Persistent perfluoroalkyl radical investigations under reductive environment: Reaction with electron-donating reagents

Article abstract of DOI:10.1016/S0022-1139(99)00046-9

The reactivity of persistent perfluoroalkyl radical, perfluoro-3-ethyl-2,4-dimethyl-3-pentyl (1), with various electron-donating reagents was investigated. It is revealed that 1 which is robust under oxidative conditions is rather vulnerable under reductive conditions. Thus, Lewis bases such as triethylamine and triphenylpnictogens (Ph₃Pn, Pn=N, P, As, Sb, Bi) and some soft anions such as iodide or tetraphenyl borate reacted with 1 to give perfluoro-3-isopropyl-4-methylpent-2-ene (2) quantitatively. Even very weak Lewis bases such as diethyl ether and diethylsulfide also reacted with 1 to give 2 and additionally a hydrido product, perfluoro-3-ethyl-3-H-2,4-dimethylpentane (4). Hydrogen gas did not react with 1 at all without a catalyst, but in the presence of metal Pd adsorbed on charcoal, smoothly reacted to give 2 in quantitative yield. Metal hydrides such as LiAlH₄, NaBH₄, NaH, BH₃ (THF complex), Bu₃SnH, Me₂PhSiH reacted with 1 to give 2 and 4. That an electron transfer mechanism is operating in the formation of 2 is obvious, but not conclusive in the formation of 4.

Full text of DOI:10.1016/S0022-1139(99)00046-9

Journal of Fluorine Chemistry 97 (1999) 173±182
Persistent per¯uoroalkyl radical investigations under reductive
environment: reaction with electron-donating reagents
Taizo Onoa,*, Haruhiko Fukayaa, Eiji Hayashia, Hiroko Saidaa, Takashi Abea,
Philip B. Hendersonb, Richard E. Fernandezc, Kirby V. Schererd
aFluorine Chemistry Laboratory, Chemistry Department, National Industrial Research Institute of Nagoya, 1-1 Hirate-cho, Kita-ku, Nagoya 462-8510, Japan
bAir Products and Chemicals, 7201 Hamilton Blvd., Allentown, PA 18195-1501, USA
cE.I. du Pont de Nemours & Co., Fayetteville Works, Fayetteville, NC 28306, USA
dE.I. du Pont de Nemours & Co., Experimental Station E302/316, Wilmington, DE 19898, USA
Received 10 December 1998; received in revised form 29 January 1999; accepted 29 January 1999
This article is dedicated to Professor Emeritus Yoshiro Kobayashi on the occasion of his 75th birthday
Abstract
The reactivity of persistent per¯uoroalkyl radical, per¯uoro-3-ethyl-2,4-dimethyl-3-pentyl (1), with various electron-donating reagents
was investigated. It is revealed that 1 which is robust under oxidative conditions is rather vulnerable under reductive conditions. Thus,
Lewis bases such as triethylamine and triphenylpnictogens (Ph3Pn, PnˆN, P, As, Sb, Bi) and some soft anions such as iodide or tetraphenyl
borate reacted with 1 to give per¯uoro-3-isopropyl-4-methylpent-2-ene (2) quantitatively. Even very weak Lewis bases such as diethyl ether
and diethylsul®de also reacted with 1 to give 2 and additionally a hydrido product, per¯uoro-3-ethyl-3-H-2,4-dimethylpentane (4).
Hydrogen gas did not react with 1 at all without a catalyst, but in the presence of metal Pd adsorbed on charcoal, smoothly reacted to give 2
in quantitative yield. Metal hydrides such as LiAlH4, NaBH4, NaH, BH3 (THF complex), Bu3SnH, Me2PhSiH reacted with 1 to give 2 and
4. That an electron transfer mechanism is operating in the formation of 2 is obvious, but not conclusive in the formation of 4. # 1999
Elsevier Science S.A. All rights reserved.
Keywords: Persistent per¯uoroalkyl radical; Reduction; Electron transfer
1. Introduction
reaction of 1 with reducing agents and electron-donating
reagents. The aim of this report is to reveal the nature of
persistent per¯uoroalkyl radical 1 under the reductive envir-
onments.
Persistent per¯uoroalkyl radical 1, per¯uoro-3-ethyl-2,4-
dimethyl-3-pentyl, is known to be amazingly stable. Radical
1 can be distilled and analyzed by gas chromatography
without decomposition and shows no reactivity with conc.
HCl, conc. H2SO4, and oxidizing agents such as oxygen,
chlorine, bromine and iodine [1]. The stability of 1 arises not
only from the inaccessibility of the radical center to the
reagent due to sterical hindrance, but also from the electro-
nic properties endowed by the per¯uoro system. Radical 1 is
prepared by reacting hexa¯uoropropene trimers, per¯uoro-
3-isopropyl-4-methylpent-2-ene (2) and per¯uoro-3-ethyl-
2,4-dimethylpent-2-ene (3), with elemental ¯uorine or in an
electrochemical ¯uorination cell [2]. Other functionalized
analogs of 1 have been prepared under similar conditions [3±
7]. Although the chemistry of 1 in oxidative environments
has been well characterized, little has been reported on the
2. Results and discussion
2.1. Reaction of 1 with hydrogen gas
The experimental conditions investigated are summarized
in Table 1. When a neat liquid containing radical 1 was
stirred under H2 atmosphere at ambient temperature, no
reaction occurred even after seven days (Run 1). The same
reaction at the forcing conditions leads to a formation of
per¯uoro-3-ethyl-3-H-2,4-dimethylpentane 4 in poor yields
(Run 2). However, when a catalytic amount of 5% Pd on
carbon is added to an ethereal solution of 1 (0.01 equivalents
of Pd) a mildly exothermic reaction occurs to cause a gentle
re¯ux of ether and 1 is quantitatively converted to 2 after
*Corresponding author. Fax: +81-52-911-2428; e-mail: ono@nirin.go.jp
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 4 6 - 9

174
T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
Table 1
Reaction of 1 with hydrogen (all unmarked bonds are to fluorines)
Run
Conditionsa
Time
Product distribution (%)b
Recovery (%)b
2
3
4
5
1
1
2
3
4
5
6
H2
H2, 808Cc
H2/cat. 5%Pd/C
Ar/cat. 5%Pd/C
H2/cat. 5%Pd/BaSO4
Ar/cat. 5%Pd/BaSO4
7 days
33 h
±
±
±
±
±
±
±
±
±
100
100
74.2
100
100
100
100
1.6
100
25.4
14.8
49.9
13.2
30 min
4.5 h
±
±
±
±
±
±
±
±
±
74.6
±
30 min
5 h
100
±
100
a Runs 3±6 were run in diethyl ether at room temperature and at atmospheric pressure of H2 or Ar, while Runs 1 and 2 were carried out neat without solvent.
b Product distribution and % of recovery (defined as the sum of yields of all products including radical 1) were calculated from GC data using perfluoro-2,4-
dimethyl-3-ethylpentane as an internal standard.
c The products other than 1±5 obtained are perfluoro-3-ethyl-2,3,4-trimethylpentane (9.9%) and an unknown with the retention time close to the former
(9.6%).
