The behaviour of chlorofluoroethanes on β-aluminium(III) fluoride: A [ 36Cl ] radiotracer study

Article abstract of DOI:10.1016/S0022-1139(01)00527-9

A [³⁶Cl] radiotracer study of the behaviour of 1,1,2-trichlorotrifluoroethane on β-aluminium(III) fluoride at elevated temperature indicates that the isomerisation of CCl₂FCClF₂ to CCl₃CF₃ occurs by an intramolecular process. Isomerisation is followed by dismutation of CCl₃CF₃ to give CCl₂FCF₃ and CCl₃CClF₂. In neither reaction, surface Al-Cl groups are formed. The compound CCl₃CClF₂ undergoes further reaction, readily, apparently also via dismutation processes.

Full text of DOI:10.1016/S0022-1139(01)00527-9

Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
The behaviour of chloro¯uoroethanes on b-aluminiumꢀIII)
36
¯
uoride: a [ Cl] radiotracer study
a
Hamid Bozorgzadeh , Erhard Kemnitz , Mahmood Nickkho-Amiry ,
a
b
c
b
Toma zÏ Skapin , Graeme D. Tate , John M. Win®eld
b,*
a
Institute of Inorganic Chemistry, Humboldt University, Hessische Strasse 1-2, D-10115 Berlin, Germany
b
Department of Chemistry, University of Glasgow, Glasgow G12 8QQ, UK
c
`
`Jo zÏ ef Stefan'' Institute, Jamova 39, SI-1000 Ljubljana, Slovenia
Received 31 May 2001; accepted 31 August 2001
Abstract
3
6
A [ Cl] radiotracer study of the behaviour of 1,1,2-trichlorotri¯uoroethane on b-aluminiumꢀIII) ¯uoride at elevated temperature
indicates that the isomerisation of CCl FCClF to CCl CF occurs by an intramolecular process. Isomerisation is followed by dismutation
of CCl CF to give CCl FCF and CCl CClF . In neither reaction, surface Al±Cl groups are formed. The compound CCl CClF undergoes
2
2
3
3
3
3
2
3
3
2
3
2
further reaction, readily, apparently also via dismutation processes. # 2001 Elsevier Science B.V. All rights reserved.
Keywords: Chloro¯uoroethane; Chlorine-36; Dismutation
1
. Introduction
pre-¯uorinated with dichlorodi¯uoromethane, and very
short catalyst contact times have con®rmed that CCl FCF
2
3
The conversion of unwanted stocks of 1,1,2-trichlorotri-
uorethane, which prior to the Montreal Protocol was used
is formed via the intermediate ꢀCCl CF ) and imply that the
3 3
¯
conversion of CCl CF to CCl FCF proceeds by a dismuta-
3 3 2 3
widely as a degreasing solvent, to the CFC-alternative
refrigerant, 1,1,1,2-tetra¯uoroethane, HFC-134a, could be
a more attractive option for its disposal than by combustion.
We have been investigating a possible route for the conver-
sion of CCl FCClF to CH FCF that does not require the
tion reaction rather than by a simple ¯uorination [2].
Additional evidence for these ®ndings is now presented
3
6
from the results of a [ Cl] radiotracer study in which events
that occur at the surface of b-AlF during its exposure to
3
CCl FCClF or other members of the chloro¯uoroethane
2
2
2
2
3
2
use of anhydrous hydrogen ¯uoride ꢀHF) and have demon-
strated [1,2] that on a laboratory scale, this can be achieved
by an heterogeneously catalysed process. Formally, the
process can be described in three steps ꢀEqs. ꢀ1)±ꢀ3)).
