Highly reactive and regenerable fluorinating agent for oxidative fluorination of aromatics

Article abstract of DOI:10.1021/op700266y

A newly synthesized copper aluminum fluoride of nominal composition CuAl₂F₈ exhibits excellent reactivity towards direct oxidative fluorination of aromatic compounds, as well as fluorodechlorination of chloroaromatics. The spent CuAl₂F₈ reagent can be regenerated by treatment with O₂ and HF, and the fluorination process has been demonstrated to retain high conversions through 20 reaction cycles. The main advantages of this new process are safety, minimal waste, and potentially low cost.

Full text of DOI:10.1021/op700266y

Organic Process Research & Development 2008, 12, 349–354
Highly Reactive and Regenerable Fluorinating Agent for Oxidative Fluorination of
Aromatics
Krishna Murthy Janmanchi and William R. Dolbier, Jr.*
Department of Chemistry, UniVersity of Florida, GainesVille, Florida 32611-7200
Abstract:
CsSO4F,7,13 various N-F8,14,15 and N+-F compounds,8,16,17 as well
as elemental F2 itself.5 However, with most of these compounds
requiring F2 for synthesis, they are all relatively expensive,
generally highly toxic, and often dangerous to use, thus
effectively precluding their use in commercial manufacture of
fluoroaromatics.
Alternatively, fluorides of transition metals have been
examined as fluorinating agents. Organic compounds, including
aromatic systems, have been oxidatively fluorinated using metal
fluorides such as CoF3, KCoF4, AgF2, CeF4, and MnF3.18–21
Unfortunately such reagents have not been found to be useful
for introducing only one or two fluorines into benzene or other
aromatics, generally acting to exhaustively fluorinate com-
pounds, with the result that saturated polyfluoro or perfluoro
compounds are usually obtained. Moreover, they require
expensive elemental F2 to prepare or regenerate. Therefore, the
discovery of a transition metal fluoride reagent that is powerful
enough to fluorinate but not so reactive as to destroy an aromatic
system and that can be regenerated without the need for F2 is
a noteworthy event.
In general, the relative fluorinating activities of transition
metal fluorides are directly related to their respective redox
potentials. For example, the fluorides of metal ions that have
E° > 1 comprise the potent group of fluorinating agents
mentioned above that give rise to exhaustive fluorination. In
contrast, fluorides of the metal ions with E° < 0, such as ZnF2,
MgF2, and AlF3, are inert towards aromatic hydrocarbons. On
the other hand, fluorides of metal ions with 1 > E° > 0, such
as CuF2, AgF, HgF2, and Hg2F2, can act as mild fluorinating
agents. In principle, all of these fluorides can be reoxidized and
the fluorides regenerated by sequential or simultaneous treatment
with O2 and HF. Among these reagents, copper(II) fluoride
appeared to have the greatest potential with respect to fluorinat-
ing benzene, as was demonstrated recently by Subramanian and
A newly synthesized copper aluminum fluoride of nominal
composition CuAl2F8 exhibits excellent reactivity towards direct
oxidative fluorination of aromatic compounds, as well as fluoro-
dechlorination of chloroaromatics. The spent CuAl2F8 reagent can
be regenerated by treatment with O2 and HF, and the fluorination
process has been demonstrated to retain high conversions through
20 reaction cycles. The main advantages of this new process are
safety, minimal waste, and potentially low cost.
Introduction
Fluorinated aromatic compounds are valuable synthetic
intermediates for the preparation of pharmaceutical drugs, agro-
chemicals, and polymers.1 However, for all practical purposes,
the only way that fluorine is currently incorporated into a
benzene ring commercially is the Balz-Schiemann reaction and
related processes involving diazonium ion chemistry2,3 or under
special circumstances by halex exchange of aromatic chlorine
by fluorine.4 The diazonium processes exhibit poor atom
economy and generate considerable waste. In the current study,
we have sought to identify and explore new, low-cost, and
environmentally benign processes for replacing aromatic hy-
drogen with fluorine.
There are already a number of fluorinating agents known
that will either electrophilically or oxidatively replace aromatic
hydrogen with fluorine,5–8 although few of them are effective
in converting relatively unreactive benzene to fluorobenzene.
