Pyrolysis reactions of 4-phenyl-2,3,5,6-tetrafluorophenyl prop-2-enyl ether and 4-trifluoromethyl-2,3,5,6-tetrafluorophenyl prop-2-enyl ether: Remarkable rearrangement reactions of intramolecular Diels-Alder products.: Mechanistic implications of a new thermal retro-Diels-Alder reaction to cyclohexa-2,4-dienylmethyl fluoroketenes and recyclisations

Article abstract of DOI:10.1016/S0022-1139(00)00405-X

The pyrolyses of 4-phenyl-2,3,5,6-tetrafluorophenyl prop-2-enyl ether (26) under flash vapour phase (FVP) conditions at 350°C and of 4-trifluoromethyl-2,3,5,6-tetrafluorophenyl prop-2-enyl ether (27) on heating in vacuo at 169°C give mixtures of products which include 3-phenyl-2,4,5,7-tetrafluorotricyclo[3.3.1.0^2,7]non-3-ene-6-one (28), and 3-trifluoromethyl-2,4,5,7-tetrafluorotricyclo[3.3.1.0 ^2,7]non-3-ene-6-one (31), respectively, the products of one of the two possible intramolecular Diels-Alder reactions of the Claisen rearrangement intermediates 2. FVP of 26 at 430°C and 27 at 450°C give the bicyclic compounds 7-phenyl-2,5β,6,7aβ-tetrafluoro-3aβ,4,5,7a-tetrahydroinden-1-one (34) and 7-trifluoromethyl-2,5β,6,7aβ-tetrafluoro-3aβ,4,5,7a- tetrahydroinden-1-one (37), respectively. It is proposed that these latter two compounds are formed via two possible intermediates produced by recyclisations of the tethered 2,4,5-trifluoro-3-(substituent)-2,4-cyclohexadienylmethyl fluoroketene (23), itself formed via new retro-cyclisation reactions of 28 and 31, respectively. Also formed from 27 via 31 is 1,2,4β,5,7-tetrafluoro-3-(trifluoromethyl)bicyclo[3.3.1]nona-2,6-dien-8-one (39).

Full text of DOI:10.1016/S0022-1139(00)00405-X

Journal of Fluorine Chemistry 108 (2001) 57±67
Pyrolysis reactions of 4-phenyl-2,3,5,6-tetra¯uorophenyl prop-2-enyl
ether and 4-tri¯uoromethyl-2,3,5,6-tetra¯uorophenyl prop-2-enyl
ether: remarkable rearrangement reactions of
intramolecular Diels±Alder products.
Mechanistic implications of a new thermal retro-Diels±Alder
reaction to cyclohexa-2,4-dienylmethyl ¯uoroketenes and
recyclisations by 4 ‡ p2s and/or 2 ‡ p2a routes
p s
p s
David M. Allen, Andrei S. Batsanov, Gerald M. Brooke*, Stephen J. Lockett
Chemistry Department, Science Laboratories, South Road, Durham DH1 2LE, UK
Received 5 October 2000; accepted 16 November 2000
Abstract
The pyrolyses of 4-phenyl-2,3,5,6-tetra¯uorophenyl prop-2-enyl ether (26) under ¯ash vapour phase (FVP) conditions at 3508C and of 4-
tri¯uoromethyl-2,3,5,6-tetra¯uorophenyl prop-2-enyl ether (27) on heating in vacuo at 1698C give mixtures of products which include 3-
phenyl-2,4,5,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-ene-6-one (28), and 3-tri¯uoromethyl-2,4,5,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-ene-6-
one (31), respectively, the products of one of the two possible intramolecular Diels±Alder reactions of the Claisen rearrangement
intermediates 2. FVP of 26 at 4308C and 27 at 4508C give the bicyclic compounds 7-phenyl-2,5b,6,7ab-tetra¯uoro-3ab,4,5,7a-
tetrahydroinden-1-one (34) and 7-tri¯uoromethyl-2,5b,6,7ab-tetra¯uoro-3ab,4,5,7a-tetrahydroinden-1-one (37), respectively. It is proposed
that these latter two compounds are formed via two possible intermediates produced by recyclisations of the tethered 2,4,5-tri¯uoro-3-
(substituent)-2,4-cyclohexadienylmethyl ¯uoroketene (23), itself formed via new retro-cyclisation reactions of 28 and 31, respectively. Also
formed from 27 via 31 is 1,2,4b,5,7-tetra¯uoro-3-(tri¯uoromethyl)bicyclo[3.3.1]nona-2,6-dien-8-one (39). # 2001 Elsevier Science B.V.
All rights reserved.
Keywords: Diels±Alder adducts; Retro-Diels±Alder products; p4s ‡ p2s and/or p2s ‡ p2a cyclisations; 2,4,5-tri¯uoro-3-(substituent)-2,4-cyclohexadienyl-
methyl ¯uoroketenes
1. Introduction
0.05 mmHg (Scheme 1) but later [2], it was found that
compound 4 was readily formed by heating 1 in re¯uxing
DMF with potassium ¯uoride as a base to effect the de¯uor-
ination of the intermediate 2. In the FVP reaction at 3658C,
the 2,5-cyclohexadienone (5) was isolated, the product of
Claisen/Cope rearrangements [3] (Scheme 2) while at 440±
4808C, two compounds were obtained from a complex
mixture: (i) the ¯uorovinyl ketone 8, formed by the decom-
position of one of the two possible intramolecular Diels±
Alder adducts 6 from 2, to 7 followed by loss of HF [4]; and
(ii) the bicyclic compound 11, the formation of which was
rationalised to take place via the intermediacy of the alter-
native intramolecular Diels±Alder adduct 9 (R=F) followed
by ring-opening to 10 (R=F) and hydrogen abstraction [1]
(Scheme 3). The Diels±Alder adduct 6 formed readily in
In an earlier paper [1], the preparation of the heterocyclic
compound 5,6,7,8-tetra¯uoro-2H-chromene (4) was
attempted from readily-available penta¯uorophenyl prop-
2-enyl ether (1) by a novel ¯ash vapour-phase (FVP) pyr-
olysis reaction through a silica tube packed with silica wool.
