Hydrodefluorinations by low-valent niobium catalyst

Article abstract of DOI:10.1016/j.jfluchem.2007.06.003

Catalytic hydrodefluorinations of organofluorine compounds by low-valent niobium catalyst were developed. In the presence of niobium(V) chloride (typically, 5 mol%), fluorobenzenes, α,α,α-trifluorotoluenes and (trifluoromethyl)pyridines were hydrodefluorinated with lithium aluminum hydride to give the corresponding benzenes, toluenes and methylpyridines in good yields, respectively.

Full text of DOI:10.1016/j.jfluchem.2007.06.003

Journal of Fluorine Chemistry 128 (2007) 1158–1167
www.elsevier.com/locate/fluor
Hydrodefluorinations by low-valent niobium catalyst
*
Kohei Fuchibe, Yoshitaka Ohshima, Ken Mitomi, Takahiko Akiyama
Department of Chemistry, Faculty of Science, Gakushuin University, 1-5-1 Mejiro, Toshima-ku, Tokyo 171-8588, Japan
Received 18 April 2007; received in revised form 6 June 2007; accepted 7 June 2007
Available online 13 June 2007
Abstract
Catalytic hydrodefluorinations of organofluorine compounds by low-valent niobium catalyst were developed. In the presence of niobium(V)
chloride (typically, 5 mol%), fluorobenzenes, a,a,a-trifluorotoluenes and (trifluoromethyl)pyridines were hydrodefluorinated with lithium
aluminum hydride to give the corresponding benzenes, toluenes and methylpyridines in good yields, respectively.
# 2007 Elsevier B.V. All rights reserved.
Keywords: C–F bond activation; Catalysis; Fluorobenzene; Hydrodefluorination; Niobium; Reduction; Trifluoromethylpyridine; Trifluorotoluene
1. Introduction
activated fluorobenzenes has been developed. Young and
Grushin reported the rhodium-catalyzed hydrodefluorination of
fluoronaphthalenes [14]. Jones and co-workers reported the
zirconocene(IV)-mediated hydrodefluorination of fluoroben-
zene and fluoroalkanes [15]. Other catalytic reactions that
involve the use of late transition metal catalysts have been
reported by Fort and co-workers [16] and others [17]. The
hydrodefluorination of fluoroalkenes has been also reported
recently [18,19].
The carbon–fluorine bond is one of the strongest single
bonds found in organic molecules [1]. Essentially because of
the high bond dissociation energies, the synthetic utility of
organofluorine compounds is limited and transition metal
catalyzed C–F bond activations have not been realized until
recently [2–4].
The hydrodefluorination of organofluorine compounds is an
important C–F bond activation [5] because it provides a way to
destroy fluorine-containing atmospheric pollutants [6] such as
CFCs (ozone-depleting materials) [7] and perfluoroalkanes
(greenhouse gases) [8]. The reductive detoxification of such
gaseous compounds having quite long half-lives in the
atmosphere [3c] is a practical way to reduce the levels of
fluorous, environmental hazardous wastes [9].
In this study, we focused on the hydrodefluorination of
organofluorine compounds with early transition metal catalysts.
In terms of Pauling’s electronegativity, fluorine is the most
electronegative element in the periodic table (EN(F) = 3.98). On
the other hand, early transition metal elements are more
electropositive than late transition metal elements (for instance,
EN(Nb) = 1.6 and EN(Pd) = 2.2). We expected that the electro-
static interaction of fluorine with an early transition metal
center is responsible for the high catalytic activity.
In spite of its importance, catalytic hydrodefluorination
emerged only in the last decade (Scheme 1). Aizenberg and
Milstein were the first to report catalytic hydrodefluorination of
hexafluorobenzene (rhodium, in 1994) [10,11]. Kiplinger and
Richmond reported the zirconocene(IV)-mediated reduction of
octafluoronaphthalene [12]. In the early stage of hydrode-
fluorination chemistry, reactions of perfluorobenzene deriva-
tives were mainly studied because these compounds are highly
electron-deficient and are therefore relatively reactive [13]. In
recent years, however, the catalytic hydrodefluorination of non-
Based on the above considerations, we examined catalytic
hydrodefluorination with early transition metal catalysts and
found that low-valent niobium species, which is generated from
niobium(V) chloride and lithium aluminum hydride in situ, is
an efficient catalyst. The synthetic utility of niobium, mainly in
the +III form, was pioneered by Pedersen and co-workers in the
mid 1980s [20–22], and the chemistry of Lewis acid
niobium(V) has been investigated in recent years [21a,23].
We disclose herein the novel catalytic activity of niobium at a
low oxidation state in the hydrodefluorination of fluoroben-
zenes [24], a,a,a-trifluorotoluenes [25] and (trifluoromethyl)-
pyridines [26].
* Corresponding author. Tel.: +81 3 3986 0221; fax: +81 3 5992 1029.
E-mail address: takahiko.akiyama@gakushuin.ac.jp (T. Akiyama).
0022-1139/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2007.06.003

K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
1159
Scheme 1.
2. Results and discussion
Table 1
2.1. Reaction conditions
Survey on catalyst
The catalytic activities of various early and late transition
metal salts were examined by using 4-fluorobiphenyl 1a as
substrate (Table 1). A 1,4-dioxane solution of 1a was treated
with lithium aluminum hydride (4 equiv.) in the presence of
20 mol% of a metal salt at room temperature. After refluxing
for 4 h, the reaction was quenched with water. Among the metal
salts examined, niobium(V) chloride gave the highest yield of
biphenyl 2a (entries 1–6, 10–13). It is worth noting that
palladium and nickel chloride gave essentially similar results to
that of the control experiment (entries 12–14). The load of
niobium(V) chloride could be reduced to 5 mol% (entries 6–9).