30 min (Run 3). It should be emphasized here that ether
which can dissolve radical 1 plays an important role in the
smooth reaction. If ether is not included, the Pd catalyst
¯oats on the surface of a per¯uoro phase and thus is not well
dispersed into it and easily expelled out on the vessel wall,
resulting in no reaction. The wettability of the Pd catalyst
and solubility of 1 in ether seem to be important factors for
the successful reaction.
tri¯uoromethyl radical [8] can combine with a sterically
shielded radical center of 1 under hydrogen gas atmosphere
to give such a yield comparable with the one of the hydrido
product 4. This result suggests the access of much smaller
hydrogen molecule to the radical center of 1 is highly
probable. In this regard, the inertness of radical 1 could
be interpreted as being not only from the sterical hindrance
but also from the nature of the per¯uoro system.
Surprisingly, even under Ar atmosphere, the conversion
of 1 to 2 still occurred when the same 5% Pd on carbon was
added to an ethereal solution of 1, although at a much slower
rate than when hydrogen was present (Run 4). Substituting
BaSO4 for carbon as the catalyst support resulted in no
reaction over argon (Run 6), but the same rapid reaction was
observed again in the presence of hydrogen to quantitatively
prepare 2 (Run 5). These results suggest that the Pd catalyst
can transfer an electron to 1 forming the anion which
eliminates a ¯uoride ion to form 2. The Pd catalyst is then
rapidly reduced by H2 or slowly reduced by the carbon
support. When the inert BaSO4 is used as the support, the
electrons lost from the Pd are not replenished and no
signi®cant reaction is observed.
Alkane 4 is formed in ca. 11% yield (calcd; recovery
Âproduct distribution/100 in Table 1) when 1 is heated to
808C in the presence of H2 for 33 h. Competing with the
addition of hydrogen are (1) the ®rst order decomposition of
1 by beta-scission to form a CF3 radical and E- and Z-
per¯uoro-3-ethyl-4-methylpent-2-ene (5) [1] and (2) radical
combination reaction between CF3 and 1 which gives per-
¯uro-3-ethyl-2,3,4-trimethylpentane (6) in the yield of 9.9%
comparable to the one of 4. It is not known if 4 is formed
from 1 by direct reaction with H2 or by reaction with
hydrogen atoms formed from the reaction between CF3
and H2 (Scheme 1). Regardless of the reaction mechanism,
it is rather amazing to know that such a sterically demanding
Exclusive formation of kinetically favored 2 over ther-
modynamically favored 3 in the electron transfer reaction is
understood by the availability of ¯uorine atoms on b-
carbons to be anti-periplanar to the lone pair electrons of
the intermediate anion. A space-®lling model of the anion
can be constructed only when the two per¯uoroisopropyl
groups are conformationally arranged in the manner
depicted in Scheme 2. Both 38-¯uorines on the per¯uoro-
isopropyl groups are orthogonal to the orbital of the lone
pair electrons, while one of the two 28-¯uorine atoms on the
per¯uoroethyl group is in the preferred anti-periplanar
position to leave as a ¯uoride ion to form 2. Alkene 2 is
subject to reversible nucleophilic attack by ¯uoride, and
when pure 2 is stirred with a catalytic amount of ¯uoride for
several hours at room temperature, it is isomerized to the
Scheme 1. All unmarked bonds are to fluorines.

T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
175
0.2 ml of H2O and 1 ml of acetone) with 1 (0.5 mmol) gave
2 almost quantitatively in 10 minutes (97.6% yield, Run 1),
but no reaction occurred with aqueous KBr/acetone (Run 2).
Radical 1 is reduced by organic bases to form 2 with loss
of a ¯uoride ion. Reaction rates tend to increase with
increasing base strength. An immediate exothermic reaction
occurred when 1 was mixed with triethylamine at room
temperature, giving 2 quantitatively (Run 3). A much more
vigorous reaction was observed between 1 and the electron-
rich tetrakis(dimethylamino)ethene [12] which resulted in
the formation of tars and a low yield of 2 and 3 in a 1:2 ratio
(Run 4). Although 2 may have been formed initially in Run
4, the ¯uoride ion and great amount of heat produced from
the reduction of 1 isomerized 2 to 3. The soft anion
tetraphenyl borate also reacted with 1 to give 2 exclusively,
but at a slower rate than the amines (Run 5). Triphenylamine
reacted smoothly with 1 over 10 min to give 2 quantitatively
(Run 6). Formation of the intermediate cation radical of
triphenylaminium by an electron transfer process is obvious
from the sudden appearance of a deep blue color by the
addition of 1 into the ethereal solution of triphenylamine at
room temperature [13]. The deep color is soon followed by a
reddish brown color, suggesting a dimerized cation radical
species [14]. It is noteworthy to point out that pyridine did
not react with 1 after 24 h stirring at room temperature even
though it is more basic and nucleophilic than triphenyl-
amine.
Scheme 2. Schematic drawing of the intermediate anion.
thermodynamic equilibrium mixture of approximately 33%
of 2 and 67% of 3 [9].
2.2. Reaction of 1 with electron-donating reagents
We next examined a series of electron-donating reagents
(Table 2). Alkyl radical 1 in acetonitrile is reported to have a
non-reversible reduction potential between 0.55 and
0.575 V by cyclic voltammetry [10]. Thus, iodide with a
redox half reaction potential of 0.535 V should reduce 1
while bromide with a potential of 1.06 V should not [11].
Mixing an aqueous KI/acetone solution (1 mmol KI in
In contrast with triphenylamine, only a faint yellow
coloration occurred with triphenylphosphine probably due
to a CT band (vide infra), but no deep coloration occurred.
Another peculiar difference is the formation of white pre-
cipitation which suddenly appeared after a few minutes of
Table 2
Reaction of 1 with various electron-donating reagents
Ether
!