It was envisaged that the initial formation of the Claisen
rearrangement product 2 would be followed by the pyrolytic
elimination of HF and electrocyclisation of the resulting o-
quinomethane-type material 3. The reaction failed to yield
the desired product both at 3658C or 440±4808C/
* Corresponding author. Fax: ‡44-191-384-4737.
0022-1139/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 0 0 ) 0 0 4 0 5 - X

58
D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
Scheme 1.
vacuo at 137±1418C over 13 days [5], and was accompanied
by its skeletally-related isomer 13 containing a plane of
symmetry and isolated as its 1,1-diol hemihydrate (14) [6].
The presence of 13 was rationalised by invoking an overall
[1,3] sigmatropic shift occurring in a stepwise manner from
6 via the diradical 12 (R=F) (Scheme 2). Schmid described
the formation of the ®rst internal Diels±Alder adduct from a
Claisen rearrangement intermediate from 2,6-dimethylphe-
nyl prop-2-ynyl ether in which the CH=C bond of the 2-
allenyl group, ±CH=C=CH2, was the dienophile entity [7].
The intramolecular Diels±Alder adduct 6 was the ®rst
tricyclic compound to be prepared from a `simple' aryl
prop-2-enyl ether and arises as a consequence of the elec-
tron-poor nature of the tetra¯uorodiene system and the
electron-rich internal alkene. The ready formation of these
unique internal adducts in poly¯uoroaromatic systems is the
starting point for the novel chemistry reported in this paper.
Recently [8], the possibility of effecting the ®rst example
of a Cope rearrangement directly involving an aromatic ring
in the [3,3] sigmatropic rearrangement was investigated
using 15, the 4-penta¯uorophenyl derivative of 1. It seemed
reasonable to suppose that using poly¯uoroaromatic rather
than aromatic hydrocarbon prop-2-enyl ethers much higher
temperatures could be used with fewer side reactions since
the C±F bond is much less prone to heterolysis than the C±H
bond. No isomerisation of 15 to 16 or further to 17 was
detected under FVP conditions at 3658C (Scheme 2). Nickon
and Aaronoff [9] similarly had failed to obtain `out of ring'
rearrangements using prop-2-enyl ethers of several 2- and 4-
hydroxybiphenols: disproportionation reactions occurred to
the corresponding phenols, showing that the second phenyl
ring was acting as a blocking group and not as a prop-2-enyl
group acceptor. The main products isolated from the reac-
tion with 15 were the internal Diels±Alder adduct 18 and its
Scheme 2.

D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
59
Scheme 3.
rearranged isomer 19 which readily formed the 1,1-diol
monohydrate (20) (Scheme 2). When compound 15 was
subjected to more vigorous FVP conditions at 4208C, the
isomeric bicyclic material 21 expected on the basis of the
mechanism proposed earlier starting from 1 [1] was not
obtained: the product was 22 in which the ¯uorine and the
C6F5 groups on the carbon±carbon double bond in 21 were
interchanged (Scheme 3). Clearly, the straightforward reac-
tion via the internal Diels±Alder adduct 9 (R=C6F5) had not
taken place, which raised an intriguing mechanistic pro-
blem. A remarkable rearrangement had obviously occurred!
The formation of 22 was explained at that time [8], to have
arisen via compound 24, isomeric with and having the same
carbon skeleton as 9, followed by ring opening to 25 and
hydrogen abstraction as before (Scheme 4). The formation
of the unexpected intermediate tricyclic compound 24 was
Scheme 4.

60
D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
rationalised by invoking, apparently for the ®rst time
[10,11], a retro-Diels±Alder reaction of the internal
Diels±Alder adduct 18 to give the cyclohexa-2,4-dienyl-
methyl ¯uoroketene 23, followed by the alternative Diels±
Alder cyclisation as shown in Scheme 4. Intermolecular
p4s ‡ p2s reactions of ketenes to form six-membered car-
bocyclic rings [10±12] are uncommon, but one intramole-
cular process of this type has been recorded [13]. In the
present paper we report further investigations of these
unusual rearrangement reactions using the 4-phenyl- and
4-tri¯uoromethyl-substituted prop-2-enyl ethers (26 and
27), respectively, and propose another mechanism based
again on an intermediate cyclohexa-2,4-dienylmethyl ¯uor-
oketene (23). A further isomerisation has also been discov-
ered with 27.
Fig. 1. X-ray crystal structure of 28. Henceforth, double bonds are shown
in black, thermal ellipsoids at 50% probability level.
2. Results and discussion
The ethers 26 and 27 were prepared from 2,3,4,5,6-
penta¯uorobiphenyl [14] and octa¯uorotoluene, respec-
tively, using sodium prop-2-enoxide. Internal Diels±Alder
compounds having structures typi®ed by 18 were key inter-
mediates in the present work but the attempted preparation
of the adduct 28 by prolonged heating of the 4-phenyl ether
26 in vacuo at 1858C failed because of the involatility of the
material; however, reaction under FVP conditions at 3508C
and 0.01 mmHg, used successfully with the 4-C6F5± ether
15 [8] gave a very complex mixture of products which when
analysed by 19F NMR spectroscopy1 indicated the presence
of analogues of 18 and 19 from chemical shift comparisons.
