case: with anisole only the ortho and para
isomers are formed, and the partial rate factor for the
para position is ∼103, with naphthalene the partial rate
factors for the R and â positions are, respectively, 9.3 ×
104 and 2.9 × 103, whereas benzonitrile and nitrobenzene
are completely unreactive toward the Me2NH+• radi-
cal.10,14
•
ing the absolute rate constants for the reactions of n-C4F9
radicals with aromatic substrates by competition kinetics.
The agreement of the results with the two alkenes is
satisfactory and we have obtained a further good control
by competition kinetics with pairs of aromatic substrates.
The rate constants and the partial rate factors with a
variety of aromatic compounds, reported in Tables 2 and
3, confirm that perfluoroalkyl radicals are generally much
more reactive than alkyl radicals toward aromatic com-
pounds (i.e. the rate constant of primary alkyl radicals
reaction with benzene,10,12 ∼102 M-1 s-1, is much lower
than that of n-C4F9• radicals, ∼105 M-1 s-1) and that they
show, as expected, a general electrophilic character; the
behavior is qualitatively similar to that with alkenes.
(13) Shelton, J . R.; Uzelmeier, C. W. J . Am. Chem. Soc. 1966, 88,
5222. Cowley, B. R.; Norman, R. O. C.; Waters, W. A. J . Chem. Soc.
1959, 1799. Corbett, G. E.; Williams, G. H. J . Chem. Soc. 1964, 3437.
(14) Minisci, F. Top. Curr. Chem. 1976, 62, 1.
(12) Citterio, A.; Minisci, F., Porta, O.; Sesana, G. J . Am. Chem.
Soc. 1977, 99, 7960.

7134 J . Org. Chem., Vol. 62, No. 21, 1997
Bravo et al.
Ta ble 4. Rela tive Ra tes for th e Hom olytic Alk yla tion of
P r oton a ted 4-X-Su bstitu ted P yr id in es14
polarizability of σ-radicals is lower than that of π-
radicals.
For example, the extent of polar effect for phenyl and
cyclopropyl radicals (σ-radicals) is lower than that of the
cyclohexyl radical (π-radical) in the addition to benzene
and protonated pyridine derivatives10,14 (Table 5). The
moderate electrophilic behavior of C4F9• is also shown by
the moderate polar selectivity observed in the free-radical
iodination of the C-H bond in 1-chlorohexane9 (eq 18),
in which the selectivity is determined by hydrogen
kX/kH
X
n-butyla
sec-butyla
CN
COCH3
H
CH3
OCH3
20.3
5.6
1
0.3
0.1
259.0
55.6
1
0.2
0.02
a
Only position 2 is substituted, and the substituent effect is
•
therefore referred to the meta position.
abstraction by C4F9 (eq 19).
Ta ble 5. Rela tive Ra tes for th e Ad d ition of
Ca r bon -Cen ter ed Ra d ica ls to P h -X a n d P r oton a ted
4-X-P yr id in es13,14
radical
Ph
Ph
Ph
X
PhX (kX/kH)a
4′-X-C5H4NH+ (kX/kH)a,b
CN
3.7
1.1
1.2
1.7
3.6
1.8
27
1.9
1.6
0.6
Cl
This selectivity is much lower than that observed in
the chlorination by protonated N-chloroamines (eq 20),
in which the selectivity is determined by hydrogen
abstraction by the strongly electrophilic R2NH+• radical.10
(eq 21).
CH3
OCH3
CN
CH3
CN
Ph
0.3
cyclopropyl
cyclopropyl
cyclohexyl
cyclohexyl
14c
1.0
256c,d
1d
CH3
0.76
Ortho, meta, and para isomers are formed. b (kX/kH) is referred
a
c
to the R position (both R and â isomers are formed). (kCN/kMe) is
referred to the R position. Only the R isomer is formed.