The low-valent niobium-catalyzed hydrodefluorination of
1a was found to proceed smoothly in ethereal solvents
(Table 2): 2a was obtained in good yields in 1,4-dioxane or
DME (entries 1 and 2). When the reaction was performed at
60 8C, the yield of 2a decreased to 52% (entry 3). The yield of
2a also decreased when the reaction was carried out in THF
(entry 4). In toluene or hexane, 2a was obtained in low yields
Entry
Catalyst,
mol%
2a (%)
Recovery
of 1a (%)
1
2
ZrCl4, 20
VCl3, 20
TaBr5, 20
TaCl5, 20
NbBr5, 20
NbCl5, 20
NbCl5, 10
NbCl5, 5
NbCl5, 2
WCl6, 20
FeCl3, 20
PdCl2, 20
NiCl2, 20
None
63
54
56
60
60
94
93
94
41
0
35
37
29
34
26
0
3
4
5
6
7
0
8
0
9
47
97
29
73
72
57
10
11
12
13
14
64
26
27
28

1160
K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
Table 2
Effect of solvents and reducing agentsa
system to give biphenyl 2a in good yields (entries 1, 2 and 4).
Not only electronically neutral fluorobiphenyls but also
electron-rich fluorobiphenyls were reduced readily to give
the corresponding products (entries 5–8). It is noteworthy that
the methyl ether moiety of 1g was tolerant to the reaction
conditions (entry 8). Arai and Nishida reported facile ether
cleavage of methyl ethers by niobium(V) chloride [31]. The
low-valent niobium catalyst was found to be less Lewis acidic
than the niobium(V). Carbon–chlorine bonds and carbon–
bromine bonds of 1h and 1i were reduced prior to the carbon-
fluorine bonds (entries 9 and 11). The complete reduction of
these fluorobenzenes was also accomplished by prolonging the
reaction (entries 10 and 12). 1-Fluoronaphthalene gave
naphthalene in satisfactory yield (entry 13). 3,4,5-Trifluor-
obiphenyl 1k was reduced completely to give biphenyl 2a in
high yield (entry 14). 1k selectively gave 3,5-difluorobiphenyl
2g when the reaction was quenched within 15 min (entry 15).
As is described in Section 1, perfluorobenzene derivatives
are relatively reactive and the catalytic hydrodefluorination of
the perfluorobenzenes was extensively studied [10,12,18a]. In
contrast, the catalytic hydrodefluorination of non-activated
fluorobenzene derivatives is fairly limited [14,16,17]. Table 3
suggests that low-valent niobium species is a highly active and
useful catalyst for the hydrodefluorination of monofluoro-
benzenes.
Entry
Solvent
Reducing
agent
2a (%)
Recovery
of 1a (%)
1
2
1,4-dioxane
DME
LiAlH4
LiAlH4
LiAlH4
LiAlH4
LiAlH4
LiAlH4
DIBAL
NaBH4
94
91
52
73
10
7
0
0
3b
DME
42
18
82
90
90
89
92
92
93
4
THF
5
6c
Toluene
Hexane
DME
7
0
8
DME
0
9
DME
n-Bu3SnH
Et3SiH
Nb turningsd
0
10
11
DME
0
DME
0
DME = 1,2-Dimethoxyethane.
a
Conditions: 1a, NbCl5 5 mol%, LiAlH4 2 equiv., reflux, 4 h.
The reaction was performed at 60 8C.
NbCl5 10 mol%.
b
c
d
LiAlH4 was not used. The reaction was performed for 7 h.
(entries 5 and 6). Although the coordination of solvent
molecule(s) to niobium seems to play an important role in the
C–F bond activation, the solubility of lithium aluminum
hydride in ethereal solvent is also likely an important factor.
Based on the yield and the temperature, we chose DME as the
most suitable solvent.
Lithium aluminum hydride was found to be the most
efficient reducing agent. Use of DIBAL, sodium borohydride,
tributyltin hydride and triethylsilane resulted in recovery of
starting material 1a (entries 7–10). Niobium turnings, in the
absence of lithium aluminum hydride, also did not give 1a
(entry 11).
2.2.2. Mechanistic considerations
It was found that the hydrogen introduced into the product
originates from lithium aluminum hydride (Scheme 3).
Hydrodefluorination of 1a with lithium aluminum deuteride
gave 4-deuteriobiphenyl (2a-d1) in quantitative yield. The
deuterium incorporation was determined to be 79% D based on
When lithium aluminum hydride was added to the dioxane
solution of 1a and niobium(V) chloride at room temperature,
the clear yellow solution instantaneously became colorless and
turned dark gray exothermically. The resulting heterogeneous
mixture violently evolved gas upon heating [27]. It has been
reported that niobium(V) chloride is reduced with lithium
aluminum hydride in diethyl ether at 0 8C to give metallic
niobium and dihydrogen [28]. We believe that zero-valent,
metallic niobium is generated in the reaction medium (Scheme
2). In this manuscript, however, we use the term ‘low-valent
niobium’ because we have not succeeded in detecting zero-
valent niobium directly. Here, it should be stated that Oshima
and Sato reported reductive reactions under the niobium(V)
chloride/sodium aluminum hydride system in 1982 [29,30].
1
the integrals on the H NMR spectrum.
Conjugation of the substrates strongly affected the rate of the
reaction. Hydrodefluorination of benzyl(fluoro)benzene 1l did
not proceed smoothly and only 13% yield of corresponding
hydrocarbon 2h was obtained (Scheme 4). Further, p- and o-
phenylated fluorobenzenes reacted faster than the corresponding
m-isomer; p- and o-fluorobiphenyls 1a and 1c gave 2a in high
yields within 4 h (Table 3, entries 1 and 4), whereas the
hydrodefluorination of m-fluorobiphenyl 1b required 6 h (entries
2 and 3). 1k underwent para-selective hydrodefluorination after a
short time (entry 15).