Three components were separated by chromatography: (i)
the Diels±Alder adduct 28 (5% yield based on a 72%
conversion of 15) the structure of which was determined
by X-ray crystallography (Fig. 1); (ii) a derivative of the
skeletally related isomer 29 (5.5% prior to chromatography),
namely, the 1,1-diol 30 (0.25% prior to chromatography Ð
probably formed by the hydration of 29 by wet solvent
during the isolation procedure) formed by reaction of 29
with adventitious water on the stationary phase during
separation; (compound 30 was readily dehydrated to 29
by heating at 1588C and the structure of this ketone deter-
mined by X-ray crystallography Fig. 2); and (iii) 4-phenyl-
2,3,5,6-tetra¯uorophenol FP (3.5%) (formed by a dispro-
portionation reaction) identical with the material prepared
previously [15] (Scheme 2). When the 4-tri¯uoromethyl
ether (27) was heated in vacuo at 1698C for 168 h, a mixture
of the intramolecular Diels±Alder adduct 31 from 2 (R=CF3)
and the skeletally identical 1,3-rearranged isomer 32 in the
ratio 32:68, respectively, was obtained in 91% yield
(Scheme 2). The structures of both 31 and 32 were deter-
mined by X-ray crystallography (Figs. 3 and 4, respec-
tively). Compound 27 was also subjected to FVP at
3508C at 0.01 mmHg but it was a less ef®cient method
for producing 31: examination of the crude product by 19F
NMR spectroscopy showed that it contained not only the
adduct 31 (21% yield based on a 40% conversion of 27) and
the product of its isomerisation, 32 (21%) but also a new
material, the Cope rearrangement product 33 (6%) formed
from 2 (R=CF3). Compound 33, isolated from the complex
mixture of products by chromatography, was slightly con-
taminated with unreacted ether 27 (Scheme 2). Clearly, the
4-CF3 group, unlike the C6F5± and the C6H5± groups does
not inhibit the Cope rearrangement reaction following the
preliminary Claisen rearrangement.
The major interest in the present work was the potential
demonstration of further examples of the formation of
1 In spite of the complexity of the 19F NMR spectra of the crude reaction
products, because of the wide range of chemical shifts ( 50 to
220 ppm), once pure materials have been obtained by chromatography
it is a simple matter to identify and quantify (molar percentages) even very
small proportions of products (see Section 3).
Fig. 2. X-ray crystal structure of 29.

D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
61
Fig. 3. X-ray crystal structure of 31.
Fig. 5. X-ray crystal structure of 34.
rearrangement products analogous to the bicyclic compound
22 formed from 15, reactions which were carried out at very
high temperatures. Pyrolysis of the 4-phenyl ether 26 at
4308C and 0.01 mmHg as before and chromatography of the
crude product aimed speci®cally for this purpose did give 34
(16.5% yield) (Scheme 4), the structure being determined by
X-ray crystallography (Fig. 5). The 19F COSY experiment
on 34 showed the expected connectivities as were found with
22. The experiment was repeated under identical conditions
and an enriched sample of a liquid was isolated and identi-
®ed as the ¯uorovinyl ketone 36 (9% yield, 85% purity)
(Scheme 3) from the chemical shifts in the 19F NMR
spectrum, particularly the vinylic ¯uorine having the large
yield). When the 4-tri¯uoromethyl ether 27 was pyrolysed
at 4508C and 0.01 mmHg and the crude product chromato-
graphed as before, a fraction enriched in suspected 4-tri-
¯uoromethyl-2,3,5,6-tetra¯uorophenol FCF [16], the
3
product of an alkyl-ether cleavage reaction, was treated
with diazomethane and the volatile product identi®ed as
the corresponding methyl ether by comparison of its IR with
that of an authentic sample. In two other separate pyrolyses
of 27 under the same conditions, a liquid compound 37 was
obtained, analogous to 22 and 34 (Scheme 4). The structure
of 37 was determined by 19F NMR spectroscopy, in parti-
cular a 19F COSY experiment which established the same
connectivities as found in 22 and 34. The ¯uorovinyl ketone
38 was readily detected in crude reaction products from the
chemical shifts and coupling patterns of the two aromatic
and one vinylic ¯uorine in the 19F NMR spectrum but could
not be isolated (Scheme 3). The 19F NMR spectrum of the
crude reaction product showed that bicyclic compound 37
and the ¯uorovinyl ketone 38 were formed in 11 and 8.5%
yields, respectively, accompanied by 4-CF3-2,3,5,6-tetra-
1
JF-antiH 43 Hz, readily measured in the H NMR spectrum.
Structurally related ¯uorovinyl ketones 8 and 35 had been
obtained previously from 1 [4] and 15 [8]. Analysis of the
crude pyrolysis product from the reaction by 19F NMR
spectroscopy showed that in addition to 34 and 36, there
was present 4-phenyl-2,3,5,6-tetra¯uorophenol FP (18%
¯uorophenol FCF (18%), and even the Diels±Alder adduct
3
31 (0.4%) and its skeletally identical isomer 32 (0.3%), and
the Claisen/Cope rearrangement product 33 (0.2%)!
One further structurally novel product 39 (2.5%) was
isolated as a solid from the FVP of 4-tri¯uoromethyl ether
27 at 4508C. The 19F NMR spectrum showed a geminal CHF
group in the molecule but at a very different shift to that
found in 37; the structure was determined by X-ray crystal-
lography (Fig. 6). A mechanism which rationalises the
formation of 39 from 27 is shown in Scheme 5 and is
analogous to that shown in Scheme 4 for the conversion
of diradicals of type 25 into 22, 34 and 37, though from
models there is less ¯exibility in the conjugated diradical 12
(R=CF3).
The work described in this paper and the earlier one [8]
shows clearly from the isomerisation of the 4-substituted-
2,3,5,6-tetra¯uorophenyl prop-2-enyl ethers 15, 26 and 27 to
Fig. 4. X-ray crystal structure of 32.

62
D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
functionality compared with the C=O group in the more
rigid 2,4-dienone functionality in the Claisen rearrangement
intermediates 2. However, we now realise that there is an
alternative mechanism to this p4s ‡ p2s route from 23 to the
conjugated diradical 25 (Scheme 4) which must be consid-
ered. This is shown in Scheme 6 and again involves the prior
formation of 23 (NB the enantiomeric structures of 22, 25,
34 and 37 are formed in this way from 40).
