An efficient and environmentally friendly synthesis of the inhalation anesthetic sevoflurane

Article abstract of DOI:10.1016/S0022-1139(00)00316-X

We report a new, high yield, single vessel synthesis of the general anesthetic sevoflurane. The new synthesis consists of a two-stage fluoromethylation of hexafluoroisopropanol. In the first stage, hexafluoroisopropanol is chloromethylated using aluminum trichloride and trioxane. In the second stage fluorine exchange is carried out using potassium fluoride and poly(ethylene glycol). Sevoflurane is distilled directly from the reaction vessel yielding material of 99.95% purity. The overall yield of the process is 65-70%.

Full text of DOI:10.1016/S0022-1139(00)00316-X

Journal of Fluorine Chemistry 106 (2000) 99±102
An ef®cient and environmentally friendly synthesis of the
inhalation anesthetic sevo¯urane
Christopher Bieniarz*, Chris Behme, Kornepati Ramakrishna
Advanced Drug Delivery, Hospital Products Division, D-97D, AP4, Abbott Laboratories, Abbott Park, IL 60064-3500, USA
Accepted 19 June 2000
Abstract
We report a new, high yield, single vessel synthesis of the general anesthetic sevo¯urane. The new synthesis consists of a two-stage
¯uoromethylation of hexa¯uoroisopropanol. In the ®rst stage, hexa¯uoroisopropanol is chloromethylated using aluminum trichloride and
trioxane. In the second stage ¯uorine exchange is carried out using potassium ¯uoride and poly(ethylene glycol). Sevo¯urane is distilled
directly from the reaction vessel yielding material of 99.95% purity. The overall yield of the process is 65±70%. # 2000 Elsevier Science
S.A. All rights reserved.
Keywords: Anesthetic; Sevo¯urane; Fluoromethyl ether; Aluminum trichloride; Chloromethylation; Fluoromethylation
1. Introduction
the yield is reported to exceed 70%, the crude product
requires extensive puri®cation to remove side-products
including bis¯uoromethyl ether, various chain length poly-
acetals, and toxic ¯uoromethyl-2,2-di¯uoro-1-(tri¯uoro-
methyl) vinyl ether. The use of toxic and caustic
reagents, particularly hydrogen ¯uoride gas, make this
synthesis environmentally unfriendly and complicates large
scale manufacture. The drawbacks of the current synthesis
prompted us to devise a new synthetic route to sevo¯urane
with several goals in mind. The process should be envir-
onmentally safe, without hazardous reagents or toxic by-
products, cost-effective, and manufacturable in industrial
quantities.
In this note we describe a new synthetic route to sevo-
¯urane which meets these goals. The synthesis involves a
novel method for the chloromethylation of HFIP followed
by ¯uorination of the intermediate chloromethyl ether. This
two-step procedure can be run in a single vessel and affords
high yields of very pure sevo¯urane.
Sevo¯urane,
(1),
[1,1,1,3,3,3-hexa¯uoro-2-(¯uoro-
methoxy) propane] is today one of the most important
and widely used general anesthetics. Sero¯urone anesthetic
combines various characteristics that are most desirable in
an inhalation anesthetic including the lowest blood/gas
partition coef®cient of 0.63, smooth induction and recovery
from anesthesia, minimal irritation to the upper respiratory
tract, low metabolic rate and rapid elimination. In addition,
sevo¯urane is suitable for out-patient surgery use.
Although the de®nitive mechanism of action has not been
elucidated, it has recently been shown that sevo¯urane
interacts with nicotinic acetylcholine receptors, by affecting
the open and closed state of the ion channels at clinical and
lower concentrations [1,2]. Sevo¯urane may also effect
reversible modulation of GABA and glycine receptors as
recently reported [3]. The above suggests that at least part of
the anesthetic action of sevo¯urane may be due to interac-
tions between sevo¯urane and speci®c voltage-gated ion
channels.
Sevo¯urane (1) is currently synthesized (Scheme 1) by
reacting 1,1,1,3,3,3-hexa¯uoro-2-propanol (HFIP) with par-
aformaldehyde in the presence of a large molar excess of
hydrogen ¯uoride in fuming sulfuric acid [4,5]. Although
2. Results and discussion
Simply put, the synthesis of sevo¯urane involves ¯uor-
omethylation of HFIP. However, direct ¯uorination did not
appear possible without the use of hydrogen ¯uoride gas or
expensive ¯uorinating agents such as diethylaminosulfur
tri¯uoride, or XeF2 [6,7]. We therefore turned our attention
toward a two-step procedure, speci®cally a chloromethyla-
* Corresponding author. Tel.: ‡1-847-937-2239; fax: ‡1-847-938-3645.
E-mail address: bienica@hpd.abbott.com (C. Bieniarz).
0022-1139/00/$ ± see front matter # 2000 Elsevier Science S.A. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 0 0 ) 0 0 3 1 6 - X