These include reagents such as XeF2,6,9,10 CF3OF,7,11 CH3CO2F,7,12
* To whom correspondence should be addressed. E-mail: wrd@chem.ufl.edu.
(1) Hewitt, C. D.; Silvester, M. J. Aldrichimica Acta 1988, 21, 3.
(2) Moilliet, J. S. In Organofluorine Chemistry: Principles and Com-
mercial Applications; Banks, R. E., Smart, B. E., Tatlow, J. C., Eds.;
Plenum: New York, 1994; pp 195–220..
(3) Boudakian, M. M. In Chemistry of Organic Fluorine Compounds II.
A Critical ReView; Hudlicky, M., Pavlath, A. E., Eds.; American
Chenmical Society: Washington, DC, 1995; pp 271–296..
(4) Hudlicky, M. In Chemistry of Organic Fluorine Compounds. A
Laboratory Manual with ComprehensiVe Literature CoVerage; John
Wiley & Sons, Inc.: New York, 1976; pp 112–127..
(5) Purrington, S. T.; Kagen, B. S.; Patrick, T. B. Chem. ReV. 1986, 86,
997–1018.
(13) Ip, D. P.; Authur, C. D.; Winans, R. E.; Appelman, E. H. J. Am. Chem.
Soc. 1981, 103, 1964–1968.
(14) Singh, S.; DesMarteau, D. D.; Zuberi, S. S.; Witz, M.; Huang, H.-N.
J. Am. Chem. Soc. 1987, 109, 7194–7196.
(15) Banks, R. E.; Khazaei, A. J. Fluorine Chem. 1990, 46, 297–305.
(16) Umemoto, T.; Fukami, S.; Tomizawa, G.; Harasawa, K.; Kawada, K.;
Tomita, K. J. Am. Chem. Soc. 1990, 112, 8563–8575.
(17) Nyffeler, P. T.; Duron, S. G.; Burkart, M. D.; Vincent, S. P.; Wong,
C.-H. Angew Chem., Int. Ed. 2005, 44, 192–212.
(6) Wilkinson, J. A. Chem. ReV. 1992, 92, 505–519.
(7) Rozen, S. Chem. ReV. 1996, 96, 1717–1736.
(8) Lal, G. S.; Pez, G. P.; Syvret, R. G. Chem. ReV. 1996, 96, 1737–
1755.
(18) Stacey, M.; Tatlow, J. C. In AdVances in Fluorine Chemistry; Stacey,
M., Tatlow, J. C., Sharpe, A. G., Eds.; Academic Press: New York,
1960; Vol. 1, pp 166–198..
(9) Anand, S. P.; Quarterman, L. A.; Pchristian, P. A.; Hyman, H. H.;
Filler, R. J. Org. Chem. 1975, 40, 3796–3797.
(10) Zupan, M.; Pollack, A. J. Org. Chem. 1975, 40, 3794–3796.
(11) Fifolt, M. J.; Olczak, R. T.; Mundhenke, R. F.; Bieron, J. F. J. Org.
Chem. 1985, 50, 4576–4582.
(19) Banks, R. E.; Tatlow, J. C. J. Fluorine Chem. 1986, 33, 227–346.
(20) Bailey, J.; Plevey, R. G.; Tatlow, J. C. J Fluorine Chem. 1988, 39,
23–38.
(12) Visser, G. W.; Bakker, C. N. M.; Halteren, B. W. v.; Herscheid,
J. D. M.; Brinkman, G. A.; Hoekstra, A. J. Org. Chem. 1986, 51,
1886–1889.
(21) Hudlicky, M. In Chemistry of Organic Fluorine Compounds II. A
Critical ReView; Hudlicky, M., Pavlath, A. E., Eds.; American
Chemical Society: Washington, DC, 1995; pp 120–132..