Concerted reactions of ketenes and alkenes to form
cyclobutanones via a 2s ‡ p2a route are well documented
p
[17]. For these antarafacial processes to take place the
LUMO of the C=C bond in the ketene has to be orientated
above and at right-angles to the HOMO of the alkenic C=C
bond for the most effective overlap of the orbitals and simple
molecular models show that this geometry with the ¯uor-
oketene 23 is realised en route to the formation of 40, a
highly strained molecule, so that its formation at 4508C
cannot be excluded. Homolytic ®ssion of the C±C bond
indicated in 40 in Scheme 6 provides the second possible
route to the diradicals 25, the precursors to the bicyclic
compounds 22, 34 and 37.
Diels±Alder reactions are, arguably, counted among the
most important C±C bond-forming reactions in organic
chemistry. The electronic properties of ¯uorine in the
substrates 1, 15, 26 and 27 have enabled unique intramo-
lecular Diels±Alder compounds 6, 18, 28 and 31 to be
Fig. 6. X-ray crystal structure of 39.
the 7-substituted-bicyclic compounds 22, 34 and 37, respec-
tively, that the process is a general one. In the earlier paper
[8] we proposed the mechanism shown in Scheme 4 for the
formation of 22 in which an intermediate 24 (R=C6F5) is
formed from the tethered ketene 23. (R=C6F5). It was noted
that it is not obvious why the circuitous route proposed to
give 24 is energetically more favourable than a direct route
to 9 having the same carbon skeleton. A closer examination
of simple molecular models shows that there is more ¯ex-
ibility in structures of the type 23, the precursor to 24,
because of the CH2 group adjacent to the 2,4-diene
Scheme 5.
Scheme 6.

D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
63
prepared and their reactions at high temperatures to be
studied. The formation of the bicyclic compounds 22, 34
and 37 demanded an interpretation which has been met by
the ®rst ever proposal of a thermal (as opposed to a photo-
chemical [10,11]) retro-Diels±Alder reaction to ketene die-
neophiles, in this case to tethered ¯uoroketenes of general
structure 23; thereafter, two possible routes to the rearranged
products have been described. It hardly needs to be stated
that the mechanisms proposed are still only hypotheses in
this area, the chemistry of which could never have been
predicted. The challenge remains to isolate the intermediates
23 and 24/or 40 to de®ne more precisely the mechanism of
the reaction; work is proceeding along these lines.
was heated to 3508C with the boat initially at one end of the
closed tube at room temperature, the other end being con-
nected to a high vacuum system via a trap cooled in liquid
air. When the pressure had reached 0.01 mmHg, the pyr-
olysis tube was moved to allow the boat and contents to enter
the hot zone of the heating oven and a hot air gun was used to
vaporise any condensate beyond the exit of the oven into the
trap. Each experiment took approximately 20 min and the
pyrolysate was washed from the trap with dichloromethane,
the solvent distilled and a 19F NMR was run of a sample in
CDCl3 to allow, ultimately, an analysis of the complex
composition (see later). The reaction products were com-
bined (1.306 and 1.752 g, respectively) and the mixture
chromatographed on silica using light petroleum (bp 40±
608C)/diethyl ether (50:50% v/v) as eluant to give a series of
complex fractions (examined by 19F NMR) followed by one
which was identi®ed as 4-phenyl-2,3,5,6-tetra¯uorophenol
(0.068 g). The solvent was changed to neat diethyl ether and
the column washed to remove all the remaining material, a
solid (ca. 0.6 g), the TLC of which showed two major spots.
This solid was chromatographed on silica using diethyl
ether/light petroleum (bp 40±608C) (50:50% v/v) as eluant
to give (i) a fraction (0.236 g) which was subjected to
sublimation at 958C/0.001 mmHg and the yellow sublimate
(0.205 g) washed with cold ether to remove the yellow
colouration and ®nally the remaining solid recrystallised
from toluene/light petroleum (bp 80±1008C) to yield
3-phenyl-2,4,5,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-ene-6-
one (28), mp 114±1158C (found: C, 63.77; H, 3.54.
C15H10F4O requires C, 63.83; H, 3.57%); dF(CDCl3)
131.9 (m, 4-F), 176.1 (m, 7-F), 193.4 (m, 5-F),
197.6 (d, 2-F); dH(CDCl3) 2.10 (d, 1-H), 2.27 (m, 1-H),
2.41 (m, 1-H), 3.00 (m, 1-H), 3.25 (m, 1-H), 7.39 (C6H5);
nmax 1770 (C=O), 1680 [C(C6H5)=CF] cm 1; and (ii) the
following fraction (0.348 g) contained a 50:50 mixture of 30
and 34 (see later) which was separated by selective crystal-
lisation from chloroform to give 3-phenyl-1,2,4,7-tetra¯uor-
otricyclo[3.3.1.02,7]non-3-ene-8,8 diol (30), mp 141.5±
142.58C (found: C, 59.83; H, 4.06. C15H12F4O2 requires
C, 60.00 H, 4.03%); dF(CDCl3) 111.8 (m, 4-F), 195.2 (d,
1-F, 7-F), 201.7 (s, 2-F); dH(CDCl3) 1.76 (m, 2-H), 2.68
(m, 2-H), 2.83 (m, 1-H), 2.90 (s, 1-H), 3.91 (d, 5-H), 7.36 (m,
C6H5); nmax 3343 (shoulder at 3553) (gem-diol), 1701
(CF=CC6H5) cm 1. Heating 30 at 1588C for 10 min fol-
lowed by sublimation in vacuo at ca. 808C/0.01 mmHg gave
3-phenyl-1,2,4,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-ene-8-
one (29), mp 141±1428C (from light petroleum (bp 80±
1008C)) (found: C, 63.56; H, 3.49. C15H10F4O requires C,
63.83; H, 3.57%); dF(CDCl3) 111.5 (M, 4-F), 187.4 (d,
1-F, 7-F), 200.9 (m, 2-F); dH(CDCl3) 2.15 (tm, 2-H), 2.62
3. Experimental
NMR spectra were recorded on the following instruments
at the frequencies listed: Varian Mercury 200 (1H,
199.991 MHz; 19F, 188.179 MHz) and Varian Unity (1H,
299.908 MHz). Chemical shifts are reported using the high-
frequency positive convention from TMS and CFCl3, hence
19F resonance values are negative; J values are in Hz); 19F
COSY experiments were carried out on compounds 34, 37
and 39 to establish connectivities. Elemental analyses were
performed on an Exeter Analytical Inc., CE440 elemental
analyser.