1 ‡
2 ‡ 3 ‡ 4
Reagents RT
Run
Reagentsa
Time
Product distribution (%)
Recovery (%)b
2
3
4
1
1
2
Aq. KI/acetone
Aq. KBr/acetone
Et3N
10 min
20 h
97.6
±
2.4
±
±
±
±
±
±
±
±
±
±
±
±
100
100
100
44.7
93.3
100
100
100
99.9
99.8
95.1
97.2
85.9
100
100
±
3
<1 min
<1 min
3 h
100
21.9
91.8
100
100
24.5
43.8
19.6
37.1
46.2
87.1
±
±
4
[(Me2N)2C=]2
‡
46.8
±
±
5
Ph4B Na
8.2
6
Ph3N
Ph3P
Ph3As
Ph3Sb
Ph3Bi
Et2O
THF
10 min
30 min
3 h
±
±
7
±
±
8
±
75.5
56.2
80.3
34.0
8.1
9.0
9
3 h
±
10
11
12
13
14
3 h
41 h
±
±
28.9
45.7
3.9
±
45 h
±
Et2S
44 h
±
DMF
48 h
±
100
a The molar ratio of the reagent to 1 is one except aq. KI/acetone and aq. KBr/acetone where the ratio is two. A large excess is used for Runs 11±14.
b Recoveries (defined as the sum of yields of all products including radical 1) are calculated from GC data using perfluoro-3-ethyl-2,4-dimethylpentane as an
internal standard.

176
T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
stirring. Although the rate of the reaction is slower, 2 is also
formed quantitatively from 1 and triphenylphosphine over
30 min (Run 7). Other triphenylpnictogens Ph3Pn (PnˆAs,
Sb, Bi) also reacted with 1 in a manner similar to Ph3P, but
with much slower rates (Runs 8±10).
phosphine was observed is yet not known. In contrast with
the triphenylphosphine, Ph3AsF2 and Ph3BiF2 were
obtained as sole products in the reaction of 1 with Ph3As
and Ph3Bi. Why most of the initial Ph3As and Ph3Bi (60±
70%) remained intact in both reactions is again not known.
The reaction of 1 with Ph3Sb gave, on the other hand, a very
complex reaction mixture.
The oxidation products of triphenylpnictogens were not
intensively studied, but it is worth commenting on these to
some extent. The main product isolated from the reaction
mixture of triphenylamine with 1 is N,N,N0,N0-tetraphenyl-
benzidine (ca. 20% yield), of which formation is well known
in both processes of electrochemical [15] and chemical [16]
oxidation of triphenylamine. Accompanied with the main
product is a mono-¯uoro derivative of N,N,N0,N0-tetraphe-
nylbenzidine, the structure of which was estimated as 7
based on the spectroscopic data (MS, 19F-NMR, 1H-NMR,
We noticed that a pale yellow color appeared when 1 was
added to diethyl ether or THF. If the ethereal solution (1 is
soluble in ether) is kept at room temperature for a prolonged
time (41 h), 68% of 1 is converted to 2 and 4 (Run 11). When
the heterogeneous mixture of 1 and THF was vigorously
stirred at room temperature over 45 h, it followed almost the
same as the above, but with a little faster reaction rate than in
diethyl ether. The end of the reaction was easily monitored
with discoloration. Diethylsul®de did not dissolve 1, but a
purple color developed in the upper sul®de layer when
mixed. After the heterogeneous mixture was vigorously
stirred for 44 h at room temperature, the color of the upper
layer became faint. GC analysis of the lower layer showed
that ca. 90% of 1 was consumed and 2 (87.1%) was formed
as a main product in addition to a small amount of by-
product 4 (Run 13). Interestingly, dimethyformamide
(DMF) did not react with 1 at all after 48 h stirring of its
heterogeneous mixture (Run 14).
Burdeniuc et al. recently reported that a hexane solution
of triethylamine and a cyclic per¯uoroalkane such as per-
¯uorodecalin shows a CT band at 270 nm with no obser-
vable changes in both reagents under the conditions for the
UV measurement [20]. The coloration by the dissolution of
1 in ether or by mixing of 1 with sul®de is very likely to be
caused by a CT complex formation between 1 and these
ethers because the color disappeared after 1 was consumed.
The relative reaction rates (N>P>Sb>As,Bi) observed for
the triphenylpnictogens series is rather surprising in the light
of literature reports of the relative basicity (As>N>Bi,Sb>P)
[21] and rates of single electron oxidation (P>As,Sb>N>Bi)
[22]. Neither order is consistent with the observed relative
1
13Cf1Hg- and 13Cf H; 19 Fg-NMR, IR). We tentatively
assigned the structure 7 by considering the reaction mechan-
ism depicted in Scheme 3. MS of 7 isolated by HPLC (see
Section 3) shows a molecular ion m/z 506 as a parent ion
with right intensities of 13C-isotopic peaks M‡1 and M‡2
[m/z 507 (40.9) and m/z 508 (7.8), respectively]. The peak
found at m/z 253 (74.9) was M2‡ molecular ion similar to
M2‡ at m/z 244 (ca. 50) reported for N,N,N0,N0-tetraphe-
nylbenzidine [17]. The fragment ions of both half of the
molecule with the compositions of C18H13FN (Ph2NC6H3F)
and C18H14N (Ph2NC6H4) were found at m/z 262 (1.4) and
244 (4.2), respectively.
A large recovery of triphenylamine (ca. 50%) is consis-
tent with the observation of the formation of more easily
oxidizable N,N,N0,N0-tetraphenylbenzidine [18]. Half of the
initial triphenylphosphine was recovered unchanged also in
the reaction of 1 with triphenylphosphine and the remaining
half was recovered as triphenylphosphine oxide after work
up, which is probably formed with reaction of moisture. No
signals corresponding to the dimerized products and M±F
fragment ion of Ph3PF2 [19] were observed in MS of the
crude reaction product measured by a direct inlet (DI) mode.
The reason why only a half consumption of the triphenyl-
Scheme 3. Mechanism of the formation of dimers 7 and 8 in the reaction of triphenylamine with persistent perfluoroalkyl radical 1.

T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
177
rates. However, considering the difference of the reaction
products in the pnictogen series used, it is realized that it is
dif®cult to make a simple comparison among all pnictogen
series. The unexpectedly fast relative reaction rate observed
for the Ph3N could be explained by taking into account the
formation of the dimer which has an electrochemical oxida-
tion rate known to be faster than the starting triphenylamine
[23]. It would be desirable to use the tri-p-tolylpnictogen
series for the kinetic study because tri-p-tolylamine is
known to be resistant toward dimerization [24,25]. This
will be the subject of a future investigation.
raised to room temperature and the sample was stirred
continually for 2 h, then quenched by H2O and 15%
aq. NaOH. GC analysis showed that most of
1
(93Â0.715ˆ66.5%) remained unchanged (Run 6). The
same reaction was conducted at room temperature through-
out the reaction, but the same result was obtained (Run 7).