series, C Cl Fn, were examined using a Geiger±M uÈ ller
2
6Àn
direct monitoring method. This technique was developed in
1
4
Glasgow for [ C] adsorption measurements relevant to
hydrocarbon catalysis by supported metals [4] and has been
3
6
35
used subsequently with the radiotracers [ Cl] and [ S] for a
variety of applications. It has been particularly useful in
probing Lewis acid±base interactions at ¯uorinated surfaces
CCl FCClF ! CCl CF
ꢀ1)
ꢀ2)
ꢀ3)
2
2
3
3
2
CCl3CF3 ! CCl2FCF3 ‡ CCl3CClF2
[5±7] and in monitoring the progress under static conditions
CCl2FCF3 ‡ 2H2 ! CH2FCF3 ‡ 2HCl
of catalytic reactions involving ¯uorine-containing com-
pounds, for example the CsF catalysed chloro¯uorination
of sulfur tetra¯uoride [8,9].
b-AluminiumꢀIII) ¯uoride, which has the hexagonaltung-
stenbronzestructure[3],isagoodcatalystfortheconversionof
CCl FCClF to CCl CF at 633 K and has reasonable selec-
2
2
3
3
tivity for the desired product, CCl FCF [1]. Experiments
2
3
under ¯ow conditions using the related catalyst, g-alumina
2. Results and discussion
*
A study of the catalytic behaviour of b-aluminiumꢀIII)
uoride towards 1,1,2-trichlorotri¯uoroethane under static
Corresponding author. Tel.: ‡44-141-3305134;
¯
conditions was carried out to aid the interpretation of the
fax: ‡44-141-3304888.
E-mail address: johnwin@chem.gla.ac.uk ꢀJ.M. Winfield).
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 1 ) 0 0 5 2 7 - 9

226
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
Table 1
Reaction of CCl
2
FCClF
2
with b-AlF
3
and fluorinated g-alumina
a
b
Catalyst
Reaction conditions
Temperature ꢀK)
Conversion ꢀ%)
Retention ꢀ%)
Time ꢀh)
CCl
2
FCCIF
2
ꢀmmol)
112 a
113
113 a
114
114 a
115
b-AlF
F-g-alumina
3
613
643
24
24
3.68
4.10
2.4
1.7
4.0
0
18.7
43.1
3.0
2.2
52.6
38.6
0.8
1.1
18.5
13.3
a
Determined by 19F NMR spectroscopy; defined as: mmol products/mmol reactant 113. Key for compounds: 115, CClF
FCF ; 113, CCl FCClF ; 113a, CCl CF ; 112a, CCl CClF
Defined as: ꢀmmol identified products À mmol reactant 113)/reactant 113. Shown by GC±MS and 19F NMR to contain CCl
2
CF
3
; 114, CClF CClF
2
2
; 114a,
Cl
CCl
2
3
2
2
3
3
3
2
.
b
3
CClF , C
2
2
6
Cl and C
2
4
.
radiotracer experiments, since radiotracer experiments per
se do not provide chemical identi®cation of the reaction
products. Therefore, it was necessary to establish the
detailed chemistry under conditions that would be similar
to those used in the radiotracer experiments to be described in
the following paragraphs. The behaviour of g-alumina, pre-
viously ¯uorinated with sulfur tetra¯uoride, was also exam-
ined, since previous work under ¯ow conditions has shown
that the behaviour of both the materials is very similar [1].
Under static conditions, 1,1,2-trichlorotri¯uoroethane
was converted to a mixture of other members of the series,
not change the observed behaviour to any great extent.
Although there was a build-up of less volatile products,
identi®ed as a mixture of CCl CClF , C Cl and C Cl , in
3
2
2
6
2
4
the metal vessel over the course of each series of reactions,
this did not appear to inhibit the catalytic activity. Product
distributions obtained using ¯uorinated g-alumina were rela-
tively constant over a series of multiple additions. Typical
product composition data, obtained from GC analysis, were
C Cl F isomers 38±56%, C Cl F isomers 60±41% and
2
3
3
2
2 4
CClF CF 2±3%. The organic material retained, quanti®ed
2
3
by mass balance and expressed as a fraction of the mass of the
reactant CCl FCClF , rose from 8.3% after the ®rst addition
C Cl F , by contact with b-AlF or g-alumina, which had
2
6Àn
been ¯uorinated with SF , also under static conditions.
n
3
2
2
to an approximately constant value of 25% in subsequent
experiments of the series. The corresponding results obtained
4
Conversion was essentially complete at 613±643 K within
2
4 h ꢀTable 1). The product distributions were similar to
those reported from catalytic reactions under ¯ow conditions
1] in that both catalysts favoured the formation of the
asymmetric isomers, CCl CF and CCl FCF . The major
using b-AlF were more variable. However, apart from one
3
experiment, the fraction of organic material retained was
always in the range of 13±47%. With neither catalyst, how-
ever, was there any discernible relationship between the
product distribution and the quantity of material retained.