2.2. Catalytic hydrodefluorination of fluorobenzenes
2.2.1. Hydrodefluorination of fluorobenzene derivatives
Low-valent niobium was found to be quite effective for the
hydrodefluorination of monofluorobenzenes (Table 3) [24]. o-,
m- and p-Fluorobiphenyls 1a–c were reduced in the optimized
Scheme 3.
Scheme 2.
Scheme 4.

K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
1161
Table 3
Catalytic hydrodefluorination of fluorobenzenes
Entry
Fluorobenzene
Time (h)
Product
Yield (%)
Recovery (%)
1
2
3
4
4-Fluorobiphenyl (1a)
3-Fluorobiphenyl (1b)
1b
4.0
6.0
4.0
4.0
Biphenyl (2a)
91
91
64
93
0
0
2a
2a
2a
34
0
2-Fluorobiphenyl (1c)
5
10.0
3-Phenyltoluene (2b)
2-Ethoxybiphenyl (2c)
90
0
6
7
8
10.2
8.0
Quant.
90
0
0
0
4.0
87a
9
0.6
19b
98d
76
0
10
1h
1i
4.0c
2a
11
12
0.5
40
90
49
0
8.0e
2a
13
14
2.2
Naphthalene (2f)
81
0
0
6
6.0e
2a
91
15
1k
0.25f
80g
a
5% yield of 4-phenylphenol was obtained.
Biphenyl 2a was not isolated.
3.5 equiv. of LiAlH4 was used.
b
c
d
e
f
4-Fluorobiphenyl 1a was not isolated.
10 mol% of NbCl5 and 6 equiv. of LiAlH4 were used.
6 equiv. of LiAlH4 was used.
g
Biphenyl 2a was not obtained. A few percents yield of m-hydrodefluorinated product was detected by GC–MS analysis.
Based on the observations, we propose an SNAr-like,
2.3. Catalytic hydrodefluorination of a,a,a-
trifluorotoluenes
nucleophilic substitution mechanism of the fluorine atom with
metal hydride species. The pendant phenyl group stabilizes the
anionic charge formed in the addition intermediate by
delocalization. Phenyl groups on p- and o-positions, in
particular, stabilize the partial anionic charge efficiently
[15b,32].
2.3.1. Hydrodefluorination of a,a,a-trifluorotoluenes
It is well recognized that the CF3 groups of a,a,a-
trifluorotoluenes are considerably stable and remain intact
under vigorous conditions. The C–F bonds are indeed benzylic,

1162
K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
but chemical transformation of the C–F bonds is extremely
difficult [33]. Hydrodefluorination of the trifluorotoluenes was
accomplished only very recently [34].
were reduced to give m-phenyltoluene 2b in 81% yield (entry
10). When the reduction of 3j was performed with a reduced
amount of lithium aluminum hydride (6 versus 3 equiv.), partial
reduction products 3b and 6, in which no chlorine atom
survived, were obtained (entry 11). On the other hand,
chlorotrifluorotoluene 3k gave considerable amounts of
chlorotoluene 4j (entry 12). Reactivities of the CF3 group,
the aromatic C–F bond, and the aromatic C–Cl bond were thus
found to follow roughly the order CAr–F < CAr–Cl ꢀ ArCF3.
The reduction of 4,40-bis(trifluoromethyl)biphenyl 3l also
proceeded smoothly to give 4c in 58% yield (entry 13).
The low-valent niobium catalyst was found to be applicable
to the hydrodefluorination of a,a,a-trifluorotoluene derivatives
(Table 4) [25]. Under conditions identical to those described in
Table 3, phenyl trifluorotoluenes 3a–f gave the corresponding
toluenes in high yields (entries 1–6). It is noteworthy that the
aromatic C–F bonds were not hydrodefluorinated under the
reaction conditions (entries 5 and 6). In the case of 2-phenyl-
a,a,a-trifluorotoluene 3g, corresponding 2-phenyltoluene 4g
was obtained but in 51% yield, and fluorene 5 was obtained in
15% yield (entry 7). 5 was generated by double activation of the
C–F bond and the aromatic C–H bond [26a]. Not only phenyl
trifluorotoluenes but also parent and benzyl trifluorotoluenes 3h
and 3i gave corresponding products 4h and 4i, respectively
(entries 8 and 9). When chlorotrifluorotoluene 3j was subjected
to the reaction conditions, both C–F bonds and the C–Cl bond
2.3.2. Partial reduction of bis(trifluoromethyl)benzene
The selective, stepwise hydrodefluorination of bis(trifluor-
omethyl)benzene 3m was also accomplished (Scheme 5). With
a large excess of lithium aluminum hydride, 3m gave
completely hydrodefluorinated product 4d in 78% yield. In
contrast, 3m gave partially hydrodefluorinated product 3d in
Table 4
Catalytic hydrodefluorination of a,a,a-trifluorotoluenesa
Entry
Trifluorotoluene
Time (h)
LiAlH4 (equiv.)
Products, yields (%)
1
2
4-Phenyl-a,a,a-trifluorotoluene (3a)
3-Phenyl-a,a,a-trifluorotoluene (3b)
4.0
6.9
4
4
4-Phenyltoluene (4a), 96
3-Phenyltoluene (2b), 94
3
4
4.0
4.0
5
5
4,40-Dimethylbiphenyl (4c), 85
3,5-Dimethylbiphenyl (4d), 82
, 82
5
6
4.0
4.0
5
3
, quant.