d
Similar features are shown by other homolytic aro-
matic substitutions dominated by the polar effect, such
as the free radical alkylation and acylation of protonated
heteroaromatic bases. Also in these cases, the selectivity
is complete in the R and γ positions and the effect of the
polar substituents is large (Table 4), due to the nucleo-
philic character of the alkyl radicals, whereas alkyl
Con clu sion
New methods of perfluoroalkylation by perfluoroalkyl
iodides of aromatic compounds and alkenes have been
developed, through iodine transfer by phenyl and methyl
radicals. Absolute rate constants have been measured
for the aromatic substitution by competition kinetics, by
taking advantage of the known values of the absolute rate
constants for the addition of perfluoroalkyl radicals to
alkenes. Perfluoroalkyl radicals are more reactive than
alkyl radicals toward aromatics. The relatively low
values of the partial rate factors would indicate that this
enhanced reactivity should be attributed mainly to the
reaction enthalpy and only to a minor extent to the
electrophilic character of the perfluoroalkyl radical.
•
radicals bearing electron-withdrawing groups (X), X-CH2 ,
are completely unreactive. We have explained10 these
behaviors by the assumption that the Reactivity-
Selectivity Principle is a generally valid criterion for
radical reactions when the reactivity is mainly governed
by the reaction enthalpy; when polar effects are dominat-
ing, the Principle is reversed, a high reactivity being
associated to a high selectivity.10
All these considerations suggest that the main factor
determining the enhanced reactivity of perfluoroalkyl
radicals toward aromatic substrates is the enthalpic
effect. This is justified by the stronger (ca. 10 kcal/mol)
C-C bond formed when Rf• versus R• adds to unsaturated
systems (CH3CH3, BDE ) 91 kcal/mol versus CF3CH3,
BDE ) 101 kcal/mol).15 We expect this effect to be more
marked in aromatic than in alkene addition, due to the
less favorable enthalpy of the aromatic addition.
Exp er im en ta l Section
Gen er a l Meth od s. Mass spectra were performed on a
GLC-MS instrument, using a gas chromatograph equipped
with SBP-1 fused silica column (30 m × 0.2 mm i.d., 0.2 µm
film thickness) and helium as carrier gas.
GLC analyses were performed on a capillary gas chromato-
graph, equipped with SBP-5 fused silica column (25 m × 0.25
mm i.d., 1 µm film thickness) at a hydrogen flow rate of 8 cm3
min-1, PTV injector, flame ionization detector.
Another factor which, in our opinion, contributes to a
relatively low polar effect in perfluoroalkylation is the
σ-nature of perfluoroalkyl radicals (markedly pyrami-
dal).6 The polar effect, being kinetic in nature, is not an
intrinsic property of a free radical; it only makes sense
to mention the polar effect associated to a particular
reaction, which is manifested as charge separation in the
transition state. In principle, and sometimes also in
practice, the same radical can behave as electrophilic,
nucleophilic or nonpolar species, depending on the par-
ticular reaction involved.10 Thus the polarity and the
polarizability of both radical and substrate are important
factors, contributing to the extent of polar effect; the
1H-NMR spectra were carried out on a spectrometer operat-
ing at 400 MHz. Chemical shifts are referenced to internal
TMS.
19F-NMR spectra were registered on a spectrometer operat-
ing at 235 MHz. Chemical shifts are referenced to internal
CFCl3. Proton NMR spectra with 19F broadband decoupling
were acquired using a supplementary frequency synthesizer.
The alkenes, the aromatic substrates, C4F9I, and the
solvents were obtained from commercial sources and were used
without further purification.
Typ ica l E xa m p les for Ar om a t ic P er flu or o-n -b u t yl-
a tion . (n -P er flu or obu tyl)ben zen e, CF 3CF 2CF 2CF 2C6H5.
Meth od i: 5 mmol of benzene, 1 mmol of C4F9I, 1 mmol of
benzoyl peroxide, and 0.1 mmol of Cu(OAc)2 in 10 mL of acetic
acid were refluxed for 4 h (115 °C). The solution was directly
analyzed by GLC, GLC-MS, and 19F-NMR (CD3COCD3 10%
v/v was added to the crude reaction mixture to obtain the lock
(15) McMillen, D. F.; Golden, D. M. Annu. Rev. Phys. Chem. 1982,
33, 493.