Therefore, it was concluded that the variability in the beha-
viour was due to physical rather than chemical reasons.
[
3
3
2
3
products formed in the presence of ¯uorinated g-alumina
were CCl CF followed by CCl FCF , with the reverse order
being found over b-AlF ꢀTable 1). In neither case were
substantial quantities of the more highly ¯uorinated product
ꢀCClF CF ) observed but a mixture of compounds,
C Cl Fn with n < 3 was apparent. They were not removed
completely from the reaction vessel by vacuum distillation at
ambient temperature ꢀTable 1) but could be removed by
distillation above room temperature under dynamic vacuum.
Repetition of the reactions, up to six times, using the same
3
3
2
3
3
3
6
A sample of b-AlF was exposed to a [ Cl]-labelled
3
mixture of CCl FCClF , CCl CF , CClF CClF and
CCl FCF ꢀmolar ratio 40:24:18:18) under comparable con-
2 3
2
3
2
2
3
3
2
2
2
6Àn
ditions in a silica, in situ Geiger±M uÈ ller counting vessel.
3
6
[ Cl] Surface count rates ꢀdue to self-absorption of the
3
6
À
[ Cl] b emission, radiation from the bulk solid is not
detected in this type of experiment) obtained at various
points during the exposure are summarised in Table 2.
sample of b-AlF or ¯uorinated g-alumina in each case, did
3
Table 2
Variation of the [36Cl] count rate during the exposure of a [36Cl]-labelled chlorofluoroethane mixture ꢀ22.7 kPa) to b-AlF
3
ꢀ5.95 mmol) at 523±73 K
[36Cl] Surface count rate ꢀcount min
À1
)
Vapour phase observation
Experimental step
a
CFC mixture added
409
1350
321
Heating for 20 h then cool to room temperature
Volatile material removed
b
3 3 2 3
Proportions of CCl CF and CCl FCF increased
c
Second aliquot of CFC added
815
Heating for 20 h then cool to room temperature
Volatile material removed
2648
Colourless liquid apparent on vessel walls
b
CCl CF and CCl FCF
3 3 2 3
Pump
After 5 min
After 2 h
After 6 h
1408
1003
Background
a
b
c
À1
.
2
2.7 kPa, count rate 6081 count min
Observation using FTIR.
.9 kPa.
8

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
227
3
6
Addition of the [ Cl]-labelled mixture led to the immediate
3
6
detection of [ Cl] activity on the surface and the count rate
increased signi®cantly after heating at 523±573 K for 20 h.
However, the surface count rate decreased markedly when
the volatile material was removed at room temperature under
static vacuum. Although the composition of the volatile
material was not determined quantitatively, FTIR examina-
tion suggested that the proportions of CCl CF3 and
3
CCl FCF had increased in the vapour. The addition of a
2
3
36
second aliquot of the [ Cl]-labelled mixture resulted in a
similar behaviour; however, additionally, a colourless liquid
condensed on the surface of the vessel after heating. The
[36
Cl] count rate from the surface decreased slowly as the
cell was pumped under dynamic vacuum and a background
count was recorded after pumping for 6 h ꢀTable 2).
By analogy with the observations made using unlabelled
3
6
compounds, it is considered that the [ Cl] activity detected
on the surface as the reaction proceeded, originated from
[36
the series, C Cl F . Formation of Al±Cl bonds did not
Cl]-CCl CClF and other relatively involatile members of
3 2
2
6Àn
n
3
6
occur, since the [ Cl] count rate determined from the
surface returned to background at the conclusion of the
experiment ꢀTable 2).