7
8
2-Phenyl-a,a,a-trifluorotoluene (3g)
4.3
8.0
4
4
2-Phenyltoluene (4g), 51 Fluorene (5), 15
, quant.b
9c
8.0
10
, 58
10
11
4.0
4.0
6
3
2b, 81
2b, 33, 3b, 12,
, 27
3j
12
4.0
4.0
4
4-Chloro-40-methylbiphenyl (4j), 38 4a, 41
4c, 58
13
10
a
Conditions: NbCl5 5 mol%, DME (1,2-dimethoxyethane), reflux.
NMR yield after distillation. Dibromomethane was used as an internal standard.
NbCl5 100 mol%.
b
c

K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
1163
Scheme 7.
Scheme 5.
Scheme 8.
1.98 atom D, Scheme 6). On the other hand, when the reaction
was quenched with deuterium oxide, 4-phenyl-a-deuterioto-
luene 4a-d1 was obtained (82% yield, 1.04 atom D, Scheme 7).
Apparently, two of the fluorines are replaced with hydride
hydrogens in situ, and (dihydro)benzylic anion equivalent 7 is
formed [37].
Scheme 6.
The precursor of 7 was found not to be benzyl fluoride 8
(Scheme 8). Separately prepared 8 was treated with lithium
aluminum deuteride, and the reaction was quenched with H2O.
We found that deuterium was incorporated into the product (4a-
d1, 1.02 atom D). This result suggests that 8 is not involved in
the hydrodefluorination of 3a.
The results of the deuterium labeling experiments and the
formation of fluorene 5 can be explained by assuming a
niobium fluorocarbenoid intermediate (Scheme 9). Fluorocar-
benoid species 9 is reductively generated from substrate 3 and
low-valent niobium species. 9 absorbs two hydride anions from
the reducing agent and liberates one fluoride anion to produce
benzylanion equivalent 7. Protonation of 7 affords toluene
derivative 4. On the other hand, when o-phenyltrifluorotoluene
3g is used, 9 undergoes formal carbene insertion into the
neighboring aromatic C–H bond to give 5 (double activation)
77% yield with 3 equiv. of the reducing agent. One of the most
general methods to prepare a,a,a-trifluorotoluene derivatives
is the fluorine–chlorine exchange of a,a,a-trichlorotoluenes
[35], which are prepared by chlorination of the corresponding
toluenes [36]. The low-valent niobium-catalyzed partial
reduction is a potentially useful method to produce aromatic
compounds that possess both CH3 and CF3 groups.
2.3.3. Mechanistic considerations
Deuterium labeling experiments suggested that the nio-
bium-catalyzed hydrodefluorination of a,a,a-trifluoroto-
luenes does not proceed via a simple SN2 mechanism;
hydrodefluorination of 3a with lithium aluminum deuteride
gave 4-phenyl-a,a-dideuteriotoluene 4a-d2 (95% yield,
Scheme 9.

1164
K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
Table 5
Hydrodefluorination of (tifluromethyl)pyridinea
Entry
1
(Trifluoromethyl)pyridine
Time (h)
4
Products, yields (%)
Complex mixture
2b
10a
10b
4
3
0.5
4
6.5c
11b, 9, 11a, 41
a
Conditions: (trifluoromethyl)pyridine, NbCl5 20 mol%, LiAlH4 4 equiv., 1,4–dioxane, reflux.
NbCl5 was not used.
b
c
4 equiv. of LiAlH4, 4 h then 2 equiv. of LiAlH4, 2.5 h.
[26a]. However, we do not have solid data to support the
formation of intermediate 9 at present.
standard. Chemical shifts (d) for 13C were referenced to a
solvent signal (CDCl3, d = 77.0 ppm). Chemical shifts (d) for
19F were referenced to a,a,a-trifluorotoluene (d = ꢁ63.9 ppm)
as an internal standard. IR spectra were recorded on FTIR-
8600PC instrument (Shimadzu Co.) using CHCl3 as a solvent.
Mass spectra (LRMS) were recorded on GCMS-QP5000
instrument (Shimadzu Co.).
2.4. Hydrodefluorination of (trifluoromethyl)pyridines
The hydrodefluorination of (trifluoromethyl)pyridine deri-
vatives was also performed. The low-valent niobium catalyst
exhibited remarkable effects in this system: 2-phenyl-5-
(trifluoromethyl)pyridine 10a underwent smooth hydrodefluor-
ination with 20 mol% of the niobium catalyst in dioxane
(Table 5, entry 1) [38]. In contrast, 10a gave a complex mixture
in the absence of the catalyst (entry 2). GC–MS analysis
suggested that the pyridine nucleus was competitively reduced
in the latter case. Fluoro(trifluoromethyl)pyridine 10b under-
went CF3-selective hydrodefluorination to give corresponding
fluoro(methyl)pyridine 11b in good yield (entry 3). Complete
hydrodefluorination of 10b also proceeded to give 11a after
prolonged reaction with an excess amount of the reducing agent
(entry 4).
4.2. Typical procedures
Caution: Addition of solid lithium aluminum hydride to
niobium(V) chloride must be carried out with great care. The
reaction mixture possibly flow out of the container due to the
exothermic, violent gas evolution.
4.2.1. Hydrodefluorination of fluorobenzenes
Reaction of 4-fluorobiphenyl 1a is described as a typical
procedure (Table 3, entry 1); to a DME solution (3 mL) of p-
fluorobiphenyl (217 mg, 1.26 mmol) and NbCl5 (17 mg,
0.06 mmol) was added LiAlH4 (96 mg, 2.52 mmol) in one
portion. The clear yellow solution turned dark gray immedi-
ately and gas evolved exothermically. After being refluxed for
4 h, the reaction mixture was quenched with H2O at 0 8C.
