Journal of Fluorine Chemistry 123 (2003) 3136

Review

Can fluorine chemistry be green chemistry?

Stewart J. Tavener*, James H. Clark
Department of Chemistry, Clean Technology Centre, University of York, Heslington, York YO10 5DD, UK
Received 17 April 2003; received in revised form 15 May 2003; accepted 19 May 2003
Dedicated to Professor R.E. (Eric) Banks on the occasion of his 70th birthday and in recognition of his many contributions to the fluorine literature

Abstract
The roles of the element fluorine and its compounds in relationship to green chemistry and clean chemical manufacturing are considered.
# 2003 Elsevier B.V. All rights reserved.
Keywords: Fluorous biphase; Microreactors; Fluorobenzene; Fluoroaromatic

1. Introduction

2. Some problems with fluorine chemistry

The main tenet of clean chemical synthesis, or green

chemistry, is that a highly efficient chemical reaction which
produces little or no waste is greatly preferable to treatment
or recycling of effluent (see, for example [1]). In general, this
means ensuring that as much of the substrate and reagents as
possible find their way into the final product, and that the use
of auxiliary compounds such as solvents and promoters is
minimised or eliminated. Frequently, the conclusions are the
same as those that would be reached from an economic point
of view; less substrate, fewer reagents, less solvent and a
consequent reduction in energy requirement will all contribute to financial savings.
Fluorine chemistry has an important role to play in clean
technology, both in catalyst and solvent replacement technologies. Fluorine is a very light element and provides
excellent value in terms of activity-per-gram. It is also recognised that fluorochemicals are frequently more effective
and are required in smaller quantities than non-fluorinated
compoundsin the case of fluorine, less is more. High
effect, low dosage fluorinated formulations allow reductions
in the use of materials and energy all along the production
chain: less non-renewable feedstocks are consumed, transport
and packaging is reduced, and the quantity of post-consumer
waste is minimised. However, fluorochemicals have also been
involved in some of the more negative aspects of the chemical
industry. In this review article, we discuss some of the ways in
which fluorine is involved in green chemistry.

One aspect of the chemical industry that has been subject

to major change in recent years is the manufacture of halogenated organic molecules. Chlorofluorocarbons (CFCs),
which were identified as having a significant contribution to
the destruction of stratospheric ozone, through liberation of
chlorine radicals via photolysis, were phased out as a result
of the 1987 Montreal Protocol [2]. HCFCs were introduced
as interim replacements for CFCs in refrigeration and
blowing agent applications, and these are now being replaced by hydrofluorocarbons (HFCs). Importantly, HFCs
are chlorine-free and are believed to have no significant
contribution to ozone destruction. However, they may still
act as greenhouse gases [3]. Of course, manufacture of these
chlorine-free replacements requires larger quantities of
fluorine than their predecessors, and manufacture of HFCs,
HCFCs, fluoropolymers and other fluorocarbon compounds
is one of the largest uses of fluorine [4].
Unfortunately, HFCs and perfluorocarbons are now also
coming under scrutiny because of their large global warming
potential (GWP). HFCs, and HFC134a in particular, have
been used since the early 1990s as replacements for CFCs in
vehicle air conditioning units, and emissions from these
systems are projected to contribute the equivalent of 50 Mt
of CO2 per annum by 2010. The EU Environment Commission is considering legislation to tackle this in the near future,
although it is not yet clear what alternatives are available at
acceptable cost [5]. A complete phase-out of HFCs and PFCs
is already being discussed in legislative circles [6], which
could be counter-productive, preventing the growth of fluorous phase chemistry and supercritical solvent systems which
have great potential as clean methods for chemical synthesis.