Free-Radical Perfluoroalkylation of Aromatics and Alkenes
J . Org. Chem., Vol. 62, No. 21, 1997 7135
Ta ble 6. MS (m /z) for P er flu or o-n -bu tyl Ar om a tics
para isomers. Thus, for example, the reaction of chlorobenzene
with C4F9I was carried out by method i. The GC and GC/MS
analyses revealed the presence of 0.13 mmol of unreacted
C4F9I, 0.82 mmol of C6H5I, and 0.47 mmol of the three isomers
of C4F9-C6H4Cl (all the isomers show the molecular ion 330
aromatic
benzene
toluene
anisole
chlorobenzene
nitrobenzene
benzonitrile
biphenyl
diphenyl ether
naphthalene
thiophene
pyrrole
M+•
(ArCF2)+
296
310
326
330
341
321
372
388
346
302
285
127
141
157
161
172
152
203
219
177
133
116
+
and the main fragment Cl-C6H4-CF2 161). The chromatog-
raphy on silica gel (hexane as eluent) allowed the separation
of the ortho isomer: NMR δ (ppm): 7.35 (m, 1H), 7.48 (m, 2H),
7.68 (d, 1H), but not the isomers meta and para. Authentic
samples of these isomers were prepared by the known proce-
dure2 from m-chloro- and p-chloroiodobenzene, C4F9I, and
copper bronze in DMSO; for the meta isomer NMR δ (ppm):
7.39 (m, 1H), 7.48 (m, 2H), 7.60 (s, 1H); for the para isomer
NMR δ (ppm): 7.42 (d, 2H), 7.53 (d, 2H). The isomers were
identified by comparison with the authentic samples and the
isomer distribution was as follows: ortho 41.7%, meta 39.6%,
and para 18.7%. The procedure of ref 2 is simple and useful
for the unequivocal synthesis of the single isomers, whose
structures can be simply verified by GC/MS analysis. In GC
analysis the retention times of the ortho isomers are always
shorter, while those of the meta and para isomers are very
close to one another so that, in the case of biphenyl, GC
separation of the meta and para isomers was not complete
enough for a quantitative analysis.
signal). The GLC and GLC-MS analyses with p-iodotoluene
as internal standard revealed the presence of 0.12 mmol of
unreacted C4F9I, 0.85 mmol of C6H5I, and 0.81 mmol of n-C4F9-
C6H5. The pure product, n-C4F9-C6H5, was isolated by dilution
of the reaction mixture with water, extraction by hexane, and
flash chromatography on silica gel (hexane as eluent); it was
identified by comparison (GLC-MS) with an authentic sample
obtained by a known procedure.16 The absolute yields of
n-C4F9-Ph were determined by 19F-NMR using tetrafluoro-p-
benzoquinone as internal standard (signal at -142.3 ppm
relative to CFCl3); known mixtures of n-C4F9-C6H5 or n-C4F9I
(-CF2I signal at -62.3 ppm) and tetrafluoro-p-benzoquinone
were utilized to verify the response of the standard. The
results (0.84 mmol of n-C4F9-Ph and 0.11 mmol of unreacted
n-C4F9I) were in good agreement with those obtained by GLC
analysis.
Meth od ii: 5 mmol of benzene, 1 mmol of C4F9I, 2 mmol of
t-BuOOH, 0.2 mmol of Fe(OAc)2OH in 10 mL of acetic acid
were refluxed for 4 h (115 °C). The solution was analyzed by
19F-NMR as in i. 0.42 mmol of unreacted n-C4F9I and 0.56
mmol of n-C4F9-C6H5 were obtained. Under the same condi-
tions, when 4 mmol of t-BuOOH were used, only 0.18 mmol of
unreacted n-C4F9I were obtained and 0.77 mmol of n-C4F9-
C6H5 were formed.
n-BuI and cyclohexyl iodide provided high yields (>80%) of
PhI, but only traces of alkylaromatics (<1%) under the
conditions of procedure i.