Sodium tartrate (0.2 g) was added and extraction with ethyl
acetate gave the crude mixture. Purification by silica gel
column chromatography (hexane) afforded biphenyl 2a
(176 mg, 1.14 mmol, 91%).
3. Conclusions
Low-valent niobium was found to be an efficient catalyst
for the hydrodefluorination of fluorobenzenes, a,a,a-trifluor-
otoluenes and (trifluoromethyl)pyridines. Using 5 mol% low-
valent niobium catalyst, high yields of the corresponding
hydrocarbons were obtained. Although the chemistry of
group 5 metal elements is underdeveloped compared to that
of late transition metal elements, the results described herein
strongly demonstrate the synthetic utility of group 5 metals as
catalysts.
4.2.2. Hydrodefluorination of a,a,a-trifluorotoluenes
Reduction of 3a is described as a typical procedure (Table 4,
entry 1): to a DME solution (5 mL) of 3a (308 mg, 1.39 mmol)
and NbCl5 (18 mg, 0.068 mmol) was added solid lithium
aluminum hydride (208 mg, 5.47 mmol) carefully. The mixture
was magnetically stirred and allowed to reflux for 4 h. The
reaction was quenched with water at 0 8C and a small amount of
sodium tartrate (ca. 200 mg) was added for the ease of
extraction. Products were extracted with ethyl acetate three
times and the combined organic layers were dried over
anhydrous sodium sulfate. The organic solvents were removed
under reduced pressure and purification by column chromato-
4. Experimental
4.1. General information
NMR spectra were recorded on Unity Inova-400 instrument
(Varian Ltd., 400 MHz for 1H, 100 MHz for 13C, 376 MHz for
19F) using CDCl3 as a solvent. Chemical shifts (d) for 1H were
referenced to tetramethylsilane (d = 0.00 ppm) as an internal

K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
1165
graphy (SiO2, hexane) gave 4a (218 mg, 1.33 mmol, 96%
yield) as a colorless crystal.
4.3.1.1. 2-Ethoxybiphenyl (2c). 1H NMR (400 MHz, CDCl3):
d = 1.37 (3H, t, J = 6.8 Hz), 4.06 (2H, q, J = 6.8 Hz), 6.99 (1H,
d, J = 8.0 Hz), 7.04 (1H. dt, J = 7.6 1.2 Hz), 7.29–7.38 (3H, m),
7.42 (2H, t, J = 8.0 Hz), 7.60 (2H, m). 13C NMR (100 MHz,
CDCl3): d = 14.7, 63.9, 112.6, 120.8, 126.7, 127.8, 128.5,
4.2.3. Complete hydrodefluorination of 3,5-
bis(trifluoromethyl)biphenyl 3m (Scheme 5)
˜
To a DME solution (9.4 mL) of 3m (225 mg, 0.778 mmol)
and NbCl5 (11 mg, 0.039 mmol) was added solid lithium
aluminum hydride (295 mg, 7.78 mmol) carefully. The mixture
was magnetically stirred and allowed to reflux for 4 h. The
reaction was quenched with water at 0 8C and a small amount of
sodium tartrate (ca. 200 mg) was added for the ease of
extraction. Products were extracted with ethyl acetate three
times and the combined organic layers were dried over
anhydrous sodium sulfate. The organic solvents were removed
under reduced pressure and purification by column chromato-
graphy (SiO2, hexane) gave 4d (111 mg, 0.607 mmol, 78%
129.5, 130.8, 130.9, 138.6, 155.8. IR (CHCl3): n ¼ 3020, 1475,
1435, 1209, 1126, 1045, 928, 731, 671 cmꢁ1. LRMS (70 eV,
EI): m/z (rel. int.) = 115 (61), 169 (98, M+ꢁEt), 170 (100,
M+ꢁEt+H), 198 (58, M+).
4.3.1.2. 3,5-Difluorobiphenyl (2g). 1H NMR (400 MHz,
CDCl3): d = 6.78 (1H, tt, J = 8.8, 2.0 Hz), 7.10 (2H, dd,
J = 8.8, 2.4 Hz), 7.39 (1H, t, J = 7.2 Hz), 7.45 (2H, t,
J = 7.2 Hz), 7.54 (2H, d, J = 7.2 Hz). 13C NMR (100 MHz,
CDCl3): d = 102.4 (t, J = 100 Hz), 109.9 (dd, J = 18, 7 Hz),
127.0, 128.4, 129.0, 138.9, 144.5 (t, J = 10 Hz), 163.3 (dd,
J = 247, 13 Hz). 19F NMR (376 MHz, CDCl3): d = ꢁ110.95 to
yield)
0.02 mmol, 3% yield).
and
3-difluoromethyl-5-methylbiphenyl (5 mg,
˜
ꢁ111.06 (m). IR (CHCl3): n ¼ 1620, 1601, 1421, 1340, 1119,
989, 860, 696 cmꢁ1. LRMS (70 eV, EI): m/z (rel. int.) = 170
(13, M+ꢁHF), 190 (100, M+).
4.2.4. Partial hydrodefluorination 3m (Scheme 5)
To a DME solution of 3m (1.23 g, 4.25 mmol) and NbCl5
(58 mg, 0.22 mmol) was added solid lithium aluminum hydride
(477 mg, 12.6 mmol) carefully. The mixture was magnetically
stirred and allowed to reflux for 4 h. The reaction was quenched
with water at 0 8C and a small amount of sodium tartrate (ca.
500 mg) was added for the ease of extraction. Products were
extracted with ethyl acetate three times and the combined
organic layers were dried over anhydrous sodium sulfate. The
organic solvents were removed under reduced pressure and
purification by column chromatography (SiO2, hexane) gave 3d
(773 mg, 3.27 mmol, 77% yield), 4d (31 mg, 0.17 mmol, 4%,
yield), and a mixture of 3,5-bis(difluoromethyl)biphenyl and 3-
difluoromethyl-5-methylbiphenyl (105 mg, 2% and 9% yields,
respectively).