10.1021/op700266y CCC: $40.75
Published on Web 02/21/2008
2008 American Chemical Society
Vol. 12, No. 2, 2008 / Organic Process Research & Development
•
349

Manzer.22 However, only modest yields of fluorobenzene were
obtained when benzene was exposed to pure CuF2 in the gas
phase at 500 °C.23
Scheme 1. First gas-phase reaction of CuF2 with benzene
Scheme 2. New process of oxidative fluorination of benzene
In subsequent work on the use of CuF2, which was a
continuation of the work reported by Subramanian and Manzer,
the fluorination of aromatics was examined using a physical
mixture of commercial fluorides, CuF2 with AlF3, MgF2, or
CaF2. Although significantly increased yields were observed
(∼48%),24 regeneration, in particular multiple regeneration of
the spent reagents, could not be effectively accomplished. Both
oxidation of Cu to CuO at 350 °C and conversion of the oxide
to CuF2 by treatment by anhydrous HF at 400 °C are known,
and such processes are dependent upon particle size.25
Typically commercial metal fluorides have a tendency to
form regular crystallites as a result of their high lattice energies.
Therefore, upon thermal treatment, metal fluoride surface areas
are on the order of 10–60 m2/g, comparable to aluminum
fluorides. The fluorination ability of CuF2 depends significantly
upon its surface area and the resistance of Cu to sintering under
high temperature reaction conditions. Normally commercial
metal fluorides are treated at high temperature and possess
thermally stabilized structures. Also, mechanical mixing of
commercial fluorides coupled with high temperature treatment
may lead to loss of fluorine content. In contrast, one can consider
the Tanabe model26 for binary metal fluoride systems, where
acid generation is caused by an excess of a positive or negative
charge in the model structure of a binary fluoride. According
to this model, when a divalent metal ion (guest) is doped with
a trivalent metal ion (host), Brønsted acidic centers will be
formed. Brønsted acid sites can play a key role in fluorination
reactions.27 The current project had as its goal the preparation
of a binary fluoride system comprising CuF2 and AlF3.
Scheme 3. Possible mechanism of oxidative fluorination of
benzene by CuF2
with electron-rich benzene derivatives, such as N,N-dimethyl-
aniline, such dissociation occurs with electron transfer to form
Cu0 plus the aromatic radical cation.29 Although there have been
no similar published calculations (or experiments) related to
Cu(II) π-complexes with benzene, one can be confident that
such π-complexes exist (with fluoride-bound Cu(II), the com-
Results and Discussion
As a result of this study, we would like to report the
discovery of a metal fluoride reagent of the nominal composition
CuAl2F8 that has proved very effective for carrying out direct,
gas-phase, oxidative fluorination of benzene and other aromatics
and which is highly regenerable. Using this CuAl2F8 reagent,
conversions of benzene as high as 70% have been obtained,
and multiple recycling of the reagent, in situ, using O2 and AHF
has been demonstrated. Process details and scope of the reaction
are discussed below.
There are few experimental or theoretical data in the literature
that might provide insight into the possible mechanism of this
oxidative fluorination process. Because we detect HF as a
coproduct in the benzene/CuF2 reaction, we presume the
stoichiometry of the reaction to be as shown. Cu(I) is known
to bind to benzene to form π-complexes that dissociate in the
gas phase, exhibiting about a 50 kcal/mol BDE.28 In complexes
2
6
plex could be η rather than η),30 and because of the larger
oxidation potential of Cu(II), it is possible that dissociation of
this complex would occur with electron transfer to form the
benzene radical cation. Thus we could propose an overall
mechanism similar to that depicted in Scheme 3. There are, of
course, mechanisms other than the simple “oxidative fluorination
process” of Scheme 3, but until evidence is acquired relevant
to this issue, this mechanism should suffice as a working
hypothesis.
Preparation of the Cu-Al Fluoride Reagents. Four solids
of different composition were prepared by using CuF2/AlF3
molar ratios of 1:1, 1:1.5, 1:2, and 1:2.5. Appropriate amounts
of Al(NO3)3 ·9H2O and Cu(NO3)2 ·xH2O were dissolved in
ethanol. The mixture was then added dropwise to a stirred 48%
hydrofluoric acid solution. After stirring for 2 h, the precipitate
was filtered, washed with a small amount of distilled water/
ethanol, dried at 120 °C overnight (16 h), and calcined at 450
°C for 6 h under Ar flow.
(22) Subramanian, M. A.; Manzer, L. E. Science 2002, 297, 1665.
(23) Subramanian, M. A. U.S. Patent 6,166,273, 2000.
(24) Dolbier, W. R., Jr.; Gopal, B. U.S. Patent 7,012,165, 2006.
(25) Moss, J. H.; Ottie, R.; Wilford, J. B. J. Fluorine Chem. 1975, 6, 493–
416.