3.1. 4-Phenyl-2,3,5,6-tetrafluorophenol FP
Compound FP was prepared in 43% yield from 2,3,4,5,6-
penta¯uorobiphenyl by the method in the literature [6]. dF
(CDCl3) 145.7 (m, 20-F, 60-F), 164.2 (m, 3-F, 5-F).
3.2. Preparation of 4-phenyl-2,3,5,6-tetrafluorophenyl
prop-2-enyl ether (26)
1,2,3,4,5-Penta¯uorobiphenyl [14] (46.4 g), sodium prop-
2-enoxide (16.7 g) and prop-2-enyl alcohol (125 ml) were
heated together under re¯ux for 18 h, to give 4-phenyl-
2,3,5,6-tetra¯uorophenyl prop-2-enyl ether (26) (47.9 g,
89%), mp 56±588C (from methanol) (found: C, 63.76; H,
3.54. C15H10F4O requires C, 63.83; H, 3.57%); dF(CDCl3)
145.2 (m, 3-F, 5-F), 156.6 (m, 2-F, 6-F). dH(CDCl3) 4.77
(d, CH2), 5.31, 5.37, 5.39 and 5.48 (four resonances of two
doublets, CH2=CH), 5.96±6.20 (m, CH=CH2), 7.46 (m,
C6H5).
3.3. Pyrolysis of 4-phenyl-2,3,5,6-tetrafluorophenyl
prop-2-enyl ether (26)
(dm, 2-H), 3.04 (d, 5-H), 7.40 (m, C6H5); nmax 1824 (C=O),
1
3.3.1. At 3508C
1687 (CF=CC6H5) cm
.
In two separate experiments the ether 26 (2.465 and
2.939 g) was distilled in vacuo from a silica boat through
a silica tube (510 mm  20 mm) packed in the middle
170 mm with silica wool. The packing only in the tube
3.3.2. At 4308C
The ether 26 (3.004 g) was pyrolysed at 4308C as
described in (a) and again the 19F NMR was run in CDCl3

64
D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
to allow, ultimately, an analysis of the complex composition
(see later). The crude reaction product (2.001 g) was chro-
matographed on silica using light petroleum (bp 40±608C)/
diethyl ether (50:50% v/v) as eluant to give a fraction
(0.601 g) which was crystallised ®rst at 188C from light
petroleum (bp 40±608C)/diethyl ether (20:80% v/v) then the
solid (0.284 g) from superheated diethyl ether in a closed
tube to give 7-phenyl-2,5b,6,7ab-tetra¯uoro-3ab,4,5,7a-tet-
rahydroinden-1-one (34), mp 100±1028C (found: C, 63.61;
H, 3.55. C15H10F4O requires C, 63.83; H, 3.57%);
dF(CDCl3) 108.8 (d, 6-F), 134.7 (d, 2-F), 151.8 (m,
7ab-F), 192.2 (dm, 5b-F); dH(CDCl3) 2.29 (m, one 4-H),
2.54 (m, one 4-H), 3.47 (dm, 4a-H), 5.11 (dm, 5a-H; J5aH±
5bF 50.0), 6.85 (m, 3-H), 7.30 and 7.42 (both m, C6H5); nmax
3089 (=C±H), 1747 (C=O), 1692 (CF=CC6H5) and
47.52:4.53:6.58:6.91:0.33:4.29:2.61:27.20, respectively. The
molecular weights of all but the unknown materials U are
known. As an approximation, the molecular weight of U is
taken to be that of the starting material 26 in this experiment
and the one below, to enable the masses of the individual
components in the crude product to be determined. (While it
is of interest to have an estimate of the yields of the several
materials identi®ed in the pyrolysis reactions, they are not of
critical importance in this work). In this FVP reaction,
compound 26 (2:465 ‡ 2:939 ˆ 5:404 g) gave the crude
product (1:306 ‡ 1:752 ˆ 3:058 g). Calculations from the
NMR data gives 1.465 g of unreacted 26 in the crude product
so that only 5:404 1:465 ˆ 3:939 g of 26 was converted to
products, i.e. 393.9/5:404 ˆ 72% (the ``conversion''). The
yields of the compounds identi®ed in the crude product
based on the amount of 26 that was consumed were as
follows: FP (3.5%), 28 (5%), 29 (5.5%), 30 (0.25%), 34
(3.5%), and 36 (2%).
1
1647 cm (CF=CH).
In a similar experiment with 26 at 4308C and chromato-
graphy on silica using light petroleum (bp 40±608C)/diethyl
ether (95:5% v/v), a fraction was obtained containing
mainly 1-¯uorovinyl-3-phenyl-2,4-di¯uorophenyl ketone
(36) and 4-phenyl-2,3,5,6-tetra¯uorophenol. The phenol
was removed from the mixture by extraction with sodium
hydroxide solution and the material remaining was rechro-
matographed on silica using light petroleum (bp 40±608C)/
diethyl ether (75:25% v/v). The main component was ``sub-
limed'' into a cup at the end of a water-cooled ®nger at 808C/
0.01 mmHg and the liquid sublimate was chromatographed
again on silica using light petroleum (bp 40±608C)/diethyl
ether (90:10% v/v) and the further enriched product ``resu-
blimed'' as before to give 36, a liquid, ca. 85% pure by 19F
3.4.2. At 4308C
There was no recovered starting material 26 in the crude
product so the conversion was 100%. The molar proportions
of all the identi®ed components and collectively unidenti-
®ed materials U in the crude product, i.e. FP, 34, 36 and U
were in the ratio 26.37:23.89:11.67:38.07, respectively.