Corresponding author. Tel.: 44-1904-434473;

fax: 44-1904-434450.
E-mail address: sjt4@york.ac.uk (S.J. Tavener).
0022-1139/$ see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-1139(03)00140-4

32

S.J. Tavener, J.H. Clark / Journal of Fluorine Chemistry 123 (2003) 3136

3. The role of fluorine compounds in clean synthesis

Table 1
Solubility of oxygen in selected perfluorinated solvents [8]

3.1. Fluorous solvent technologies

Solvent

Solubility of O2 (ml)
per 100 ml solvent

n-Perfluorooctane (C8F18)
Perfluorotributylamine ((C4F9)3N)
Perfluorooctyl bromide (C8F17Br)

52.1
38.4
52.7

It would be unfortunate if organofluorine compounds

were to become the subject of such comprehensive restrictions that now affect chlorine-containing compounds, as the
unique properties of fluorine give it an important role to play
in the advancement of clean chemical synthesis. In particular, the immiscibility of compounds containing perfluoroalkyl groups with hydrocarbon and polar organic solvents
has led to the development of fluorous biphasic chemistry,
and other fluorous solvent technologies, in which the ease of
separation of product or catalyst is greatly improved. This
gives better efficiency and reduces waste. Fig. 1 shows how
fluorous liquids may be utilised in a biphasic system to
facilitate recycling of a catalyst and solvent: at low temperatures the phases are immiscible, but on heating become
a single phase allowing reaction to proceed at homogeneous
rates. Cooling causes phase separation and allows separation
of solvent and catalyst from the product.
Many laboratory and even some industrial scale oxidations were historically performed using stoichiometric,
toxic, metal-based oxidants such as KMnO4 and K2Cr2O7
[1,7]. However, the use of small-molecule sources of oxygen
such as air, O2 and H2O2 with an additional metal catalyst if
required, is preferable on both economic and environmental
grounds. An oxidation reaction will typically give products
of greater polarity than the substrates, and these changes in
polarity make oxidation reactions very suitable candidates
for fluorous biphasic reactions. For example, if an alkene is
oxidised to a more polar product (e.g. an epoxide or diol) in a
fluorous biphase, the product should be less soluble in the
fluorous phase than the substrate, and this should lead to
improved separation. In addition, oxygen has a remarkably
high solubility in perfluorinated solvents (Table 1) and so the
fluorous biphase is an ideal medium in which to conduct
oxidation reactions using elemental oxygen.
For example, the oxidation of a range of primary alcohols
to the corresponding aldehydes may be conducted under
fluorous biphasic conditions, using a perfluoroalkylated
bipyridine as ligand to ensure that the catalyst remains in
the fluorous phase [9] (Scheme 1). The catalyst may be

C 8 F18 , C6 H 5 Cl
90 o C, O2

OH
O2N

O
O 2N

catalyst

93%

(CH 2 )4C 8 F17

catalyst
Br

N
Cu

Me 2 S

(CH 2 )4C 8 F17

Scheme 1. Alcohol oxidation under fluorous conditions.

successfully recycled by phase separation, with analytically

pure products being isolated even after eight cycles.
Similarly, aldehydes may be converted to carboxylic acids
using a fluourous Ni complex as the oxidant, as shown in
Scheme 2 [10]. This system is also active for the oxidation of
sulfides to sulfoxides and sulfones.
The fluorous approach is not limited to oxidations using
elemental oxygen. A fluorous analogue of DMSO has been
used to perform Swern reactions [11], a widely used method
of oxidising an alcohol to an aldehyde which is unsatisfactory from a green chemistry point of view due to its production of stoichiometric quantities of dimethyl sulphide. Using
fluorous biphase methodology, this stoichiometric reaction
may be made pseudo-catalytic, as shown in Scheme 3. After
reaction, the sulphide is extracted into perfluorohexane and
recycled. However, the system is still not ideal as it requires
an additional chlorinated or aromatic solvent.

Ni(F-acac)2 , O2
o
PhCH 3 /F-decalin, 64 C

fluorous
solvent
& catalyst

heat
& stir

product
phases
mix

O
Cl

Cl
substrate

N O

OH

cool

87%
fluorous
solvent
& catalyst

Ni(F-acac)2
C 7 F15

F15 C7
O

O
Ni

O
recycle catalyst & fluorous solvent

Fig. 1. Schematic diagram of a fluorous biphase process in which a

fluorous solvent and fluorous catalyst are recycled together.