4.3.2. Hydrodefluorination products of a,a,a-
trifluorotoluenes
Characterization data of 4a, 4c–f, 4i, 3b and 6 were
described in our previous communication [25]. Characteriza-
tion of 4g and 5 was described in our previous communication
[26a]. 1H NMR spectrum of 4h was in complete agreement with
the authentic toluene.
4.3.2.1. 4-Chloro-40-methylbiphenyl (4j). 1H NMR (400 MHz,
CDCl3): d = 2.39 (3H, s), 7.25 (2H, d, J = 8.0 Hz), 7.38 (2H, d,
J = 8.0 Hz), 7.45 (2H, d, J = 8.0 Hz), 7.50 (2H, d, J = 8.0 Hz).
13C NMR (100 MHz, CDCl3): d = 21.1, 126.8, 128.2, 128.8,
˜
129.6, 133.0, 137.1, 137.4, 139.6. IR (CHCl3): n ¼ 2926, 2855,
1489, 1481, 1096, 1007, 810 cmꢁ1. LRMS (70 eV, EI): m/z (rel.
int.) = 167 (55, M+ꢁCl), 202 (100, M+(C13H1135Cl+)), 204 (33,
M+(C13H1137Cl+)).
4.2.5. Hydrodefluorination of (trifluoromethyl)pyridines
Reaction of 5-trifluoromethyl-2-phenylpyridine 10a is
described as a typical procedure (Table 5, entry 1). To a 1,4-
dioxane solution of 10a (273 mg, 1.23 mmol) and NbCl5
(66 mg, 0.24 mmol) was added solid lithium aluminum hydride
(188 mg, 4.95 mmol) carefully. The mixture was magnetically
stirred and allowed to reflux for 4 h. The reaction was quenched
with water at 0 8C and a small amount of sodium tartrate (ca.
200 mg) was added for the ease of extraction. Products were
extracted with ethyl acetate three times and the combined
organic layers were dried over anhydrous sodium sulfate. The
organic solvents were removed under reduced pressure and
purification by column chromatography (SiO2, hexane/ethyl
acetate = 10/1) gave 11a (133 mg, 0.79 mmol, 65% yield) as a
colorless crystal.
4.3.3. Hydrodefluorination products of
(trifluoromethyl)pyridines
4.3.3.1. 5-Methyl-2-phenylpyridine (11a). 1H NMR (400
MHz, CDCl3): d = 2.38 (3H, s), 7.39 (1H, t, J = 7.2 Hz),
7.46 (2H, t, J = 7.2 Hz), 7.56 (1H, dd, J = 8.0, 2.0 Hz), 7.63
(1H, d, J = 8.0 Hz), 7.96 (2H, dd, J = 8.4, 1.2 Hz), 8.52 (1H, s).
13C NMR (100 MHz, CDCl3): d = 18.1, 120.0, 126.7, 128.6,
128.7, 131.6, 137.3, 139.4, 150.0, 154.8. IR (CHCl3):
n ¼ 3020, 1479, 1213, 752, 667 cmꢁ1. LRMS (70 eV, EI):
˜
m/z (rel. int.) = 169 (100, M+), 168 (53, M+ꢁH), 154 (9,
M+ꢁCH3).
4.3.3.2. 2-(4-Fluorophenyl)-5-methylpyridine (11b). 1H NMR
(400 MHz, CDCl3): d = 2.37 (3H, s), 7.14 (2H, t, J = 8.8 Hz),
7.54–7.58 (2H, m), 7.94 (2H, dd, J = 9.2, 5.2 Hz), 8.50 (1H, s).
13C NMR (100 MHz, CDCl3): d = 18.1, 115.5 (d, J = 21.3 Hz),
119.7, 128.4 (d, J = 8.4 Hz), 131.5, 135.6, 137.4, 150.0, 153.8,
163.2 (d, J = 246 Hz). 19F NMR (376 MHz, CDCl3): d = ꢁ115.0
4.3. Characterization of the products
4.3.1. Hydrodefluorination products of fluorobenzenes
1H NMR spectra of 2a [39], 2b [40], 2d [41], 2e [42], 1a
[43], 1c [44], 2f [45] and 2h [46] were in complete agreement
with those in the literature.
˜
to ꢁ114.9 (m). IR (CHCl3): n ¼ 1605, 1514, 1479, 1157,

1166
K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
826 cmꢁ1. LRMS (70 eV, EI): m/z (rel. int.) = 187 (100, M+),
186 (42, M+ꢁH).
(b) N.A. Jasim, R.N. Perutz, A.C. Whitwood, T. Braun, J. Izundu, B.
Neumann, S. Rothfeld, H.-G. Stammler, Organometallics 23 (2004) 6140–
6149.
[14] R.J. Young Jr., V.V. Grushin, Organometallics 18 (1999) 294–296.
[15] (a) B.M. Kraft, R.J. Lachicotte, W.D. Jones, J. Am. Chem. Soc. 122
(2000) 8559–8560;
Acknowledgements
(b) B.M. Kraft, R.J. Lachicotte, W.D. Jones, J. Am. Chem. Soc. 123
(2001) 10973–10979;
This work was supported in part by Grant-in-Aid for
Scientific Research from the Ministry of Education, Culture,
Sports, Science, and Technology, Japan (No. 18750036). This
work was also supported by Saneyoshi scholarship foundation
(No. 1818).
(c) W.D. Jones, Dalton Trans. (2003) 3991–3995.
[16] S. Kuhl, R. Schneider, Y. Fort, Adv. Synth. Catal. 345 (2003) 341–344.