(26) Tanabe, K. Bull. Chem. Soc. Jpn. 2002, 47, 1064.
(27) Kemnitz, E.; Zhu, Y.; Adamczyk, B. J. Fluorine Chem. 2002, 114,
163–170.
(29) Ruan, C.; Yang, Z.; Rodgers, M. T. Phys. Chem. Chem. Phys. 2007,
9, 5902–5918.
(28) Bauschlicher, C. W., Jr.; Partridge, H.; Langhoff, S. R. J. Phys. Chem.
1992, 96, 3273–3278.
(30) Zhang, S.-L.; Liu, L.; Fu, Y.; Guo, Q.-X. J. Mol. Struct. (Theochem)
2005, 757, 37–46.
350
•
Vol. 12, No. 2, 2008 / Organic Process Research & Development

Figure 1. Schematic diagram of reaction/recycle setup.
Figure 3. Performance of copper aluminum fluoride in time
on stream analysis; 2.5 g of CuAl2F8 is reacted with 0.3 mL of
benzene in each experiment. Ar flow was 30 mL.
Figure 2. Conversion of benzene to fluorobenzene using
different metal fluorides. The y-axis is the yield of fluorobenzene
(%).
Table 1. Effect of Cu to Al ratio upon benzene conversiona
conversion of benzene (%)
CuF2/AlF3 ratio
initial activity
recycled activity
1:1
28
53
57
36
9
14
42
30
1:1.5
1:2
1:2.5
a Conditions: 2.5 g of reagent, 500 °C, 0.5 mL of benzene.
Typical Fluorination and Recycle Experimental Condi-
tions. The fluorination activities of these materials were
evaluated in a fixed bed reactor in the gas phase (Figure 1).
About 2.5 g of reagent was loaded into a hastalloy reactor tube
(8 mm o.d.) in a dry box between two plugs of silver wool.
The reactor was placed in an electrically heated furnace, and
the reagent was pretreated under a flow of argon gas (40 mL/
min) for 4 h at 500 °C. Benzene was introduced into the reactor
using a syringe pump (flow rate, 15 mL/min; volume, 0.5 mL)
along with the carrier gas (Ar, 25 mL/min) at 450–500 °C. The
product mixture coming out of the reactor was condensed in a
cold trap kept at 0 °C for 30 min for each pass of benzene, and
the outlet from this trap was connected to an alkali trap to
Figure 4. XRD patterns of copper aluminum fluoride: (A)
calcined at 500 °C; (B) reacted with benzene after 18 cycles;
(C) recycled with HF and O2 for 4 h after 20 cycles.
neutralize the biproduct HF. The composition of the product
1
mixture was determined by gas chromatography and H and
19F NMR. Fluorobenzene was the only product at 450 °C,
whereas small amounts of other products such as 1,3- and 1,2-
difluorobenzene were also obtained at 500 °C. Once the reagent
was expended (i.e., conversion of benzene fell to zero), the
Vol. 12, No. 2, 2008 / Organic Process Research & Development
•
351

Figure 5. SEM/EDS patterns of copper aluminum fluoride: (A) calcined at 500 °C; (B) reacted with benzene after 18 cycles; (C)
recycled with HF and O2 for 4 h after 20 cycles.
Table 2. Yield comparison for different Cu reagents
carrier gas was switched off, and the reactor was connected to
a flow of gaseous anhydrous hydrogen fluoride (8 mL/min) and
oxygen (10 mL/min) for 2.5 h. Then the reactor was flushed
with argon for 0.5 h, and the reaction cycle (fluorination of
benzene) was repeated without loss of actiVity in the fluorination
process.
352
•
Vol. 12, No. 2, 2008 / Organic Process Research & Development

Table 3. Yields of fluorotoluenes from chlorotoluenes
which generation of the CuF2 in situ from copper metal by
successive reaction with oxygen and hydrogen fluoride at 500
°C was followed by reaction with benzene, the cycle being
completed by reconversion of the copper metal residue back to
CuF2 followed by further reaction with benzene to obtain
virtually identical yields of fluorobenzenes in each cycle. The
binary fluoride system was found to be very stable through 20
recycled experiments, exhibiting constant activity of ∼45%
conversion with almost 80% selectivity towards monofluo-
robenzene (Figure 3).