F15 C7

O
C 7 F15

Scheme 2. Aldehyde oxidation under fluorous conditions.

S.J. Tavener, J.H. Clark / Journal of Fluorine Chemistry 123 (2003) 3136
I

C 4F 9

NaBH 4
Me 2S2
S

C 4F 9

H 2 O2 , MeOH
C 4F 9

C 6 F14
OH

O
CH 2Cl2 or toluene

Scheme 3. A Swern oxidation under fluorous conditions allows reuse of

the sufoxide reagent.

33

been applied to separation of chiral mappicines [14], and to

separate and reuse metal catalysts with fluorous ligands [15].
However, whilst FRPSG may allow efficient recycling of
catalysts, this improvement must be offset against the
increased quantities of solvents and reagents that must be
used to prepare the fluorous silica and perform the chromatography.
Fluorous solvents have also recently found applications as
liquid membranes to limit the rate of addition of reagents
and so control exothermic reactions such as alkene bromination, and demethylation of anisoles by reaction with boron
tribromide [16]. This has potential as a clean route as it may
improve the selectivity of the system.
3.2. The use of fluorinated groups in supercritical CO2

Fluorous biphase reactions have been reviewed extensively in the past few years, and most important types of
reaction may now be conducted under fluorous conditions
[8,12]. However, partitioning of catalysts and reagents into
the fluorous phase is seldom perfecteven a loss of 12% of
an expensive catalyst may be unacceptableand the GWPs
and relatively high prices of the perfluorocarbon reagents
may ultimately limit their use to specialist applications.
Silica gel may be made compatible with fluorous reagents
and solvents by grafting a perfluoroalkyl group to the surface
using a silane reagent [13]this is a fluorous technology
which requires no perfluorinated solvent. The resultant solid,
know as fluorous reverse phase silica gel (FRPSG), may
then be used as a chromatography support for separation and
purification of fluorous reagents. Compounds without fluorous groups are eluted first using a polar solvent (in which
the fluorous compounds have little solubility), and then a
less polar solvent is used to elute the fluorous compound.
Whereas effective partitioning into a fluorous liquid phase
generally requires two or three perfluoroalkyl groups to be
attached, FRPSG techniques require much lower fluorine
content for separation to be effective. This approach has

The use of fluorous groups also finds application in

supercritical CO2 systems because of their solubility in
the supercritical phase. The cause of this compatibility is
not well understood, although there is some evidence that
specific interactions between the fluid and the fluorous
group may exist [17]. Supercritical CO2 is finding many
applications as a solvent in clean synthesis as it is non toxic
and is completely removed from the reaction product
simply by releasing the reactor pressure. Hydrogen peroxide, which produces only water and O2 as by-products
and is generally considered to be a clean oxidant, may be
generated in supercritical CO2. The anthraquinone (or AQ)
process, in which an alkyl anthraquinone is first hydrogenated and then oxidised, is used to supply almost all of the
global demand for H2O2 [18]. The AQ process may be
successfully performed in supercritical CO2 if the anthraquinone catalyst is made compatible with the fluid phase by
functionalisation with perfluorinated chains [19]. Moreover, the H2O2 produced in this way may be utilised in
the same reactor (i.e. a one-pot process), for the epoxidation
of alkenes (Scheme 4).

OH

F C 3 F6 O

OC3 F6 F
n
n
OH
O2
supercritical
CO 2

H2

H 2 O2

OC3F 6 F
n

C 3 F6 O
n
O

H2 O

Scheme 4. Generation of H2O2 in supercritical CO2 using a fluorous anthraquinone catalyst, and in situ use as an oxidant.

34

S.J. Tavener, J.H. Clark / Journal of Fluorine Chemistry 123 (2003) 3136

10% F2 in N2
OEt

99% conversion
73% selectivity

O
OEt
F

Scheme 5. Fluorination of ethyl acetoacetate with elemental fluorine using

a microreactor.