These observations indicate that the isomerisation of
CCl FCClF to CCl CF occurs by an intramolecular pro-
2
2
3
3
cess, a situation that has also been found for isomerisation on
uorinated chromia [10] and on partially ¯uorinated alumi-
Fig. 1. Room temperature adsorption isotherm of ꢀa) [36Cl]-CCl
b) [36Cl]-CCl
2 3
FCF and
¯
ꢀ
2
2
3
FCClF on b-AlF .
niumꢀIII) chloride [11,12]. By analogy with isomerisation
on ¯uorinated chromia [10], it is concluded that, although
CCl FCClF may be dissociatively adsorbed via C±F bond
3
6
[ Cl]-labelledCCl FCF ,CCl FCClF andCCl CF towards
2
2
2
3
2
2
3
3
®
ssion, the surface F that is formed does not become
b-AlF was examined at room temperature. The Geiger±
3
36
equivalent to surface Al±F species. The [ Cl] results are
also consistent with the formation of CCl CClF and
CCl FCF directly from CCl CF , by a process in which
halogen exchange involving the replacement of F on the b-
AlF surface by Cl is not involved. This is consistent with the
dynamic behaviour of CCl FCClF over CCl F -¯uorinated
M uÈ ller direct monitoring technique enables adsorption iso-
3
6
therms to be determined by derivation of the [ Cl] surface
count rate over a range of pressure, the latter being re¯ected
3
2
2
3
3
3
3
6
by the [ Cl] count from the vapour above the surface.
The room temperature adsorption isotherm of [ Cl]-
36
3
CCl FCF on b-AlF ꢀFig. 1a) was indicative of physical
2
2
2
2
2
3
3
g-alumina under ¯ow conditions using very short contact
times [2]. Unfortunately, an analogous [ Cl] experiment
adsorption. Since the surface area of b-AlF is relatively
3
36
2
À1
small, 26.3 m g [1], the adsorption isotherm was not parti-
cularly well de®ned and long counting times were required
to minimise the counting errors. There was no evidence for
a chemical reaction, although when volatile material was
using SF -¯uorinated g-alumina proved impossible to per-
4
form due to contamination of the counters, presumably by
trace HF that is evolved from this material during heating.
However, the very similar product distributions obtained
using the two catalysts ꢀTable 1) imply that the reaction
pathway is identical. It appears that the behaviour shown by
these surfaces towards CCl FCClF is the result of Lewis
3
6
removed under static vacuum, the [ Cl] surface count took
several hours to decrease to the background value ꢀTable 3).
2
2
Table 3
Variation in the [36Cl] surface count rate during the exposure of [36Cl]-
acid sites that are strong enough, both to isomerise
CCl FClF and activate the CF group of CCl CF , enabling
a subsequent dismutation reaction to occur.
labelled 1,1-dichlorotetrafluoroethane ꢀ0.4 mmol) to b-AlF
at 295 K
3
ꢀ5.95 mmol)
2
2
3
3
3
a
Since these ®ndings were of key importance, additional
experiments were carried out. The behaviour of the three
compounds, CCl FCClF , CCl CF and CCl FCF , each
Time ꢀh)
[36Cl] Surface count rate
Vapour phase
observation
À1
ꢀ
count min
)
2
2
3
6
3
2
3
1 h
1140
3
b
individually labelled with [ Cl], towards b-AlF was
Vapour removed
After ca. 6 h
804
Background
CCl FCF
3
3
2
examined at 643 K, a temperature at which isomerisation/
dismutationbehaviour should be readily observed ꢀTable 1).
As a preliminary to these experiments, the behaviour of
a
À1
.
Initial count rate: 595 count min
By FTIR.
b

228
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
Table 4
Variation of the [36Cl] surface count rate during the exposure of [36Cl]-
labelled 1,1,2-trichlorotrifluoroethane ꢀ0.35 mmol) to b-AlF
at 295 K
3
ꢀ5.95 mmol)
a
Time ꢀh)
[36Cl] Surface count rate
Vapour phase
observation
À1
ꢀ
count min
)
1
1391
831
b
Vapour removed
2 2
CCl FCClF
2
3
170
74
24
Background
a
À1
Initial count rate: 511 count min
By FTIR.