100
C. Bieniarz et al. / Journal of Fluorine Chemistry 106 (2000) 99±102
Scheme 1.
Scheme 2.
tion of HFIP followed by ¯uorination of the intermediate
HFIP chloromethyl ether.
in PEG 400 and 4 equivalents of spray-dried KF were added.
The reaction mixture was then heated at 958C for 1±2 h.
Smooth halide exchange yielded sevo¯urane, which after
distillation gave a 71% yield of 99.95% pure anesthetic.
Other solvents, including DMF, NMP, and DMSO likewise
afforded the product, but in lower yields and product iso-
lation was dif®cult. The use of glyme or diglyme furnished
only traces of the desired product. Since the conversion of
the chloromethylated intermediate to sevo¯urane is essen-
tially quantitative as determined by GC±MS, the evaporative
loss of sevo¯urane, which boils at 598C, undoubtedly low-
ered the yield. Although we used spray-dried KF, less
expensive commercial grade KF gave a product of compar-
able purity and yield.
Our initial attempts to chloromethylate HFIP using tradi-
tional methods met with failure. Infusing a slurry of paraf-
ormaldehyde in HFIP with anhydrous HCl gas afforded a
complex mixture of polyacetals. The addition of Lewis acid
catalysts, i.e. AlCl3 or anhydrous SnCl4 and/or lowering the
reaction temperature had little effect on the product dis-
tribution. The use of other Lewis acid chlorides, including
BCl3 and FeCl3, gave complex mixtures of chloromethyl
polyacetals. Phosphorous trichloride furnished the chloro-
methyl ether in low (7%) yield in addition to numerous
polyacetals and unreacted HFIP. A successful high yielding
synthesis of the chloromethyl ether of HFIP was ®nally
achieved with the use of AlCl3 (Scheme 2A). When paraf-
ormaldehyde was added to a neat slurry of AlCl3 in HFIP, a
rapid exothermic reaction took place, generating the desired
chloromethyl ether 2 as the major product. Several mod-
i®cations of our initial attempts led to a facile, high yielding
protocol. In the optimized procedure, equimolar amounts of
HFIP and anhydrous AlCl3 were cooled to 08C for the
addition of 1/3 equivalent of trioxane. The latter reagent
gave a more controlled, less exothermic reaction than par-
aformaldehyde. The mixture was stirred overnight at ambi-
ent temperature, then carefully quenched with cold 6N HCl
to decompose the hydroxydichloroaluminate gel which
forms during the reaction. The addition of water caused
the mixture to separate into a clear aqueous phase containing
the dissolved low molecular weight aluminum salts and a
lower phase consisting of 95% pure HFIP chloromethyl
ether 2. In addition, 4.5% of the bis-acetal of HFIP (3)
and about 0.5% of higher molecular weight polyacetals were
present as side products.
The two component reactions of our procedure deserve
further comment. The chloromethylation protocol appears to
have little if any direct precedent in the chemical literature.
For example, AlCl3 has previously been used as a chlorinat-
ing agent in the preparation of b-chloroglycosides [8], and in
para-chloromethylation of thioanisole [9]. In our synthesis,
AlCl3 plays a triple role, as illustrated in Scheme 3. First, it
serves as a Lewis acid in the complexation to trioxane and
hydroxymethylation of HFIP, probably through intermediate
‡
formation of CH2O AlCl3 . Second, it functions as a
chlorinating agent. And third, it acts as a dehydrating agent,
a role which is ful®lled by fuming sulfuric acid in the
previous synthesis. Although we did not study the mechan-
ism of the reaction in detail, we depict the mechanism of
halogenation in Scheme 3 as SNi. Other mechanisms invol-
ving intermolecular chloride attack may also be operative.1
Alternatively, attack of a second molecule of HFIP on the
putative intermediate aluminate affords instead the bis-
HFIP-acetal. Indeed, increasing the ratio of AlCl3 to 2:1
After separation of the aqueous phase the chloromethyl
ether was washed repeatedly with water to remove remain-
ing aluminum salts. In the second stage of the process
(Scheme 2B), the crude chloromethyl ether was dissolved
1 For instance, it is possible that AlCl3 initially coordinates with the
oxygen of HFIP followed by electrophilic attack of the protonated
formaldehyde on the oxygen of HFIP.