Supercritical fluid reactors are no longer merely laboratory

curiosities, and the compatibility of fluorinated compounds
with supercritical CO2 has facilitated the replacement of
CFCs as a medium for polymerisation of Teflon and other
fluoropolymers on an industrial scale. DuPont recently built a
US$ 275 million plant capable of making 1000 t of polymer
per annum. The plant uses carbon dioxide technology in a
process that generates less waste during manufacture, and
produces a grade of polymer that the manufacturer claims has
enhanced performance and processing capabilities compared
with that produced using existing technology [20].
3.3. Selective fluorination using microreactors
Fundamentally, there is no reason why elemental fluorine
should not be viewed as a suitable reagent for clean synthesis. Because of its reactivity, F2 may be used without
additional compounds such as catalysts, and may be used in
the gas phase without solvent. However, this supreme reactivity is itself a problem because of the large quantities of
heat liberated during the reaction, which can lead to
increased reaction rates, reduced selectivity and, in the worst
case scenario, a runaway reaction. Whilst this must be
avoided on safety grounds, the loss of selectivity is also
of concern from a green chemistry of view. Fluorinations
may be made significantly more selective simply by dilution
of F2 in an inert carrier gas (usually nitrogen) [21]. Microreactors potentially allow scale up of reactions simply by
running many tiny reactors in parallel, and could in the
future intensify processes and reduce the size of chemical
plants [22]. The extremely small reactor volume means that
heat may be removed from the reaction much more efficiently, moderating the exotherm and thus giving much
greater selectivitywhich is ideal for fluorinations, and
microreactor systems are already showing great potential
for selective reactions using elemental fluorine. For example, the fluorination of ethyl acetoacetate, using 10% F2 in
N2, in a microreactor gave 73% selectivity for the monofluorinated product, as shown in Scheme 5 [23], which
compares very favourable with the 15% conversion and
85% selectivity observed in a similar bulk reactor [24].
Fluorinations of toluene and other aromatics in microreactors have also been reported [25].

4. A case study: fluorination of aromatics

Fluoroaromatics are an important class of chemicals
with applications including pharmaceuticals, agricultural

chemicals, polymers and liquid crystals [26]. Out of several

methods for producing fluoroaromatics, four of the most
importantHalex, Balz-Schiemann, direct fluorination, and
a recently reported route involving CuF2 (Scheme 6)are
discussed here in the context of their cost efficiency, risk,
and other important criteria. An ideal synthesis for fluoroaromatics might involve a starting material from a sustainable source, a good fluorinating agent in terms of reactivity,
cost, ease of handling and safety, and would give high yields
with low levels of by-products.
The halex, or halogen exchange, route to fluoroaromatics
starts with a chloroaromatic substrate (or nitro group in
analogous fluorodenitration reactions), which is inherently
wasteful from an atom efficiency perspective, as the chlorine will not appear in the final product. A fluoride salt,
usually KF, and quaternary ammonium or phosphonium salt
catalyst are also required. Halex reactions work best in polar
aprotic solvents such as DMF and DMSO, which are toxic
and tend to break down under reaction conditions. The
quaternary salt is generally not reusable, and high temperatures are required.
The Balz-Schiemann route involves two steps, and
requires treatment with HCl, NaNO2 and HBF4, although
the nitrogen, chlorine and boron that appear in these reagents
are not found in the final fluorinated product. The aniline
substrate may already have produced substantial waste
during its manufacture which via nitration and subsequent
reduction. It is, however, applicable to a wider range of
substrates than some of the other methods, and remains a
common method for fluoroaromatic manufacture.
Direct fluorination is potentially very efficient in terms of
utilisation of reagents, requiring no existing functional
group: the CH is replaced directly with CF. The drawbacks are that the reaction is frequently unselective, giving
polysubstituted products (although see Section 3.3), energy
is required to cool the system, and acidic HF waste is
produced. Unlike chlorination and bromination reactions,
it is not possible to oxidise the HF back to F2 by in situ use of
H2O2the oxidation potential for F2 is too high. Whilst the
cleanest option is to recycle the HF back to F2 by electrolysis, it is frequently just neutralised and disposed of.
The recently reported oxidative fluorination route has
extremely high atom efficiency, incorporating all of the
fluorine into the aromatic, and producing only water as a
by-product [27]. Because it replaces an aromatic hydrogen
directly with a fluorine, no precursor group is required.
Copper(II) fluoride, used as the fluorinating agent, was
selected on the basis of examination of metal redox potentials. The metal must be sufficiently oxidising that it reacts
with the CH bond, yet not so oxidising that it is impossible
to regenerate. This system gives very good selectivity
(>95%) towards the monofluorinated aromatic product,
requires no solvent, and, at first glance, appears to be very
promising as a clean, atom efficient route to fluoroaromatics.
However, very high temperatures are required for both the
reaction (450550 8C) and regeneration (400 8C) steps, and