[17] (a) H. Yang, H. Gao, R.J. Angelici, Organometallics 18 (1999) 2285–
2287;
(b) P.P. Cellier, J.-F. Spindler, M. Taillefer, H.-J. Cristau, Tetrahedron
Lett. 44 (2003) 7191–7195.
References
[18] (a) J. Vela, J.M. Smith, Y. Yu, N.A. Ketterer, C.J. Flaschenriem, R.J.
Lachicotte, P.L. Holland, J. Am. Chem. Soc. 127 (2005) 7857–7870;
(b) B.M. Kraft, W.D. Jones, J. Am. Chem. Soc. 124 (2002) 8681–8689;
(c) M.S. Kirkham, M.F. Mahon, M.K. Whittlesey, Chem. Commun.
(2001) 813–814;
[1] S.J. Blanksby, G.B. Ellison, Acc. Chem. Res. 36 (2003) 255–263.
[2] (a) T. Schaub, M. Backes, U. Radius, J. Am. Chem. Soc. 128 (2006)
15964–15965;
(b) A. Steffen, M.I. Sladek, T. Braun, B. Neumann, H.-G. Stammler,
Organometallics 24 (2005) 4057–4064;
(d) D. Noveski, T. Braun, M. Schulte, B. Neumann, H.-G. Stammler,
Dalton Trans. (2003) 4075–4083;
´
(e) E. Clot, C. Megret, B.M. Kraft, O. Eisenstein, W.D. Jones, J. Am.
(c) J. Terao, A. Ikumi, H. Kuniyasu, N. Kambe, J. Am. Chem. Soc. 125
(2003) 5646–5647;
Chem. Soc. 126 (2004) 5647–5653.
´
[19] See also: D. Guijarro, P. Martınez, M. Yus, Tetrahedron, 59 (2003) 1237–
(d) N. Yoshikai, H. Mashima, E. Nakamura, J. Am. Chem. Soc. 127
(2005) 17978–17979;
1244.
(e) T. Saeki, Y. Takashima, K. Tamao, Synletters (2005) 1771–1774;
(f) M. Fujiwara, J. Ichikawa, T. Okauchi, T. Minami, Tetrahedron Lett. 40
(1999) 7261–7265;
[20] (a) E.J. Roskamp, S.F. Pedersen, J. Am. Chem. Soc. 109 (1987) 6551–
6553;
(b) J.B. Hartung, S.F. Pedersen, Oganometallics 9 (1990) 1414–1417.
[21] For reviews:
(g) Y. Ishii, N. Chatani, S. Yorimitsu, S. Murai, Chem. Lett. (1998) 157–
158;
(a) C.K.Z. Andrade, Curr. Org. Synth. 1 (2004) 333–353;
(b) J.A. Labinger, in: G. Wilkinson, F.G.A. Stone, E.W. Abel (Eds.),
Comprehensive Organometallic Chemistry, vol. 3, Pergamon Press, New
York, 1982, pp. 705–782;
(h) Y.M. Kim, S. Yu, J. Am. Chem. Soc. 125 (2003) 1696–1697;
(i) See also: Y. Kiso, K. Tamao, M. Kumada, J. Organomet. Chem. 50
(1973) C12–C14.
[3] For reviews, see:
(c) D.E. Wigley, S.D. Gray, in: E.W. Abel, F.G.A. Stone, G. Wilkinson
(Eds.), Comprehensive organometallic Chemistry II, vol. 5, Pergamon,
New York, 1995, pp. 57–153.
(a) T.G. Richmond, Activation of unreactive bonds and organic synthesis,
in: S. Murai (Ed.), Topics in Organometallic Chemistry, vol. 3, Springer,
Berlin, 1999, pp. 243–269;
[22] See also:
(b) J.L. Kiplinger, T.G. Richmond, C.E. Osterberg, Chem. Rev. 94 (1994)
373–431;
(a) Y. Kataoka, K. Takai, K. Oshima, K. Utimoto, Tetrahedron Lett. 31
(1990) 365–368;
(c) J. Burdeniuc, B. Jedlicka, R.H. Crabtree, Chem. Ber. 130 (1997) 145–
154.
(b) Y. Kataoka, J. Miyai, M. Tezuka, K. Takai, K. Oshima, K. Utimoto,
Tetrahedron Lett. 31 (1990) 369–372;
[4] See also:
(c) S. Arai, S. Takita, A. Nishida, Eur. J. Org. Chem. (2005) 5262–5267;
(d) Y. Obora, M. Kimura, M. Tokunaga, Y. Tsuji, Chem. Commun. (2005)
901–902;
(a) S.A. Begum, J. Terao, N. Kambe, Chem. Lett. 36 (2007) 196–197;
(b) K. Hirano, K. Fujita, H. Yorimitsu, H. Shinokubo, K. Oshima,
Tetrahedron Lett. 45 (2004) 2555–2557.
¨
(e) J. Szymoniak, J. Besanc¸on, C. Moıse, Tetrahedron 50 (1994) 2841–
[5] O. Paleta, Replacement of fluorine by hydrogen, in: B. Baasner, H.
Hagemann, J.C. Tatlow (Eds.), Organo-Fluorine Compounds/Methods of
Organic Chemistry, vol. E 10b/Part 2, Houben-Weyl, Stuttgart, 2000, pp.
306–344.
2848;
(f) B. Kammermeier, G. Beck, D. Jocobi, H. Jendralla, Angew. Chem. Int.
Ed. Engl. 33 (1994) 685–687;
(g) A.C. Williams, P. Sheffels, D. Sheehan, T. Livinghouse, Organome-
tallics 8 (1989) 1566–1567.
[6] F. Alonso, I.P. Beletskaya, M. Yus, Chem. Rev. 102 (2002) 4009–4138.