C. Bieniarz et al. / Journal of Fluorine Chemistry 106 (2000) 99±102
101
minimization of evaporative losses throughout the process
will undoubtedly result in a considerably higher yield. We
believe that the new synthesis will be used in the future
manufacture of sevo¯urane.
3. Experimental
3.1. General methods
1H NMR spectra were recorded at 300 or 400 MHz on
Varian NMR spectrometers. 13C spectra were recorded at 75
or 100 MHz. Chemical shifts are reported in ppm down®eld
from tetramethylsilane (TMS, d 0.00). Qualitative and
semiquantitative analyses of reaction mixtures were per-
formed on an HP 6890 gas chromatograph equipped with an
HP 5973 mass selective detector. All reagents were pur-
chased from Aldrich and used without further puri®cation.
Scheme 3.
and 4:1 resulted in the predominant formation of the bis-
acetal. The success of the reaction is a result of the very
nonnucleophilic character of HFIP, which allows for chlor-
ide attack to effectively compete. This reaction is therefore
probably best suited for the chloromethylation of low pKa
alcohols, since our attempts at applying this method to
chloromethylate other alcohols without electron withdraw-
ing groups failed, giving instead only the bis-acetal of the
alcohol. In the chloromethylation of HFIP, the concentration
of the bis HFIP-acetal is the highest at the early stages of the
reaction and diminishes as the reaction proceeds. We sus-
pected that the bis-acetal was being cleaved by the AlCl3
once it had formed. Indeed, when a solution of pure bis-
HFIP-acetal was stirred at 238C with an equivalent of AlCl3,
equimolar amounts of HFIP and HFIP-chloromethyl ether
were produced. The second stage of the process, the con-
version of the HFIP chloromethyl ether to the corresponding
¯uoromethyl ether, represents an optimization of previously
documented methods for ¯uorination. In particular, we
found that polyethers are excellent solvents in the second
stage of the reaction. The use of poly(ethylene glycol) (PEG)
as the solvent allows for lower reaction temperature and
shorter reaction times and results in a remarkable increase in
yield. In addition, the use of PEG afforded the purest crude
product in comparison with other solvents. These effects
may be due to the reported complexation of potassium ion to
the PEGs and the concomitant increase in the nucleophilicity
of the ¯uoride ion [10]. While the use of KF as an inexpen-
sive reagent in chloride to ¯uoride exchange is well docu-
mented [7] there are very few literature examples of the use
of PEGs as solvents in these reactions [11±13]. Moreover,
the existing ¯uorination methods which utilize KF in the
absence of cation-complexing solvents invariably call for
much harsher reaction conditions [7].
3.2. Synthesis of 1,1,1,3,3,3-hexafluoro-2-(fluoromethoxy)-
propane
3.2.1. Stage 1: synthesis of 1,1,1,3,3,3,-hexafluoro-2-
(chloromethoxy)-propane
To a 100 ml ¯ask containing anhydrous aluminum
trichloride (18.56 g, 139.2 mmol) was added cold (08C)
HFIP (14.66 ml, 139.2 mmol) and the resultant slurry was
stirred for 10 min in an ice bath. To the reaction mixture was
added 1,3,5-trioxane (4.18 g, 46.60 mmol) in a single por-
tion and the ¯ask was capped with a rubber septum and
vented through a drying tube. The reaction was allowed to
warm to ambient temperature while stirring overnight. The
mixture was then cooled to 08C in an ice bath and an ef®cient
dry ice/acetone condenser was placed on the ¯ask. The
reaction was quenched with 50 ml of ice-cold 6N HCl,
added in small portions. Water was added, suf®cient to
dissolve any remaining aluminum salts, and the mixture
was partitioned. The bottom layer consisted of 27.0 g of
95% pure chloromethyl ether, for a yield of 90%. The
product, 1,1,1,3,3,3-hexa¯uoro-2-(chloromethoxy)-propane
may be further puri®ed by washing with 1N NaOH, drying
over MgSO4 and distilling at atmospheric pressure (bp
768C) to yield pure (99.9%) hexa¯uoroisopropylchloro-
methyl ether: 1H NMR (CDCl3, 400 MHz): d 5.55 (s,
2H), 4.54 (septet, 1H, JFCCH ˆ 5:7 Hz); 13C NMR (CDCl3,
100 MHz): d 121.3 (dq, JFC ˆ 283 Hz, JFCCC ˆ 3:0 Hz),
80.2 (s), 73.1 (septet, JFCC ˆ 33:4 Hz). In the one-vessel
procedure, the crude product of the ®rst reaction is simply
rinsed with water (2 Â 100 ml) to remove residual alumi-
num salts, and carefully decanted.
In summary, we have developed a new synthesis of
sevo¯urane, a very important inhalation anesthetic. The
new synthesis avoids the use of toxic and environmentally
hazardous reagents, and can be carried out in a single vessel
in two stages. Puri®cation via a simple distillation affords
99.95% pure sevo¯urane in 65±70% overall yield. The
3.2.2. Stage 2: fluorination of 1,1,1,3,3,3-hexafluoro-2-
(chloromethoxy)-propane
To
a
solution of 1,1,1,3,3,3-hexa¯uoro-2-(chloro-
methoxy)-propane (2.16 g, 10 mmol) in PEG 400 (10 ml)
was added spray-dried KF (2.32 g, 40 mmol) at room tem-