S.J. Tavener, J.H. Clark / Journal of Fluorine Chemistry 123 (2003) 3136

35

HALEX

BALZ-SCHIEMANN

NH2
HCl, NaNO2 , HBF 4

+ MF
R

catalyst,
solvent,
temperature

N2 BF4

MX

HCl, H2 O
R

N 2 , BF3
R
"F 2"

HF

HF + Cu

H
CuF2

HF + O2
R

R
OXIDATIVE
FLUORINATION

H 2O

DIRECT
FLUORINATION

Scheme 6. Summary of four common routes for preparing fluoroaromatics (see also Table 2).

Table 2
Overall evaluation of manufacturing methods for fluoroaromatics
Route

Waste

Efficiency

Risk

Energy and materials

Cost

Non-renewables

Simplicity

Halex

Salt lost or
decomposed
solvent
Salt lost solvent
BF3
Salt
(neutralised HF)
Water only

High selectivity

Corrosion,
decomposition of
(toxic) solvent
Unstable
intermediate HF
F2, HF

High temperature

Activated aromatics,
expensive solvents

Fluorides,
aromatics

Cooling special
reactors
Cooling highly
specialised reactors
Very high temperatures
special reactors

Anilines

One step (not counting

chloroaromatic
manufacture)
Two steps (not counting
aniline manufacture)
One step

Fluorides,
aromatics
Fluorides,
aromatics
Fluorides, copper, One step with recycling
aromatics
of fluoride

Balz-Schiemann
Direct fluorination
Oxidative
fluorination

High selectivity
Difficult to make
selective
Largely unproven

HF

only low conversions (530%) are observed. The high

temperatures mean that whilst this may well be a suitable
method for the fluorination of benzene and simple alkylbenzenes, few functional groups would survive these conditions. In addition, any advantages gained in terms of atom
efficiency must be offset against the energy required to heat
the reactor.
There are many factors which contribute to the greenness of a chemical reaction, and there is no clear winner in
this contest, although the two systems that replace CH
directly are, on paper at least, cleaner than those that require

F2 generation and
handling
Energy costs and
HF handling

functional group exchange. These reactions are summarised

in Table 2.

5. Conclusions
Fluorine-containing compounds are found widely in the
environment, and although some of these may be attributed to the activities of human beings in general, and the
chemical industry in particular, their are also many natural
processes contribute to fluorine reservoirs. Although certain