The yields were as follows: FP (18.5%), 34 (16.5%), 36
(9%).
3.5. Preparation of 4-trifluoromethyl-2,3,5,6-
tetrafluorophenyl prop-2-enyl ether (27)
‡
NMR (found: M , 262. C15H9F3O requires M, 262);
dF(CDCl3) 106.4 (m, unassigned F), 112.2 (m, unas-
Octa¯uorotoluene (42.61 g) and sodium prop-2-enoxide
(14.24 g) in prop-2-enyl alcohol (170 ml) were heated
together under re¯ux for 42 m, cooled to room temperature,
diluted with water and the mixture extracted with ether. The
ether extracts were washed six times with water (200 ml) to
remove the alcohol, dried (MgSO4), the solvent evaporated
in vacuo at room temperature and the residue distilled to
give 4-tri¯uoromethyl-2,3,5,6-tetra¯uorophenyl prop-2-
enyl ether (27) (44.52 g, 90%), bp 68.58C at 7.5 mmHg
(found: C, 43.78; H, 1.79. C10H5F7O requires C, 43.81; H,
1.84%); dF(CDCl3) 56.3 (t, CF3), 142.4 (m, 2-F, 6-F),
155.6 (m, 3-F, 5-F); dH(CDCl3) 4.85 (dm, CH2), 5.4 (dm,
1-H of CH2=CH), 5.38 (m, 1-H of CH2=CH), 5.9±6.16 (m, ±
CH=CH2).
signed F), 115.0 (m, Fa, JF
Ha overlapping with Hb, JF
43); dH(CDCl3) 5.49 (dd,
Hb
a
13.8, JH
3.6 in CFa=-
Ha
Hb
a
a
CHaHb where Fa and Ha are Z), 5.59 (dd, Hb overlapping
with Ha, JF 43); 7.10 (td) and 7.57 (q) (unassigned 5-H
Hb
a
1
and 6-H) overlapping with 7.47 (m, C6H5); nmax 1687 cm
(C=O).
3.4. Yields of products from the pyrolysis reactions of 26
3.4.1. At 3508C
In the 19F NMR spectrum of the crude product, the sum of
all the integrations for every ¯uorine resonance (Itotal) and
the integration equivalent to one ¯uorine of a unique absorp-
tion in every identi®able component (Icomp x) in the mixture
were measured. The assumption is now made that every
compound in the complex product retains four ¯uorines, the
only known exception being the ¯uorovinyl derivative 36
which has three ¯uorines, the unique single ¯uorine reso-
nance being Icomp 36, say. Consequently, (Itotal ‡ Icomp 36)
divided by 4 provides a correct measure of all the molecular
species present in the mixture. On this basis, the molar
proportions of all the identi®ed components and the collec-
tively unidenti®ed materials U in the crude product, i.e.
26, FP, 28, 29, 30, 34, 36 and U were in the ratio
3.6. Preparation of 4-trifluoromethyl-2,3,5,6-
tetrafluorophenyl methyl ether
Octa¯uorotoluene was reacted with sodium methoxide in
methanol as above to give 4-tri¯uoromethyl-2,3,5,6-tetra-
¯uorophenyl methyl ether (68%), bp 588C at 13 mmHg
(found: C, 38.52; H, 1.10. C8H3F7O requires C, 38.73; H,
1.22%); dF(CDCl3) 56.2 (t, CF3), 142.4 (m, 2-F, 6-F),
157.2 (m, 3-F, 5-F); dH(CDCl3) 4.19 (t, CH3).

D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
65
3.7. 4-Trifluoromethyl-2,3,5,6-tetrafluorophenol FCF
3.9.2. At 4508C
3
In two separate experiments the ether 27 (2.558 and
2.607 g) was pyrolysed at 4508C as in (a) and the combined
crude products (4.459 g) chromatographed on silica using
light petroleum (bp 40±608C)/diethyl ether (50:50% v/v) to
give an enriched fraction (0.36 g), recrystallisation of which
from chloroform gave 1,2,4b,5,7-tetra¯uoro-3-(tri¯uoro-
methyl)bicyclo[3.3.1]nona-2,6-dien-8-one (39), mp 118±
1208C (found: C, 43.68; H, 1.74. C10H5F7O requires C,
43.81; H, 1.84%); dF(CDCl3) 61.8 (d, 3-CF3), 104.2 (m,
2-F), 125.5 (m, 7-F), 167.6 (dd, 4b-F, J4aH±4bF 48),
184.8 (bs, 1-F); dH(CDCl3) 2.64 (m, 1-H), 2.95 (m, 1-
H), 3.36 (m, 1-H), 5.30 (dd, 4a-H, J4aH±4bF 48), 6.57 (dd, 6-
This compound was prepared by the method described
previously [16]; dF(CDCl3) 55.9 (t, CF3), 142.3 (m, 2-F,
6-F), 162.0 (m, 3-F, 5-F).
3.8. Static thermolysis of 4-trifluoromethyl-2,3,5,6-
tetrafluorophenyl prop-2-enyl ether (27)
The ether 27 (4.948 g) was sealed in a 10 l round-
bottomed ¯ask in vacuo and heated at 1698C for 168 h.
The products were condensed into a side arm cooled in
liquid air and washed from the opened vessel with ether and
dichloromethane to give the crude product (4.535 g) the 19F
NMR of which showed the presence of essentially two
products in the ratio 2.13:1 Recrystallisation from chloro-
form effected the separation of the major component 3-
tri¯uoromethyl-1,2,4,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-
ene-8-one (32), mp 145±1478 (sealed tube) (found: C, 43.66;
H, 1.71. C10H5F7O requires C, 43.81; H, 1.84%); dF(CDCl3)
58.7 (dd, 4-CF3), 94.1 (m, C=CF), 187.0 (dm, 1-F, 7-
F), 208.1 (m, 2-F); dH(CDCl3) 2.10 (m, 2-H), 2.62 (dm, 2-
H), 3.02 (dm, 5-H); nmax 1844 (C=O), 1703 (CF=CCF3)
cm 1. The ®rst mother liquors from the recrystallisation of
32 were evaporated to dryness, recrystallised ®ve times from
toluene/light petroleum (bp 60±808C), seven times from
toluene at 188C and three more times from toluene at 5,
0 and 58C, respectively, to give the minor component 3-
tri¯uoromethyl-2,4,5,7-tetra¯uorotricyclo[3.3.1.02,7]non-3-
ene-6-one (31), mp 86.5±87.08C (found: C, 43.67; H, 1.80.