X-ray diffraction patterns of the copper aluminum fluoride
reagent are shown in Figure 4. In the fresh sample (calcined at
450 °C under Ar for 4 h), before reacting with benzene,
crystalline phases due to AlF3 are mainly seen along with very
weak signals due to CuF2, which indicates that the CuF2 is
mostly dispersed within the AlF3. In the XRD pattern of the
spent system (after reacting with benzene), strong crystalline
signals due to Cu(0) appear as a result of the reaction of CuF2
with benzene to form fluorobenzene. Crystalline signals of CuF2
and loss of Cu(0) signals in the XRD pattern of the recycled
material can be seen. This indicates that CuF2 has been
regenerated by treatment with oxygen and HF.
The copper aluminum fluoride has also been characterized
by scanning electron microscopy (SEM) and energy dispersive
spectrometry. A fresh sample (before reaction with benzene)
exhibits uniform morphology and the presence of a high
concentration of fluoride on the surface (Figure 5a). In the
partially spent binary fluoride material (after limited reaction
with benzene), a fine distribution of copper can be seen (Figure
5b). This indicates that the copper is not sintered or agglomer-
ated even at the high reaction temperature. Thus, as mentioned
above, it has apparently been possible to prevent copper
sintering by use of the highly dispersed binary reagent when
using limited amounts of benzene, under which conditions the
Cu(II) is not exhaustively reduced. In the regenerated binary
fluoride material, after 20 cycles, surface analysis shows a
uniform morphology with a high concentration of fluoride
(Figure 5c).
This new copper aluminum fluoride reagent has also been
found to be active in fluorinating monofluorobenzene and
pyridine (Table 2), with 1,3-difluorobenzene and 2-fluoropy-
ridine being obtained as major products in those reactions,
respectively.
In addition, the reagent of nominal composition CuAl2F8 has
been found to be effective in converting chloroaromatics to
fluoroaromatics. Results for the reactions of chlorobenzene,
dichlorobenzenes, chlorotoluenes and chlorofluorobenzenes are
summarized in Tables 3 and 4, with the fluorination of
1-chloronaphthalene being shown in Scheme 4.
Other than results presented in our recent related patent,
which claimed low conversions of chloroaromatics to fluoro-
aromatics using either CuF2 (500 °C) or AgF (350 °C), there
appear to be no reports of fluorodechlorination of chloro
aromatics, other than classic SNAr reactions, such as the reaction
of p-chloronitrobenzene with fluoride ion.31 On the other hand,
cuprous salts (sometimes with Cu(II) catalysis) have long been
known to promote replacement of unactivated aryl halogens
yield of products (%)
reactanta
o-fluorotoluene m-fluorotoluene p-fluorotoluene
o-chlorotoluene
m- chlorotoluene
p-chlorotoluene
60
0
5
7
61
10
4
2
71
a Conditions: 0.5 mL of reactant, 2.5 g of catalyst, at 500 °C.
Using this same procedure, a series of supported and
unsupported CuF2 systems were evaluated with respect to the
fluorination of benzene. Pure copper(II) fluoride gave only low
conversion (11%) to fluorobenzene, whereas among the sup-
ported, physically mixed systems, CuF2/AlF3 was the most
effective towards fluorination as compared to MgF2, CaF2, and
AgF, giving rise to 41% conversion at 500 °C (Figure 2). All
such metal fluorides were commercial samples (Aldrich).
Among the coprecipitated binary fluoride systems, CuF2/AlF3
at a ratio of 1:2 was found to exhibit optimal fluorinating ability
in the fluorination of benzene (Table 1).
Regarding benzene fluorination, AlF3 was found to be totally
inactive, which indicates that it acts only as a support to disperse
CuF2. Comparing the coprecipitated reagent of nominal com-
position CuAl2F8 with the CuF2/AlF3 mixture, the binary
fluoride exhibits greater reactivity (56%) compared to that of
the physically mixed, commercial CuF2 and AlF3 under identical
conditions (Table 1). We believe that the observed high activity
of the binary fluoride might be due to the CuF2 being highly
dispersed within the AlF3 support. At the present time, we
cannot be certain whether the fluorinating reagent constitutes a
separate CuAl2F8 phase or whether the synthesis conditions
simply provide a good basis for forming a very amorphous,
finely distributed CuF2 reagent.