36

S.J. Tavener, J.H. Clark / Journal of Fluorine Chemistry 123 (2003) 3136

fluorine-containing compounds have been identified as

being involved in ozone depletion and global warming
effects, legislative measures have been, or are being, put
in place where necessary to reduce this impact. Fluorine
containing compounds have a great potential for clean
synthesis, and utilisation of the high activity that fluorinated
groups can impart molecules may help to reduce the quantities required of certain classes of chemicals that are used in
the biosphere, including pesticides and pharmaceuticals.
References
[1] J.H. Clark, Chem. Br. (1998) 43;
J.H. Clark, Green Chem. 1 (1999) 1;
M. Lancaster, Green Chemistry: An Introductory Text, RSC,
Cambridge, 2002.
[2] A. McCulloch, J. Fluorine Chem. 100 (1999) 163.
[3] Inventory of US Greenhouse Gas Emissions and Sinks 19902000,
US Environmental Protection Agency, Washington, 2002.
[4] M.M. Miller, US Geological Survey Minerals Yearbook 2001:
Fluorspar Report;
M.M. Miller, US Geological Survey Mineral Commodity Summaries, January 2002.
[5] Environmental Data Service, ENDS Rep. 337 (2003) 46.
[6] Environmental Data Service, ENDS Rep. 328 (2002) 50.
[7] R.A. Sheldon, J. Chem. Technol. Biotechnol. 68 (1997) 381.
[8] D.J. Adams, P.J. Dyson, S.J. Tavener, Chemistry in Alternative
Reaction Media, Wiley, Chichester, 2003.
[9] B. Betzemeier, M. Cavazzini, S. Quici, P. Knochel, Tetrahedron Lett.
41 (2000) 4343.
[10] I. Klement, H. Leutjens, P. Knochel, Angew. Chem. Int. Ed. Engl. 36
(1997) 1454.

[11] D. Crich, S. Neelamkavil, J. Am. Chem. Soc. 123 (2001) 7449;

D. Crich, S. Neelamkavil, Tetrahedron 58 (2002) 3865.
[12] A.P. Dobbs, M.R. Kimberley, J. Fluorine Chem. 118 (2003) 3;
E.G. Hope, A.M. Stuart, J. Fluorine Chem. 100 (1999) 75;
D.P. Curran, Z. Lee, Green Chem. 3 (2001) G3.
[13] D.P. Curran, Synlett 9 (2001) 1488.
[14] D.P. Curran, Z. Lee, Green Chem. 3 (2001) G3.
[15] B. Croxtall, E.G. Hope, A.M. Stuart, FBC7 in Abstracts of the RSC
Annual Conference, 2001.
[16] I. Ryu, H. Matsubara, S. Yasuda, H. Nakamura, D.P. Curran, J. Am.
Chem. Soc. 124 (2002) 12946.
[17] A. Dardin, J.B. Cain, J.M. DeSimone, C.S. Johnson, E.T. Sanulski,
Macromolecules 30 (1997) 3592.
[18] W.R. Sanderson, in: J.H. Clark, D.J. Macquarrie (Eds.), Handbook
of Green Chemistry and Technology, Blackwell, Oxford, 2002,
p. 256.
[19] D. Ha ncu, J. Green, E.J. Beckman, Acc. Chem. Res. 35 (2002) 757.
[20] M. McCoy, Business 77 (1999) 11;
M. McCoy, Chem. Eng. News (1999) 11;
M. McCoy, Chem. Week, 27 March 2002 (unattributed article).
[21] G. Sandford, Tetrahedron 59 (2003) 437.
[22] P.D.I. Fletcher, S.J. Haswell, E. Pompo-Villar, B.H. Warrington, P.
Watts, S.Y.F. Yong, X. Zhang, Tetrahedron 58 (2002) 4735;
S.J. Haswell, R.J. Middleton, B. OSullivan, V. Skelton, P. Watts, P.
Styring, Chem. Commun. (2001) 391.
[23] R.D. Chambers, R.C.H. Spink, Chem. Commun. (1999) 833;
R.D. Chambers, D. Holling, R.C.H. Spink, G. Sandford, Lab. Chip 1
(2001) 132.
[24] R.D. Chambers, M.P. Greenhall, J. Hutchinson, Tetrahedron 52
(1996) 1.
[25] N. de Mas, A. Gu nther, M.A. Schmidt, K.F. Jensen, Ind. Eng. Chem.
Res. 42 (2003) 698.
[26] J.H. Clark, D. Wails, T.W. Bastock, Aromatic Fluorination, CRC
Press, Boca Raton, 1996.
[27] M.A. Subramanian, L.E. Manzer, Science 297 (2002) 1665.