C10H5F7O requires C, 43.81; H, 1.84%); dF(CDCl3) 59.3
(dd 4-CF3), 112.5 (qm, C=CF), 175.9 (m, 7-F), 193.8
(m, 5-F), 204.5 (m, 2-F); dH(CDCl3) 2.13 (d, 1-H), 2.27
H); nmax 1725 (C=O), 1701 [CF=C(CF3)] and 1653
1
(CF=CH) cm
.
In two further separate experiments the ether 27 (2.416
and 2.752 g) was pyrolysed at 4508C and the combined
crude products (5.057 g) separated as in (d) to give an
enriched liquid fraction (0.167 g) which was `sublimed'
at 858C/0.02 mmHg and the liquid `sublimate' resublimed
at 608C/0.02 mmHg to give 7-tri¯uoromethyl-2,5b,6,7ab-
tetra¯uoro-3ab,4,5,7a-tetrahydroinden-1-one (37), a liquid
(found: C, 43.71; H, 1.73. C10H5F7O requires C, 43.81; H,
1.84%); dF(CDCl3) 57.3 (tm, 7-CF3), 93.5 (m, 6-F),
133.6 (d, 2-F), 153.3 (m, 7ab-F), 193.8 (dm, 5b-F);
dH(CDCl3) 2.14 (m, one 4-H), 2.53 (m, one 4-H), 3.45 (dm,
4a-H), 5.06 (dm, 5a-H; J5aH±5bF 48.2 Hz), 6.87 (m, 3-H);
nmax 1757 (C=O), 1690 (CF=C(CF3)) and 1655 (CF=CH)
1
cm
.
In another two further separate pyrolyses of the ether 27 at
4508C (5.165 g), the crude products (4.459 g) were sepa-
rated as in (d) to give an enriched fraction of suspected 4-
tri¯uoromethyl-2,3,5,6-tetra¯uorophenol (0.8 g) which was
treated with excess diazomethane in ether, the solvent
evaporated on a rotatory evaporator and the residue sub-
jected to high vacuum (0.05 mmHg); the material which
condensed in a trap cooled in liquid air was identi®ed as 4-
tri¯uoromethyl-2,3,5,6-tetra¯uorophenyl methyl ether by
comparison of its IR spectrum with that of an authentic
sample.
(m, 1-H), 2.39 (m, 1-H), 3.04 (m, 1-H), 3.26 (m, 1-H); nmax
1
1777 (C=O), 1705 [C(CF3)=CF] cm
.
3.9. Pyrolysis of 4-trifluoromethyl-2,3,5,6-tetrafluoro-
phenyl prop-2-enyl ether (27)
An examination of the 19F NMR spectra of some of the
fractions obtained by chromatography of the pyrolysis
reactions of 27 at 4508C showed the presence of 1-¯uor-
ovinyl 3-phenyl-2,4-di¯uorophenyl ketone (38), identi®ed
by the presence of three characteristic ¯uorine resonances
with chemical shifts close to those in related compounds 8,
35 and 36: dF(CDCl3) 56.9 (3-CF3), 103.2 and 109.4
3.9.1. At 3508C
The ether 27 (2.510 g) was pyrolysed at 3508C as
in (a), the crude product (2.440 g) isolated as before
and the 19F NMR run in CDCl3 to allow, ultimately, an
analysis of the complex composition (see later). The mixture
was subjected to chromatography on silica using light
petroleum (bp 40±608C)/diethyl ether (95:5% v/v) to give
an enriched fraction (0.053 g), a liquid, containing ca. 3.5%
of 27 and 96.5% of 2,3,5,6-tetra¯uoro-4-(prop-2-enyl)-
(both m, unassigned 2-F, 4-F), 115.5 (m, Fa, JF
CFa=CHaHb where Fa and Hb are E.
43) in
Hb
a
4-(tri¯uoromethyl)-cyclohexa-2,5-dienone (33) (found:
‡
3.10. Yields of products from the pyrolysis reactions of 27
M
(t, 4-CF3), 127.7 (m, two unassigned F), 150.5 (dm,
274. C10H5F7O requires M, 274): dF(CDCl3) 69.6
3.10.1. At 3508C
In the 19F NMR spectrum of the crude product, the sum of
all the integrations for every CF3 group resonance between
55 and 70.6 ppm (Itotal) and the integration equivalent to
two unassigned F); dH(CDCl3) 3.09 (d, CH2), 5.28±5.67
(overlapping m, CH2=CH); nmax 1709 (C=O), 1687
1
(CF=CF) cm
.

66
D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
Table 1
Crystal data (T ˆ 120 K)
Compounda
28
29
31
32
34
39
Crystal system
Space group
Monoclinic
P21 (# 4)
6.217(1)
6.926(1)
14.225(2)
93.02(1)
611.7(2)
2
Monoclinic
P21/n (# 14)
12.006(2)
6.856(1)
14.851(1)
94.05(1)
1219.4(3)
4
Monoclinic
P21/n (# 14)
7.714(1)
12.587(1)
10.799(1)
109.26(1)
989.9(2)
4
Ortho rhombic
Pnma (# 62)
11.010(1)
8.007(1)
10.756(1)
90
Ortho rhombic
Pbca (# 61)
11.488(3)
12.761(3)
16.747(4)
90
Monoclinic
P21 (# 4)
7.229(1)
9.270(1)
7.692(1)
112.21(1)
477.2(1)
2
Ê
a (A)
Ê
b (A)
Ê
c (A)
b (8)
U (A )
Ê 3
948.2(2)
4
2455(1)
8
Z
Reflections total
Refletcions unique
Reflections, F2 > 2s(F2)
2ymax (8)
3235
1476
14094
3210
8883
2269
6983
1339
20939
3269
5982
2510
1325
2848
2006
1175
2924
2363
55
58
55
58
58
58
Rint
0.048
0.040
0.038
0.037
0.033
0.035
0.038
0.031
0.027
0.042
0.030
0.031
R[F2 > 2s(F2)]
wR(F2), all data
CCDC dep. no.