The reaction was conducted in a concentration (0.1 M in
THF) much less than the solubility of LiAlH4 in THF (1.0 M
solution is commercially available); thus the reaction sys-
tem is almost homogeneous. However, a vigorous efferves-
cent reaction occurred in the quenching process in both
cases, suggesting that only a part of hydrogen among the
four hydrogens of LiAlH4 is active for this reaction and the
rest are in a dormant state until being quenched. When the
molar ratio of LiAlH4 was increased to 1, ca. 26% of 1 still
remained unchanged. While doing this reaction, we noticed
that THF is a rather poor solvent for 1 and thus a per¯uoro
phase existed not only at low temperature but also at room
temperature. We roughly estimated the solubility of the
stock radical solution by adding it into THF or diethyl ether
with vigorous stirring at room temperature until the ®rst tiny
per¯uoro drop appeared. Solubilities thus obtained were
only 17 mg in 1 ml of THF and 262 mg in 1 ml of diethyl
ether. This surprising solubility difference prompted us to
repeat the same reduction by using diethyl ether as solvent.
Again, about the same amount of 1 (ca 20%) remained
unreactive (Run 9). It is rather surprising to know such a
radical system even susceptible to the solvent like ether can
survive through contact with a very strong reducing agent
like LiAlH4. The unexpected high recovery of 1 in the
reaction of 1 with powerful reducing agents such as LiAlH4,
NaBH4 (Run 10) and NaH (Run 11), even at room tem-
2.3. Reaction of 1 with various metal hydrides
Tin and silicon hydride reducing agents which readily
undergo hydrogen abstraction [26±28] (Runs 1±3) gave a
mixture of 4 and 2 from 1. With tributyltin hydride at room
temperature 4 was obtained as a main product in addition to
2 in a combined yield of 78%. Dimethylphenylsilane
reacted similarly to tributyltin hydride, while diphenylsilane
gave 2 and 4 in a 1: 1 ratio. There is no doubt that an electron
transfer (ET) mechanism worked for the formation of 2,
but it is beyond the scope of this paper to refer to the
mechanism for the formation of 4. Both tetrabutyltin and
tetramethyltin did not react with 1 at all (Runs 4 and 5),
suggesting the Sn±H bond is responsible for the ET process
of the reaction.
Since vigorous reaction was expected, the reaction of 1
with LiAlH4 was ®rst conducted in dry-ice acetone bath by
using a molar ratio of LiAlH4:1ˆ0.25:1. However, GC
analysis of the lower per¯uoro phase of the 20 min sample
showed no reaction at all. Therefore, the temperature was
Table 3
Reaction of 1 with various metal hydrides
Solvent
!
1 ‡
2 ‡ 3 ‡ 4
Reagents RT
Run
Reagents
Molar ratioa
Solvent
Time (h)
Product distribution (%)
Recovery (%)c
2
3
4
1
Othersb
1
2
Bu3SnH
Me2PhSiH
Ph2SiH2
Bu4Sn
1
Neat
Et2O
Et2O
Neat
Neat
THF
THF
THF
Et2O
Et2O
Et2O
THF
2.0
37.0
40.4
47.3
±
±
63.0
59.6
44.9
±
±
±
±
78.2
81.5
84.7
100
1
4.0
±
±
3
1
3.5
2.3
±
5.5
±
4
1
7 days
5 days
2.0
100
100
±
5
Me4Sn
1
±
±
±
±
100
d
6
LiAlH4
0.25
0.25
18.6
16.3
54.0
21.2
9.4
±
9.9
5.5
7.2
50.9
6.3
±
71.5
±
93.0
98.0
96.0
94.5
98.7
100
7
LiAlH4
LiAlH4
LiAlH4
NaBH4
NaHd
1.0
±
76.3
26.9
20.4
83.8
86.6
±
1.9
10.6
6.4
0.5
±
8
1
1
1
1
1
1.0
1.3
1.1
±
9
1.0
10
11
12
1.0
4.0
13.4
57.0
±
BH3/THF
4.0
±
43.0
±
100
a The molar ratio to the reagent of 1.
b Not investigated further.
c Recoveries (defined as the sum of yields of all products including radical 1) are calculated from GC data using perfluoro-3-ethyl-2,4-dimethylpentane as an
internal standard.
d The reaction was conducted at 788C at first, then warmed to room temperature and continued for the time described in the table.

178
T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
perature, was probably caused by passivation due to the
formation of a passive coat on the surface of the suspended
reagents. Both, that vigorous effervescence occurred in the
quenching process of the alkali metal hydrides and that a
BH3/THF adduct soluble in THF completely reacted with 1
in the homogeneous system to give a mixture of the products
2 and 4 (Run 9), are consistent with this explanation of
passivation.
There are two things which should be mentioned on the
product distribution. One is a remarkable solvent effect
found in reduction with LiAlH4. Thus, the ole®n 2 is the
main product (54.0%) in the case of THF (Run 8); in
contrast, the hydrido product 4 is main (50.9%) in the case
of diethyl ether (Run 9). The other is the selective formation
of 2 in the reduction of 1 with NaH.
In summary, it was found that the persistent per¯uoro-
alkyl radical 1, which is very stable under oxidative con-
ditions, reacts with various kinds of electron-donating
reagents to form 2 exclusively through an ET process, or
additionally a hydrido product 4 depending upon the
reagents used. Thus, alkyl amines reacted rapidly with 1;
however, for the series of triphenylpnictogens the reaction
rates did not correlate with the reported basicities or rates of
single electron oxidations due to the divergence of the
reaction paths observed among the pnictogens used. Diethyl
ether, the weakest electron-donating reagent among the
reagents surveyed was found to give a coloration when
mixed with 1 probably due to the formation of the charge
transfer complex and was found to be reactive with 1.