[7] M.J. Molina, F.S. Rowland, Nature 249 (1974) 810–812.
[8] (a) D.G. Victor, G.J. MacDonald, Climat. Change 42 (1999) 633–662;
(b) C.M. Roehl, D. Boglu, C. Bru¨hl, G.K. Moortgat, Geophys. Res. Lett.
22 (1995) 815–818;
[23] (a) S. Kobayashi, K. Arai, H. Shimizu, Y. Ihori, H. Ishitani, Y. Yamashita,
Angew. Chem. Int. Ed. 44 (2005) 761–764;
(b) K. Arai, M.M. Salter, Y. Yamashita, S. Kobayashi, Angew. Chem. Int.
Ed. 46 (2007) 955–957.
(c) P.L. Timms, J. Chem. Soc., Dalton Trans. (1999) 815–822.
[9] Commodore Chemical Co. in Columbus Ohio developed reductive dis-
posals of CFCs.
[24] K. Fuchibe, T. Akiyama, Synletters (2004) 1282–1284.
[25] K. Fuchibe, Y. Ohshima, K. Mitomi, T. Akiyama, Org. Lett. 9 (2007)
1497–1499.
[10] (a) M. Aizenberg, D. Milstein, Science 265 (1994) 359–361;
(b) M. Aizenberg, D. Milstein, J. Am. Chem. Soc. 117 (1995) 8674–
8675.
[26] See for our preliminary C–C bond forming reactions using CF3 groups,
also:
(a) K. Fuchibe, T. Akiyama, J. Am. Chem. Soc. 128 (2006) 1434–1435;
(b) K. Fuchibe, K. Mitomi, T. Akiyama, Chem. Lett. 36 (2007) 24–
25.
[11] See also: T. Imamoto, T. Takeyama, T. Kusumoto, Chem. Lett. (1985)
1491–1492.
[12] (a) J.L. Kiplinger, T.G. Richmond, Chem. Commun. (1996) 1115–1116;
(b) J.L. Kiplinger, T.G. Richmond, J. Am. Chem. Soc. 118 (1996) 1805–
1806.
[27] Dihydrogen was detected by gas indicator (GASTEC corporation, gas
indicator tube No. 30).
[28] M.E. Kost, A.I. Golovanova, Zh. Neorg. Khim. 22 (1977) 977–979.
[29] M. Sato, K. Oshima, Chem. Lett. (1982) 157–160.
[30] Other reductive reactions by (transition) metal/lithium aluminum hydride
systems:
[13] See also:
(a) B.L. Edelbach, W.D. Jones, J. Am. Chem. Soc. 119 (1997) 7734–
7742;

K. Fuchibe et al. / Journal of Fluorine Chemistry 128 (2007) 1158–1167
1167
(a) J.O. Osby, S.W. Heinzman, B. Ganem, J. Am. Chem. Soc. 108 (1986)
(b) F.D. Marsh, W.B. Farnham, D.J. Sam, B.E. Smart, J. Am. Chem. Soc.
104 (1982) 4680–4682.
67–72;
(b) E.C. Ashby, J.J. Lin, J. Org. Chem. 43 (1978) 1263–1265;
and reference [11].
[37] It was confirmed that 7 was not generated by deprotonation of the product.
Treatment of isolated 4a under the reaction conditions and quenching the
reaction with deuterium oxide did not afford deuterionated 4a at all (88%
recovery).
[31] (a) See also: Y. Sudo, S. Arai, A. Nishida, Eur. J. Org. Chem. (2006) 752–
758;
(b) Q. Guo, T. Miyaji, R. Hara, B. Shen, T. Takahashi, Tetrahedron 58
(2002) 7327–7334.
[38] When the reaction was performed in DME 57% yield of 11a was obtained.
[39] C.P. Pouchert, J. Behnke (Eds.), The Aldrich Library of 13C and 1H FT
NMR Spectra, Edition I, vol. 2, Aldrich Chemical Company, Inc., 1993, p.
28.
[32] B.M. Kraft, W.D. Jones, J. Organomet. Chem. 658 (2002) 132–140.
´
[33] (a) C. Saboureau, M. Troupel, S. Sibille, J. Perichon, J. Chem. Soc. Chem.
Commun. (1989) 1138–1139;
[40] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993, p. 32.
[41] N.H. Lajis, M.N. Khan, H.A. Hassan, Tetrahedron 49 (1993) 3405–3410.
[42] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993, p. 216.
[43] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993, p. 157.
[44] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993 , p.
156.
[45] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993, p. 40.
[46] C.P. Pouchert, J. Behnke (Eds.), I ed., The Aldrich Library of 13C and 1H
FT NMR Spectra, vol. 2, Aldrich Chemical Company, Inc., 1993, p. 7.
(b) D. Ferraris, C. Cox, R. Anand, T. Lectka, J. Am. Chem. Soc. 119
(1997) 4319–4320;
(c) H. Amii, Y. Hatamoto, M. Seo, K. Uneyama, J. Org. Chem. 66 (2001)
7216–7218;
See also:
(d) K. Uneyama, H. Amii, J. Fluorine Chem. 114 (2002) 127–131.
[34] V.J. Scott, R. C¸ elenligil-C¸ etin, O.V. Ozerov, J. Am. Chem. Soc. 127 (2005)
2852–2853.
[35] (a) J.H. Simons, C.J. Lewis, J. Am. Chem. Soc. 60 (1938) 492–1492;
(b) A.J. Bloodworth, K.J. Bowyer, J.C. Mitchell, Tetrahedron Lett. 28
(1987) 5347–5350.
[36] (a) P.G. Harvey, F. Smith, M. Stacey, J.C. Tatlow, J. Appl. Chem. 4 (1954)
319–325;