102
C. Bieniarz et al. / Journal of Fluorine Chemistry 106 (2000) 99±102
perature. The reaction mixture was heated at 958C, for 1 h,
then cooled to room temperature and diluted with water
(30 ml). Distillation directly from the reaction vessel
afforded 1.42 g (71%) of 1,1,1,3,3,3-hexa¯uoro-2-(¯uoro-
methoxy)-propane, bp 558C. The purity of this material was
determined at 99.95% by GC analysis). El-MS: m/z 199
References
[1] M. Scheller, J. Bufler, H. Sclmeck, E. Kochs, C. Franke,
Anesthesiology 86 (1997) 118±127.
[2] N.P. Franks, W.R. Lieb, Nature 367 (1994) 607±614.
[3] T. Kira, N. Harata, T. Sakata, N. Akaike, Neuroscience (Oxford) 85
(1998) 383±394.
‡
([M ±H], 5%, 181 (30%), 131 (100%), 79 (35%), 69 (50%),
[4] C.L. Coon, R.L. Simon, US Patent 4,250,334 (10 February 1981).
[5] C.L. Coon, R.L. Simon, US Patent 4,469,898 (4 September 1984).
[6] M.A. Tius, J.K. Kawakami, Tetrahedron 51 (1995) 3997.
[7] M. Hudlicky, A.E. Palath, Chemistry of Organic Fluorine Com-
pounds II ACS Monograph 187, Washington, DC, 1995.
[8] M. Fieser, L. Fieser, Reagents for Organic Synthesis, Vol. II, Wiley/
Interscience, New York, 1969, p. 22.
and 51 (27%). 1H NMR (CDCl3, 300 MHz): d 5.42 (d, 2H,
JHF ˆ 53:4 Hz, 4.42 (m, 1H, JFCCH ˆ 5:7 Hz). 13C NMR
(CDCl3, 75 MHz): d 121.1 (q, JFC ˆ 283 Hz), 103.0 (d,
JFC ˆ 226 Hz), 74.2 (septet, JFCC ˆ 33:8 Hz).
[9] S.H. Pines, R.F. Czaja, N.L. Abramson, J. Org. Chem. 40 (1975)
1920±1923.
Acknowledgements
[10] C.M. Starks, C.L. Liotta, M. Halpern, Phase Transfer Catalysis,
Chapman & Hall, New York, 1994, pp. 158±170.
We thank Professor David M. Lemal of Dartmouth Col-
lege for his helpful discussions. We also thank Dr. Curt
Kirkemo of Abbott Laboratories for his support in the course
of this work.
[11] D. Badone, G. Jommi, R. Pagliarin, P. Tavecchia, Synthesis 10 (1987)
920±921.
[12] T. Kitazume, N. Ishikawa, Chem. Lett. 3 (1978) 283±284.
[13] R. Neumann, S. Dermeik, Y. Sasson, Isr. J. Chem. 26 (1985) 239±242.