Monohydrido alkane 4 can be formed from 1 in modest
yields with tin and silicon hydrides; however, direct reaction
between 1 and H2 to form 4 is observed in poor yields under
forcing conditions and when the reaction between 1 and H2
is conducted in the presence of Pd catalyst, quantitative
formation of 2 through an ET process is observed.
and negative values for up®eld shifts. Chemical shifts of 1H-
and 13C-NMR are reported on the ꢀ scale, with TMS as an
internal standard. Triphenyl phosphate was used as an
internal standard for 31P-NMR. A glass column (3 mm
i.d.Â5 m) packed with 12.1% Fomblin (YH/VAC 40/11)
on Chromosorb P (AW-DMCS, 100±120 mesh) was used for
GC analysis. Per¯uoro-3-ethyl-2,4-dimethylpentane was
used as an internal standard for GC calculation. Preparative
HPLC (Shimadzu LC-6A) was performed with a Chroma-
torex B-5 column (Fuji Silysia Chemical, 5 mm SiO2,
10 mm i.d.Â250 mm) eluted with 2.5% CH2Cl2 in hexane
at 2.0 ml/min, using a Shimadzu UV detector (SPD-6A)
operating at 254 nm. TLC (0.25 mm thick Silica gel 60 F254
Merck) was used.
,
A mixture of hexa¯uoropropene trimers 2 and 3 was
prepared by the reported method [29]. All chemicals were
reagent grade and used as received. Pd on charcoal (5%) was
purchased from Kawaken Fine Chemicals. Diethyl ether and
THF were distilled over sodium/benzophenone ketyl just
before use. DMF was dried over CaH2. The liquid phase
used for GC analyses, Fomblin (YH/VAC 40/11), was a gift
from Ausimont K.K.
3.2. Preparation of the stock solution of 1
A hexa¯uoropropene trimer mixture (186.7 g) consisting
of 2 (82.0%) and 3 (18.0%) and a Te¯on-coated magnetic
stirring bar were placed in
a glass tube (24 mm
i.d.Â400 mm). Undiluted ¯uorine gas was introduced into
the bottom of the tube at a rate of 4.2 ml/min at room
temperature (ca. 278C) with vigorous stirring. After 34.5 h
¯uorination, all of 2 and 3 were consumed to give a radical
solution (185.5 g), of which GC analysis showed it contains
86% of 1 as a main product and 14% of F2 adduct, per¯uoro-
3-ethyl-2,4-dimethylpentane, as a ``by-product'' (82%
yield!). The content of 1 in this stock solution did not
change for three years when stored in a refrigerator. At
room temperature, content of 1 gradually decreased to 85%
of the initial level after six months.
3. Experimental details
3.1. Measurements and materials
3.3. Reaction of 1 with hydrogen gas
Infrared spectra were measured in a KBr disk method
with a Shimadzu FTIR-8000. Mass spectra (EI, 70 eV) were
measured on a Shimadzu QP-5000 quadrupole mass spec-
trometer using a capillary column (60 mÂ0.25 mm i.d.,
1.5 mm thick NEUTRA BOND-1, GL Sciences) for volatile
compounds and using a DI mode for non-volatile com-
pounds. Negative chemical ionization (NCI) mass spectra
(70 eV) were recorded on a Perkin-Elmer 281 spectrometer
with HP accessory 18962A (NCI packages). Argon gas was
used as the reagent gas. 19F-, 13C-, 1H- and 31P-NMR spectra
were measured with a Varian FT-NMR spectrometer
(UNITY INOVA-300) operated at 282.01, 75.423, 299.95
and 121.42 MHz, respectively. Deuterated chloroform was
used as an NMR solvent. Chemical shifts of 19F-NMR are
reported on the ꢀ scale with CFCl3 as an internal standard
3.3.1. Reaction in the presence of Pd catalyst
Catalytic amount of 5% Pd on charcoal or on BaSO4
(35 mg) was placed in a 20 ml two-necked ¯ask and dried
by a heat gun under reduced pressure. After ®lling the
reaction vessel with hydrogen gas, dried ether (2 ml) and
the stock radical solution obtained above (1.09 g, 2 mmol 1)
were introduced by a syringe with vigorous stirring. A pale
yellow color developed and an exothermic reaction occurred
to re¯ux. The re¯ux almost subsided after 15 min, and the
GC analysis of the reaction mixture showed all the radical
was quantitatively converted to 2 after 30 min stirring.
When the same reaction using 5% Pd on charcoal was
conducted without ether solvent, no reaction occurred after
24 h stirring at room temperature.

T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
179
3.3.2. Thermal reaction in the absence of Pd catalyst
3.4.2. Reaction with triethylamine and
tetrakis(dimethylamino)ethene
1 g of the radical solution (86% content) and a Te¯on-
coated magnetic stirring bar were placed in a two-necked
¯ask ®tted with a Dimroth condenser and a rubber septum.
The air was replaced with H2 by three cycles of freeze-and-
thaw. Hydrogen gas was supplied by a hydrogen balloon
connected at the end of the condenser. The ¯ask was
immersed into an oil bath maintained at 838C and the
contents were vigorously stirred. The reaction mixture
was periodically sampled by a syringe through a septum
and was monitored by GC. Product distributions were
Into an ethereal radical 1 solution (1.09 g, 2 mmol 1 in
5 ml of dry ether) was added triethylamine with vigorous
stirring. An exothermic reaction occurred immediately after
addition to give a brown colored solution. The accompany-
ing viscous brown material was not characterizable. The
ethereal phase was analyzed by GC. The same kind of
reaction was conducted by using tetrakis(dimethylami-
no)ethene (400 mg, 2 mmol). A vigorous exothermic reac-
tion occurred with fuming. A blackish dense color appeared
immediately after the addition, but the color soon disap-
peared with the formation of orange-colored tarry material
on the vessel of which characterization was not tried. The
ethereal phase was analyzed by GC. Results of the GC
analyses are summarized in Table 2.
calculated
by
per¯uoro-3-ethyl-2,4-dimethylpentane
included in the stock radical solution as an internal standard.
Relative sensitivity of the products and the internal standard
toward TCD are known to be approximately the same [2],
and calculated accordingly. After 33 h (about 3 half-lives of
1), the reaction mixture consisting of compounds 1±5 [30],
one new compound per¯uoro-3-ethyl-2,3,4-trimethylpen-
tane (6) and one unknown was obtained. Spectroscopic
data for 6 were the following. ꢀF: 54.32 (3F, brs),
65.96 (12F, brs), 75.09 (3F, brs), 93.29 (2F, m),
157.58 (2F, m); MS: m/z (%), 331 (2.6), 281 (8.6), 262
(1.6), 243 (4.4), 231 (8.2), 193 (3.5), 181 (8.9), 169 (2.7),
151 (3.9), 143 (3.2), 124 (1.6), 119 (6.2), 100 (3.4), 93 (5.8),
82 (5.4), 74 (1.6), 73 (3.5), 69 (100); NCI±MS: m/z (%), 419
(6.4), 370 (7.8), 369 (100), 350 (6.1). Compound 4 is known,
but its spectroscopic data is not available in literature and
thus included here [2]. ꢀF: 71.1 (6F, bs), 75.7 (6F, bs),
84.7 (3F, t, Jˆ18.6 Hz), 107.8 (2F, bs), 167.4 (2F, bs);
3.4.3. Reaction with sodium tetraphenylborate
Into a suspension of sodium tetraphenylborate (684 mg,
2 mmol) in 5 ml of dry ether was added the stock radical
solution (1.09 g, 2 mmol 1) with vigorous stirring. No
visible change except the usual pale yellow coloration
occurred. GC analysis showed most of 1 was consumed
after 3 h. Data are summarized in Table 2.