P er flu or o-n -bu tyla tion of Cycloh exen e. (A) A mixture
of 5 mmol of cyclohexene, 1 mmol of C4F9I, 1 mmol of benzoyl
peroxide, and 0.1 mmol of Cu(OAc)2 were refluxed for 4 h in
10 mL of acetic acid. The solution was filtered on silica gel in
order to eliminate Cu(OAc)2, and then it was directly analyzed
by GLC, GLC-MS, and 19F-NMR. Two fluorinated products
were formed: the main product arises from the allylic substi-
tution of cyclohexene by the C4F9 group, 3-(n-perfluorobutyl)-
cyclohexene, 1 (85%). The minor product (15%) is (n-perfluoro-
butyl)cyclohexane, 2. MS (m/z) are characterized, for 1 and
2, respectively, by the molecular ions (300 and 302) and the
+
+
main fragments C6H9 (81) and C6H11 (83). 19F-NMR for 1
shows an AB system for the CF2 attached to the allylic position
[A from AB ) -114.5 ppm, B from AB ) -117.0 ppm, J (AB)
) 300 Hz], the CF2 in the â position at -123.0 ppm (m), the
CF2 in the γ position at -127.6 ppm (m), and the CF3 at -82.2
ppm. For 2 the CF2 attached to the cyclohexyl group at -119.5
ppm (s), the CF2 in the â position at -123.0 ppm (s), the CF2
in the γ position at -127.2 ppm (s), and the CF3 at -82.4 ppm
(s). The conversion of C4F9I is 98%; the overall yield of
perfluorobutyl derivatives is 87%, based on converted C4F9I.
PhI was obtained in 96% yield based on converted C4F9I.
(B) A mixture of 5 mmol of cyclohexene, 1 mmol of C4F9I, 2
mmol of t-BuOOH, and 0.2 mmol of Fe(OAc)3 were refluxed
for 4 h in 10 mL of acetic acid. The solution was filtered on
silica gel and directly analyzed by GLC, GLC-MS, and 19F-
NMR. Only 2 is formed in 68% yield based on C4F9I. MS and
19F-NMR are identical to those of the pefluoro-n-butyl cyclo-
hexane obtained in A. The conversion of C4F9I is 86%; the
overall yield of perfluorobutyl derivatives is 73%, based on
converted C4F9I.
Meth od iii: 5 mmol of benzene, 1 mmol of C4F9I, 3 mmol
of H2O2 (35% in water), and 6 mmol of CF3COOH in 10 mL of
acetone were refluxed for 8 h (57 °C). The solution was
analyzed by 19F-NMR as in i. 0.53 mmol of unreacted n-C4F9I
and 0.44 mmol of n-C4F9-Ph were obtained. Under the same
conditions, when 6 mmol of t-BuOOH were used, only 0.25
mmol of unreacted n-C4F9I were obtained and 0.71 mmol of
n-C4F9-Ph were formed.
Meth od iv: 3 mmol of H2O2 (35% in water) were dropped
while stirring at room temperature into a mixture of benzene
(5 mmol), n-C4F9I (1 mmol), FeSO4‚7H2O (0.7 mmol) in 7 mL
of DMSO. After 10 min the mixture was diluted in water and
extracted with hexane. The hexane solution was analyzed by
19F-NMR as in i: 0.58 mmol of unreacted n-C4F9I and 0.40
mmol of n-C4F9-C6H5 were obtained. Under the same condi-
tions, when 6 mmol of H2O2 were used, only 0.28 mmol of
unreacted n-C4F9I were obtained and 0.67 mmol of n-C4F9-Ph
were formed.
Procedure i was utilized for all the aromatic compounds
reported in Table 1, while procedures ii-iv were utilized for
benzene and anisole.
The MS of all the perfluoro-n-butyl aromatic derivatives are
specific for the molecular ion (M+) and the main fragment
(ArCF2)+; the results are reported in Table 6.
(C) The reaction was carried out as in B in the presence of
0.1 mmol of Cu(OAc)2; the results are similar to those obtained
in A (83% of 1 and 17% of 2; the overall yield is 57% based on
converted C4F9I.