Fluorination of aromatics using inorganic metal fluorides in
the gas phase is a “solid-gas” reaction. Normally in solid–gas
reactions, once the surface CuF2 has become reduced to Cu(0),
sintering of copper can begin to occur at the high temperature
of reaction. As a result of the process of coprecipitating CuF2
with AlF3 followed by heating and treatment with HF, it has
been possible to obtain a high dispersion of CuF2 within the
AlF3 support, such that sintering can be minimized, if not
prevented, even at high temperature. By selecting aluminum
fluoride, an acidic (Lewis acid) support, to deposit copper
fluoride, it has been possible to obtain highly dispersed copper
fluoride species on aluminum fluoride.
The copper aluminum fluoride systems (both the commercial
and coprecipitated) were tested in multiple cycles, in the
oxyfluorination of benzene. After the initial activity experiment
(single pass of benzene), the carrier gas was switched off, and
the reactor was connected to the flow of anhydrous hydrogen
fluoride (8 mL/min) and oxygen gases (10 mL/min) for 2.5 h.
Then the reactor was flushed with argon for 0.5 h, and the
reaction cycle (fluorination of benzene) was repeated. The
coprecipitated copper aluminum fluoride remained very ac-
tive and stable during multiple cycles, whereas when using the
commercial metal fluoride mixture, the benzene conversion fell
to zero within 3 cycles.
Thus the benzene-CuF2 (copper aluminum fluoride) reaction
has been demonstrated to be a two-phase, cyclic process in
(31) Finger, G. C.; Kruse, C. W. J. Am. Chem. Soc. 1956, 78, 6034–6037.
Vol. 12, No. 2, 2008 / Organic Process Research & Development
•
353

Table 4. Yields of fluorobenzenes from chlorobenzenes
Scheme 4. Fluorination of 1-chloronaphthalene
Scheme 5. CuCl reaction with aryl bromides
reaction is still in its preliminary stages, and thus we will refrain
from speculating about a possible mechanism.
Conclusions
The newly synthesized copper aluminum fluoride of nominal
composition CuAl2F8 exhibits improved activity towards fluo-
rination of aromatic compounds in comparison to the physical
mixture of commercial CuF2 and AlF3. It has also been shown
to maintain reactivity through 20 reaction cycles of regeneration.
The high fluorination ability and recyclability of the binary
fluoride might be due to the high dispersion of CuF2 on AlF3,
which combined with the use of limited amounts of benzene
in each run seems to inhibit sintering of copper.
with nucleophiles,32 the best known example being the conver-
sion of aryl bromides to aryl nitriles through reaction with
CuCN.33,34 Less known is the little studied conversion of aryl
bromides to chlorides through treatment with CuCl,35,36 which
is shown in Scheme 5.
To our knowledge, however, there have been no previous
reports of the use of a transition metal fluoride to convert a
nonactivated aryl chloride to an aryl fluoride. As one can see
from the tables, the reaction is not regiospecific and thus cannot
involve a simple Cl-F replacement process. The study of this
The main advantages of this new process are (1) safety; the
process design avoids the possibility of runaway reactions that
can be encountered when using the diazonium ion based
synthesis, and (2) minimal waste; with virtually no waste, the
new process is extremely environmentally friendly. When this
is combined with the ability to recycle the Cu reagent, the cost
of the process should also be very attractive.
Acknowledgment
Support of this research in part by DuPont is acknowledged
with thanks. We also wish to thank Dr. Ed Mahler for his very
helpful insights regarding the recycling process.
(32) Bacon, R. G. R.; Hill, H. A. O. Quart. ReV. 1965, 19, 95–125.
(33) Koelsch, C. F.; Whitney, A. G. J. Org. Chem. 1941, 6, 795–803.
(34) Newman, M. S.; Boden, H. J. Org. Chem. 1961, 26, 2525–2525.
(35) Hardy, W. B.; Fortenbaugh, R. B. J. Am. Chem. Soc. 1958, 80, 1716–
1718.
Received for review November 20, 2007.
OP700266Y
(36) Bacon, R. G. R.; Hill, H. A. O. J Chem. Soc. 1964, 1097–1107.
354
•
Vol. 12, No. 2, 2008 / Organic Process Research & Development