3.4.4. Reaction with triphenylamine
Into an ethereal solution of triphenylamine (490.6 mg,
2 mmol in 5 ml ether) was added the stock radical solution
(1.09 g, 2 mmol 1) with vigorous stirring over 1 min. Only a
drop of 1 gave a blue color to the solution, and soon turned
to reddish purple by further addition, then green for a short
period of time, and ®nally a very dense persistent blue color.
GC analysis of the 10 min sample showed that 1 was
quantitatively converted to 2. Ether was removed by eva-
poration and the remaining brown viscous material
(420 mg) was subjected to column chromatography
(SiO2, CH2Cl2:Hexˆ1:3). Triphenylamine was recovered
(Rfˆ0.54, 244 mg, 50%) and a mixture of N,N,N0,N0-tetra-
phenyl-2-¯uorobenzidine 7 and N,N,N0,N0-tetraphenylben-
‡
ꢀH: 4.63 (1H, t, Jˆ11.9 Hz); MS: m/z 451 (M , 0.2), 313
((CF3)2C=CH±C(CF3)2, 3.4), 263 (CF3CF=CH±C(CF3)2,
5.5), 213 (CF2=CH±C(CF3)2, 7.0), 169 (9.6), 163
(CF2=CH±CFCF3, 5.4), 119 (43.9), 113 (CF2=CH±CF2,
4.7), 100 (4.5), 69 (100).
3.4. Reaction of 1 with electron-donating reagents
Unless otherwise stated all reactions of Sections 3.4 and
3.5 were carried out at room temperature under Ar atmo-
sphere.
zidine
8
were obtained (133 mg, Rfˆ0.37). Each
component was isolated by preparative HPLC using a
Chromatorex B-5 column in the yields of 7.1% and
19.9%, respectively. Spectroscopic data for 7 were the
following. ꢀF: 117.07 (dd, Jˆ16.5, 10.4 Hz); ꢀH: 7.40
(2H, dd, Jˆ8.4, 1.5 Hz), 7.24ꢀ7.32 (9H, overlapped),
7.12ꢀ7.16 (8H, overlapped), 7.0ꢀ7.1 (6H, overlapped),
3.4.1. Reaction with aqueous KI and KBr
An aqueous KI acetone solution was prepared by dissol-
ving KI (166 mg, 1 mmol) in 200 ml of water, then adding
1 ml of acetone. Into this aqueous KI acetone solution was
added the stock radical solution (273 mg, 0.5 mmol 1) with
vigorous stirring at room temperature. Brown color devel-
oped immediately after the addition of the radical due to I2
formation. After 10 min stirring the bottom layer of the per-
¯uoro phase was analyzed by GC. An aqueous KBr acetone
solutionwaspreparedbydissolvingKBr(119 mg,1 mmol)in
400 mlofwater, thenadding1 mlofacetone.Intothisaqueous
KBr acetone solution was added the same amount of the stock
radical solution as the above. GC analysis of the per¯uoro
phase after 20 h showed no reaction. Results of both analyses
are summarized in Table 2.
1
6.78ꢀ6.90 (2H, overlapped); 13Cf Hg ꢀC: 160.22 (d,
JCFˆ247 Hz), 148.30 (d, JCFˆ10.1 Hz), 147.72, 147.20,
146.94, 130.40 (d, JCFˆ5.5 Hz), 129.68 (d, JCFˆ11.2 Hz),
129.44, 129.39 (d, JCFˆ3.2 Hz), 129.28, 125.03, 124.54,
124.35 (d, JCFˆ72.5 Hz), 123.66, 123.43, 122.97, 118.39
(d, JCFˆ2.9 Hz), 110.00 (d, JCFˆ26.2); MS: m/z (%), 508
(7.8), 507 (40.9), 506 (M, 100), 338 (M-NPh2, 2.3), 262 (M-
C6H4NPh2, 1.4), 253 (M2, 74.9), 244 (M-C6H3FNPh2, 4.2),
214 (9.2), 168 (Ph2N, 7.3), 77 (Ph, 8.5); IR (KBr, ꢁmax
cm 1) 1587, 1489, 1275, 818, 752.

180
T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
3.4.5. Reaction with triphenylphosphine
3.4.6.3. Reaction with triphenylbismuth. A pale brownish
color developed in this reaction. After 35 h stirring at room
temperature, 1 was completely converted to 2 (87.2%) and 4
(12.8%). 19F-NMR (CDCl3) of the white powder obtained in
the same manner as the above showed only one broad signal
Into an ethereal solution of triphenylphosphine
(524 mg, 2 mmol in 5 ml ether) was added the stock radical
solution (1.09 g, 2 mmol 1) with vigorous stirring over
1 min. A pale yellow colored solution was obtained. A
white precipitate suddenly appeared 1.5 min after the
completion of the addition. When the color disappeared
after 30 min stirring at room temperature, the reaction
mixture was analyzed by GC using a FOMBLIN column.
The GC showed that 1 was completely converted to 2. The
white precipitate (240 mg) was collected by ®ltration,
dried, and analyzed by DI-MS. The MS was consistent
with the one of the authentic sample of triphenylphosphine
oxide. 31P-NMR showed the signals of triphenyl phosphine
oxide and one unknown at 33.8 and 49.1 ppm with the
integrations of 13:1. The unknown signal was quartet
with a coupling constant of 7.0 Hz. The ®ltrate was con-
densed by evaporation and dried in vacuo. The white powder
(284 mg) obtained was analyzed by 31P-NMR and found to
comprise 84% of the initial triphenyl phosphine and 16% of
triphenyl phosphine oxide. Thus, the mass balance roughly
calculated on the basis of 31P-NMR are triphenyl phosphine
(46%), triphenyl phosphine oxide (48%) and an unknown
(6%).
at
161.0 ppm, consistent with the reported data
1
( 158 ppm) of Ph3BiF2. H-NMR showed that this white
powder comprised the starting Ph3Bi and the fluorinated
product Ph3BiF2 with a ratio of 66:34. No further effort to
isolate the product was done.