P er flu or o-n -bu tyla tion of 1-Octen e. (A) The reaction
and analyses were carried out as for cyclohexene. The trans
The absolute yields were determined by 19F-NMR, using
tetrafluorobenzoquinone as standard, the chemical shifts,
relative to CFCl3, being in the range -104/-109 ppm for the
R-CF2, -121 ppm for the â-CF2, -125 ppm for the γ-CF2, and
-81 ppm for CF3. The isomer distribution has been deter-
mined by GLC and CLG-MS by comparison with results
obtained by known1,2,16 procedures, assuming the gas chro-
matographic response to be similar for the ortho, meta, and
isomer of the allylic perfluorobutylation, C4F9CH2CHdCHC5H11
,
3, was the main reaction product (82%), the cis isomer 4 being
16%, while only traces (<2%) of 5 were obtained. The
conversion of C4F9I is quantitative and the yields, based on
C4F9I, were 96%. C6H5I was obtained in 93% yield based on
C4F9I. MS (m/z) shows the molecular ion (330), and 19F-NMR
gives four signals at -114.0, -124.6, -126.5, and -81.8 ppm,
respectively, for R, â, and γ CF2 and CF3 for 3. The structure
of the main reaction product, 3, was elucidated on the basis
(16) Yoshida, M.; Amemiya, H.; Kobayashi, M.; Sawada, H.; Hagii,
H.; Aoshima, K. J . Chem. Soc., Chem. Commun. 1985, 234. Sawada,
H.; Nakayama, M.; Yoshida, M.; Yoshida, T.; Kamigata, N. J . Fluorine
Chem. 1989, 46, 423. Kamigata, N.; Ohtsuka, T.; Fukushima, T.;
Yoshida, M.; Shimizu, T. J . Chem. Soc., Perkin Trans. 1 1994,1339.
1
of H-NMR and with 19F broadband decoupling. The multi-
plets at 5.80 and 5.42 ppm were assigned to the vinylic
hydrogens, respectively, in the â- and γ-position relative to the

7136 J . Org. Chem., Vol. 62, No. 21, 1997
Bravo et al.
1
F igu r e 1. Expansion of the vinylic region of the H-NMR spectrum of compound 3. Lower trace: normal NMR spectrum. Upper
trace: Lorentz-Gauss resolution enhancement for spectral analysis. Ha: δ (ppm) 5.80, dtt, J ab 15.3 Hz, J ac 6.7 Hz, J ad 1.5 Hz; Hb:
δ (ppm) 5.42, dttt, J ab 15.3 Hz, J bd 7.2 Hz, J bc 1.4 Hz, J HF 0.7 Hz.
Rf moiety. The multiplicity of the low field multiplet can be
explained with the presence of a vicinal coupling to the other
vinylic proton (3J ) 15.3 Hz), a second vicinal coupling to CH2
(3J ) 6.7 Hz) and an allylic coupling to the CH2 group directly
attached to Rf (4J ) 1.5 Hz).
mmol of C4F9I, 1 mmol of benzoyl peroxide, and 0.1 mmol of
Cu(OAc)2 in 15 mL of AcOH were refluxed for 4 h (116 °C).
The solution was filtered through silica gel and analyzed by
GC by using p-iodotoluene as internal standard. The reaction
products are as follows: 1 (0.47 mmol), 2 (0.08 mmol), ortho,
meta, and para n-(perfluorobutyl)anisoles, respectively, 0.06,
0.15, and 0.15 mmol. 1 and 2 and the aromatic isomers were
identified in the preparative experiments. MS (m/z) of the
three isomers show the molecular ion 296 and the main
The multiplet at 5.42 is further complicated by long-range
heteronuclear coupling to CF2 and can be analyzed on the basis
of the following coupling constants: 3J ) 15.3 Hz; 3J ) 7.2
4
4
Hz; J ) 1.4 Hz; J H-F ) 0.7 Hz. These data were confirmed
by acquiring H spectra with 19F broadband decoupling.