0.100
0.099
0.089
0.089
0.118
0.076
150540
150541
150542
150543
150544
150545
a Formulae: C15H10F4O, M ˆ 282.23 (28, 3-phenyl-2,4,5,7-tetrafluorotricyclo[3.3.1.02,7]non-3-ene-6-one; 29, 3-phenyl-1,2,4,7-tetrafluorotricy-
clo[3.3.1.02,7]non-3-ene-8-one; and 34, 7-phenyl-2,5b,6,7ab-tetrafluoro-3ab,4,5,7a-tetrahydroinden-1-one). C10H5F7O, M ˆ 274.14 (31, 3-trifluoromethyl-
2,4,5,7-tetrafluorotricyclo[3.3.1.02,7]non-3-ene-6-one; 32, 3-trifluoromethyl-1,2,4,7-tetrafluorotricyclo[3.3.1.02,7]non-3-ene-8-one; 39, 1,2,4b,5,7-tetrafluoro-
3-(trifluoromethyl)bicyclo[3.3.1]nona-2,6-dien-8-one).
one ¯uorine of a unique absorption in every identi®able
component (Icomp x) in the mixture were measured (there is
no problem here about how many ¯uorine atoms are present
in each species since every one will contain a CF3 group).
The value of Itotal divided by 3 provides the correct measure
of all the molecular species present in the mixture. Using
exactly the same method of calculation as was used with the
work using compound 26, the conversion of compound 27 to
products was 40% and the yields of products were as
follows: FCF3 (4%), 31 (21%), 32 (21%), 33 (6%), and
37 (1%).
were re®ned in anisotropic, H atoms in isotropic approx-
imation (in 28, all H atoms were treated as `riding'). Crystal
data and experimental details are summarised in Table 1,
full structural information has been deposited with the
Cambridge Crystallographic Data Centre as supplementary
publication CCDC-150540 to 150545.
Acknowledgements
We thank Dr. A. Kenwright and Ian McKeag for 19F
COSY experiments and Mrs. J. Dostal for elemental
analyses.
3.10.2. At 4508C
The conversion of compound 27 to products was 95% and
the yields of products were as follows: FCF (18%), 31
(0.4%), 32 (0.3%), 33 (0.2%), 37 (11%), 38 (8.5%) and 39
(2.5%).
3
References
[1] G.M. Brooke, J. Chem. Soc., Perkin Trans. I (1974) 233.
[2] G.M. Brooke, J. Fluorine Chem. 22 (1983) 283.
[3] G.M. Brooke, Tetrahedron Lett. (1971) 2377.
[4] G.M. Brooke, D.H. Hall, J. Chem. Soc., Perkin Trans. I (1976)
1463.
3.11. X-ray crystallography
All X-ray diffraction experiments were carried out at
T ˆ 120 K on a SMART 3-circle diffractometer with a
1 K CCD area detector, using graphite-monochromated
[5] G.M. Brooke, D.H. Hall, J. Fluorine Chem. 20 (1982) 163.
[6] G.M. Brooke, D.H. Hall, H.M.M. Shearer, J. Chem. Soc., Perkin
Trans. I (1978) 780.
Ê
Mo Ka radiation (l ˆ 0:71073 A) and a Cryostream
[7] J. Zsindely, H. Schmid, Helv. Chim. Acta 16 (1968) 1510.
[8] A.S. Batsanov, G.M. Brooke, D. Holling, A.M. Kenwright, J. Chem.
Soc., Perkin Trans. 1 (2000) 1731.
(Oxford Cryosystems) open-¯ow N2 gas cryostat. A com-
bination of four or ®ve sets of o scans; each set at different j
and/or 2y angles, nominally covered a hemisphere of the
reciprocal space (28, 32, 34), full sphere (29) or ca. 70%
thereof (31, 39). Re¯ection intensities were integrated using
SAINT program [18]. The structures were solved by direct
methods and re®ned by full-matrix least squares against F2
of all data, using SHELXTL software [19]. Non-H atoms
[9] A. Nickon, B.R. Aaronoff, J. Org. Chem. 29 (1964) 3041.
[10] D.J. Pollart, H.W. Moore, J. Org. Chem. 54 (1989) 5444.
[11] T.T. Tidwell, Ketenes, Wiley, New York, 1995.
[12] J.A. Hyatt, P.W. Raynolds, Org. React. 45 (1994) 159.
[13] T. Miyashi, H. Kawamoto, T. Nakajo, T. Mukai, Tetrahedron Lett.
(1979) 155.
[14] M.T. Chaudhry, R. Stephens, J. Chem. Soc. (1963) 4281.

D.M. Allen et al. / Journal of Fluorine Chemistry 108 (2001) 57±67
67
[15] J.M. Birchall, R.N. Haszeldine, H. Woodfine, J. Chem. Soc., Perkin
Trans. I (1973) 1121.
[18] SMART & SAINT, Area detector control & integration software, Ver.
6.01, Bruker Analytical X-ray Systems, Madison, WI, USA, 1999.
[19] SHELXTL, An integrated system for solving, refining and displaying
crystal structures from diffraction data, Ver. 5.10, Bruker Analytical
X-ray Systems, Madison, WI, USA, 1997.
[16] V.C.R. McLoughlin, J. Thrower, Chem. Ind. (1964) 1577.
[17] J. March, Advanced Organic Chemistry, 4th Edition, Wiley, New
York, 1992, pp. 857±859.