3.4.7. Reaction of 1 with solvents
This reaction was conducted as a blank test. Into 5 ml of
dry diethyl ether was added the stock radical solution
(1.09 g, 2 mmol 1) with vigorous stirring. The pale yellow-
ish solution obtained was continually stirred. Reactions with
other solvents (THF, diethyl sul®de and DMF) were carried
out in a heterogeneous state due to the mutual immiscibility.
In the case of diethyl sul®de, purple color developed in the
upper diethyl sul®de layer while a tea-like brown color
appeared in the lower layer. No coloration occurred in the
DMF case. The reactions were followed by GC for 3 days
with ca. 24 h intervals. The results on the 2-day samples are
summarized in Table 2.
3.4.6. Reaction with other triphenylpnictogens
The reaction was done in the same manner as the above.
After 3 h stirring at room temperature the volatile products
were analyzed by GC. The results are summarized in
Table 2.
3.5. Reaction of 1 with metal hydrides
3.5.1. Reaction with some organotin derivatives and some
organosilane derivatives
3.5.1.1. Reaction with tributyltin hydride and tetra-
alkyltins. Into the stock radical solution of 1 (1.09 g,
2 mmol 1) was added tributyltin hydride (600 mg, 554 ml,
2 mmol, purity 97%) by syringe with vigorous stirring. An
exothermic reaction occurred immediately after addition of
the hydride to give a white precipitate. Into the slurry
obtained after 2 h stirring at room temperature was added
5 ml of ether and filtered through a glass wool. The obtained
colorless solution was analyzed by GC. Both reactions of 1
(1.09 g, 2 mmol 1) with tetrabutyltin (694 mg, 2 mmol) and
with tetramethyltin (357.7 mg, 2 mmol) were conducted in
the same manner. Both tetraalkyltins are immiscible with 1,
thus the reactions were conducted in a heterogeneous state
with vigorous stirring at room temperature. The upper
tetraalkytin layers were very slightly colored from the
beginning (brownish in the case of tetrabutyltin and
yellowish in the case of tetramethyltin). Lower perfluoro
phases were analyzed by GC over 5±7 days. The GC results
are summarized in Table 3.
3.4.6.1. Reaction with triphenylarsine. Colorless crystals
appeared from the tea-colored reaction mixture after 21 h
vigorous stirring. GC analysis showed that 1 was completely
converted to 2 (96.9%) and 4 (3.1%). Colorless crystals
(162 mg) were found to be difluorotriphenylarsine (23.5%
yield). A large amount of triphenylarsine was recovered
intact (467 mg, 76.3%). Spectroscopic data of the colorless
crystal were the following. ꢀF: 87.8 ppm (s) (reported data
86.5) [31]; ꢀH: 7.40±7.64 (9H, m), 8.08±8.22 (6H, m); MS
(DI) of a colorless crystal m/z: 344 (Ph3AsF2, 0.6), 325
(M±F, 3.1), 306 (M±2F, 2.4), 249 (5.9), 248 (M±PhF, 42.4),
227 (7.3), 171 (11.5), 155 (13.6), 154 (Ph±Ph, 100), 153
(9.1), 152 (PhAs, 14.8), 151 (8.8), 77 (30.0), 51 (34.2), 50
(9.6).
3.4.6.2. Reaction with triphenylantimony. A brownish
yellow color developed in this reaction. After 18 h
stirring at room temperature, 1 was completely converted
to 2 (97.3%) and 4 (2.7%). Volatile compounds and the
solvent were removed by evaporation, and the remaining
white powder was examined by 19F-NMR (CDCl3). Many
signals were found at 65.05 (s), 74.99 (d, Jˆ6.2 Hz),
76.61 (s), 83.51 (s), 87.80 (s), 154.33 (s) with the
integration of 3:6:12:26:37:15. No further characterization
was carried out.
3.5.1.2. Reaction with dimethylphenylsilane and dip-
henylsilane. Into an ethereal solution of radical 1 (1.09 g,
2 mmol
1
in 5 ml of dry ether) was added
dimethylphenylsilane (272.5 mg, 2 mmol) with vigorous
stirring. The yellow color became faint with time and
disappeared after 1 h. The reaction with diphenylsilane

T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
181
(380.0 mg, 2 mmol, purity 97%) was conducted in the same
manner. The ether layer was analyzed by GC. GC results are
summarized in Table 3.
dry-ice acetone bath was added the radical stock solution
(1.09 g, 2 mmol) with vigorous stirring over 2 min. After
completion of the addition, the bath was removed. After
30 min and 4 h stirring at room temperature, the ethereal
layer was sampled by a syringe and analyzed by GC. Both
30 min and 4 h GC data were totally the same and are quoted
as 4 h data in Table 3. When methanol was added into the
4 h reaction mixture, a vigorous effervescent reaction
occurred.
3.5.2. Reaction with alkali metal hydrides
3.5.2.1. Reaction with lithium aluminum hydride. Reaction
of 1 with lithium aluminum hydride was conducted in four
reaction conditions (a)±(d) depending on temperature,
molar ratio of reagents, and solvent: (a) 788C at first,
then at room temperature, 0.25, THF; (b) room temperature,
0.25, THF; (c) room temperature, 1, THF; (d) room
temperature, 1, diethyl ether, respectively.
(a) Into a suspended solution of LiAlH4 (57 mg,
1.5 mmol) in 5 ml of dry THF cooled in dry-ice acetone
bath was added the stock radical solution (3.27 g, 6 mmol 1)
over 5 min with vigorous stirring. Per¯uoro phase iced out
and the stirring was very erratic. The bath was removed after
10 min and continually stirred for 2 h until quenched with
H2O (60 ml), 15% aq. NaOH (60 ml), and H2O (100 ml). GC
analysis of the lower per¯uoro phase showed that volatile
compounds are only 1, 2, 4 with recovery of 93.0% calcu-
lated by using internal standard (per¯uoro-3-ethyl-2,4-
dimethylpentane). The distribution among 1, 2, and 4 is
shown in Table 3.
3.5.2.4. Reaction with borane±THF complex. Into the
radical stock solution (2.73 g, 5 mmol 1) was added 5 ml
of 1 M borane±THF solution with vigorous stirring at room
temperature. A slightly exothermic reaction occurred to give
a heterogeneous two-layer mixture. The upper THF layer
was slightly yellow. GC analysis of the lower perfluoro
phase was conducted. After 2 h stirring 77% of 1 was
converted to 2 (42%) and 4 (58%). After 4 h stirring 1
was completely consumed (see Run 12 in Table 3).