+
1
fragment MeO-C6H4-CF2 157; almost identical are the 19F
The (E) stereochemistry of the double bond can be inferred
by the large value (15.3 Hz) of the vicinal J coupling between
the vinylic protons, while the presence of long-range H-H
allylic couplings on the proton not coupled to CF2 allows us to
rule out the formation of the addition product Rf-
CHdCH(CH2)5CH3. In Figure 1 the 1H-NMR spectrum of 3
NMR spectra relative to CFCl3, with R-CF2 -104/-106, â-CF2
-12, γ-CF2 -125, CF3 -81 ppm. 1H-NMR of the meta isomer
gave δ (ppm): 3.81 (s, 3H), 7.05 (d, 1H), 7.13 (s, 1H), 7.2 (d,
1H), 7.38 (t, 1H); for the para isomer δ (ppm): 3.80 (s, 3H),
7.02 (d, 2H), 7.44 (d, 2H). A competition between cyclohexene
and 1-octene has shown that 1-octene is 4.92 times more
reactive than cyclohexene, in good agreement with the value
of 5.17 derived from the absolute rate constants.
with
a complete analysis of vinyl and allylic protons is
reported, as suggested by the referee.
(B) The reaction and analyses were carried out as in B with
cyclohexene; the reduced perfluoro-n-butyl derivative was
obtained in 73% yield, based on converted C4F9I. MS (m/z)
shows the molecular ion (332) and 19F-NMR gives four signals
at -119.2 (s, CF2CH2), -123.1 (s, CF2 in the â position), -127.5
(s, CF2 in the γ position), and -82.1 ppm (CF3, s).
(C) Benzene and another aromatic compound (10 mmol on
the whole), C4F9I (1 mmol), benzoyl peroxide (1 mmol), and
Cu(OAc)2 (0.1 mmol) in 15 mL of AcOH were refluxed for 4 h
(115 °C). The solution was analyzed as in A. The ratios
between benzene and the other aromatic substrates were 1:5
for chlorobenzene, benzonitrile, and nitrobenzene, 1:1 for
toluene, 4:1 for anisole, diphenyl ether, thiophene, and bi-
phenyl, and 10:1 for naphthalene. The results are reported
in Table 2. In all the procedures A, B, and C the conversion
of C4F9I is >85%, and the overall yields based on converted
C4F9I are in the range 94-97% for n-C4F9-Ar + 1 + 2 (or + 3
+ 4 + 5) and 93-98% for C6H5I, supporting the validity of the
competition kinetics and excluding the presence of significant
amounts of other byproducts. With pyrrole the procedures A,
B, and C were not suitable because the benzoyloxyl radical
adds to the pyrrole ring. In this case method iv (H2O2/DMSO/
Fe(II)) has been utilized in a competition between benzene and
pyrrole. Even with a 10:1 benzene/pyrrole ratio, the attack
to the R-position of the pyrrole ring was higher than 90%.
(C) The reaction was carried out as in C with cyclohexene;
the results are similar to those obtained in A.
Com p etition Kin etics. (A) 1-Octene and the aromatic
substrate (10 mmol on the whole), C4F9I (1 mmol), benzoyl
peroxide (1 mmol), and Cu(OAc)2 (0.1 mmol) in 15 mL of AcOH
were refluxed for 4 h (116 °C). The solution was filtered
through a silica gel column for separating Cu(OAc)2 and
analyzed by GLC and 19F-NMR. The ratios between the
aromatic substrate and 1-octene were 10:1 with benzene and
toluene, 6:1 with anisole, diphenyl ether, thiophene, and
biphenyl, and 2:1 with naphthalene.
(B) As in A using cyclohexene as alkene. The ratios between
the aromatic substrates and cyclohexene were 5:1 for benzene
and toluene, 2:1 for anisole and diphenyl ether, 1:1 for
thiophene and biphenyl, and 1:2 for naphthalene. A typical
example: 5 mmol of cyclohexene and 5 mmol of anisole, 1
J O970302S