References
[1] K.V. Scherer, T. Ono, K. Yamanouchi, R. Fernandez, P. Henderson, J.
Am. Chem. Soc. 107 (1985) 718.
(b) Into a solution of LiAlH4 (19 mg, 0.5 mmol) in 5 ml
of dry THF was added the radical stock solution (1.09 g,
2 mmol 1) with vigorous stirring at room temperature.
Slightly exothermic reaction occurred to give a gray colored
suspension. After 60 min stirring at room temperature, 20 ml
of H2O, 20 ml of 15% aq. NaOH, and 40 ml of H2O were
added in that order. Per¯uoro phase separated at the
bottom of the vessel was analyzed by GC (see Run 7 in
Table 3).
[2] T. Ono, H. Fukaya, M. Nishida, N. Terasawa, T. Abe, J. Chem. Soc.,
Chem. Commun. (1996) 1579.
[3] V.F. Cherstkov, E.A. Avetisyan, B.L. Tumanskii, S.R. Sterlin, N.N.
Bubnov, L.S. German, Izv. Akad. Nauk SSSR, Ser. Khim. (1990)
2450 [Bull. Acad. Sci. USSR, Div. Chem. Sci. 39 (1990) 2223].
[4] E.A. Avetisyan, B.L. Tumanskii, V.F. Cherstkov, S.R. Sterlin, L.S.
German, Izv. Akad. Nauk, Ser. Khim. (1993) 226 [Russ. Chem. Bull.
42 (1993) 207].
[5] L. TumanskiI, V.F. Cherstkov, N.I. Delyagina, S.R. Sterlin, N.N.
Bubnov, L.S. German, Izv. Akad. Nauk, Ser. Khim. (1995) 991
[Russ. Chem. Bull. 44 (1995) 962].
(c) Into a suspension of LiAlH4 (75.9 mg, 2 mmol) in
5 ml of dry THF was added the stock radical solution
(1.09 g, 2 mmol 1) with vigorous stirring at room tempera-
ture. After 1 h stirring, 80 ml of H2O, 80 ml of 15% aq.
NaOH, and 160 ml of H2O were added in that order. Per-
¯uoro phase separated at the bottom of the vessel was
analyzed by GC (see Run 8 in Table 3).
(d) The same reaction as (c) was carried out by using 5 ml
of dry diethyl ether instead of THF. The ether layer was
analyzed by GC (see Run 9 in Table 3).
[6] E.A. Avetisyan, B.L. Tumanskii, V.F. Cherstkov, S.R. Sterlin, Izv.
Akad. Nauk, Ser. Khim. (1995) 576 [Russ. Chem. Bull. 44 (1995)
558].
[7] U. Gross, S. Rudiger, A. Dimitrov, J. Fluorine Chem. 76 (1996)
139.
[8] B.E. Smart, in: M. Hudlicky (Ed.), Chemistry of Organofluorine
Compounds, Am. Chem. Soc., Washington, DC, 1995, p. 25.
[9] N. Ishikawa, M. Maruta, Yuki Gosei Kagaku Kyokaishi 39 (1981)
51.
[10] R.E. Fernandez, Perfluoroalkyl radical studies, Dissertation Thesis
for University of Southern California, 1987, p. 15.
[11] J.A. Dean (Ed.), Lange's Handbook of Chemistry, 13th ed.,
McGraw-Hill, New York, 1985, Table 6-2.
3.5.2.2. Reaction with sodium borohydride. Into
a
[12] N. Wiberg, Angew. Chem. Internat. Edit. 7 (1968) 766.
[13] G.N. Lewis, E. Lupkin, J. Am. Chem. Soc. 64 (1942) 2801.
[14] J. Robinson, R.A. Osteryoung, J. Am. Chem. Soc. 102 (1980) 4415.
[15] H. Debrodt, K.E. Heusler, Z. Phys. Chem. 125 (1981) 35.
[16] W.H. Bruning, R.F. Nelson, L.S. Marcoux, R.N. Adams, J. Phys.
Chem. 71 (1967) 3055.
suspension of NaBH4 (76 mg, 2.0 mmol) in 5 ml dry
diethyl ether was added the stock solution of 1 (1.09 g,
2.0 mmol) with vigorous stirring at room temperature over
3 min. No exothermic reaction occurred, and the ether layer
was colored pale yellow. After 1 h stirring, the ether layer,
the color of which was still sustained, was analyzed by GC
(see Run 10 in Table 3).
[17] M.J. Tricker, D.T.B. Tennakoon, J.M. Thomas, J. Heald, Clay and
Clay Minerals 23 (1975) 77.
[18] H. Debrodt, K.E. Heusler, Electrochim. Acta 25 (1980) 545.
[19] T.A. Blazer, R. Schmutzler, I.K. Gregor, Z. Naturforschg. 24b (1969)
1081.
3.5.2.3. Reaction with sodium hydride. Into a suspension of
NaH (48 mg, 2 mmol) in 5 ml of dry diethyl ether cooled in
[20] J. Burdeniuc, M. Sanford, R.H. Crabtree, J. Fluorine Chem. 91
(1998) 49.

182
T. Ono et al. / Journal of Fluorine Chemistry 97 (1999) 173±182
[21] O.W. Kolling, E.A. Mawdsley, Inorg. Chem. 9 (1970) 408.
[22] V.Y. Atamanyuk, V.G. Koshechko, V.D. Pokhodenko, Zh. Org.
Khim. 16 (1980) 1901.
[27] D.P. Curran, Synthesis (1988) 417.
[28] J. Lusztyk, B. Maillard, K.U. Ingold, J. Org. Chem. 51 (1986) 2457.
[29] W. Dmowski, W.T. Flowers, R.N. Haszeldine, J. Fluorine Chem. 9
(1977) 94.
[23] E.T. Seo, R.F. Nelson, J.M. Fritsch, L.S. Marcoux, D.W. Leedy, R.N.
Adams, J. Am. Chem. Soc. 88 (1966) 3498.
[30] T. Ono, H. Fukaya, K.V. Scherer, Reports of the National Industrial
Research Institute of Nagoya XLV No. 7, 1996, p. 321.
[31] E.L. Muetterties, W. Mahler, K.J. Packer, R. Schmutzler, Inorg.
Chem. 3 (1964) 1298.
[24] R.I. Walter, J. Am. Chem. Soc. 77 (1955) 5999.
[25] D.W.A. Sharp, J. Chem. Soc. (1957) 4804.
[26] W.P. Neumann, Synthesis (1987) 665.