.
b
Fig. 2. Variation of surface ꢀ&) and gas ꢀ^) [36Cl] count rates with
3
6
2 2 3
multiple additions of [ Cl]-CCl FCClF aliquots to b-AlF .
This observation and the results reported in the following
paragraphs, indicate that the adsorbate±adsorbent interac-
tion was stronger that might have been expected on the basis
of physical adsorption alone. It is concluded that some of the
heating indicated, qualitatively, that transformations to other
members of the series, C Cl F had occurred. This was
2
6Àn
particularly apparent in reactions that involved [ Cl]-
n
3
6
3
6
CCl FCF is strongly adsorbed. Exposure of [ Cl]-
2
3
3
6
CCl FCClF or [ Cl]-CCl CF ꢀTables 7 and 8). During
3
2
2
3
CCl FCClF to b-AlF under comparable conditions ꢀpres-
2
2
3
the reaction of CCl CF , a relatively involatile, colourless
3
3
sure in the range 0±6700 Pa) did not result in the observation
3
6
liquid formed on the surface of the counting vessel ꢀTable 8),
reminiscent of the observation made ꢀTable 2) during the
of a saturation [ Cl] surface count rate ꢀFig. 1b), unlike the
6
3
behaviour of [ Cl]-labelled CCl FCF ꢀFig. 1a). The
2
3
6
3
6
3
reaction of a [ Cl]-chloro¯uoroethane mixture.
adsorption isotherm determined for [ Cl]-CCl CF was
3
3
3
6
These experiments have provided important con®rmatory
evidence to support our previous kinetic studies [2]. The
results indicate that the behaviour of CCl FCClF on b-AlF
very similar to that for [ Cl]-CCl FCClF and in both
2
2
cases, adsorbed material was removed relatively slowly
from the surface ꢀTables 4 and 5). However, in neither case
was there any evidence for the formation of permanently
2
2
3
involves isomerisation to CCl CF ꢀEq. ꢀ1)) followed by
dismutation of the latter to give CCl FCF and CCl CClF
2
3
3
2
3
3
retained species or reaction products on b-AlF after expo-
3
ꢀEq. ꢀ2)). The formation of these directly from CCl CF ,
3 3
sure. For example, the effect of successive exposures of
Cl]-CCl FCClF aliquots to the same sample of b-AlF is
[36
shown in Fig. 2. Had even a small proportion of the adsorbed
without the intervention of other species, is indicated by the
behaviour of CCl CF itself as a reactant under identical
conditions.
2
2
3
3
3
36
species been permanently retained, the [ Cl] count rate from
the surface would have increased over the series of additions.
However, as in the case of CCl FCF , it is concluded that
Estimates of the free energy change for Eq. ꢀ2) indicate
that the reaction is favourable thermodynamically but it does
pose a mechanistic dif®culty, since the CF group is a very
2
3
3
some CCl FCClF and CCl CF is strongly adsorbed.
2
2
3
3
[36
of the vapour phase that were observed when [ Cl]-labelled
stable entity. As we have pointed out elsewhere [2], there are
numerous precedents in the literature for activation of a C±F
bond, although the great majority of cases are not particu-
larly good analogies for the reaction studied here. However,
Cl] Surface count rates and changes in the composition
36
CCl FCF , CCl FCClF or CCl CF were exposed to b-
2
3
2
2
3
3
AlF at 643 K are contained in Tables 6±8. In each case, a
3
3
6
signi®cant [ Cl] surface count was observed after heating a
3
6
Table 6
Variation in the [36Cl] surface count rate during the exposure of [36Cl]-
sample of a [ Cl]-labelled chloro¯uoroethane in the pre-
sence of b-AlF . This was reduced to the background count
rate on pumping. The IR spectra of the vapour phase after
3
labelled 1,1-dichlorotetrafluoroethane ꢀ0.3 mmol) to b-AlF
at 643 K
3
ꢀ5.95 mmol)
a
Experimental step
[36Cl] Surface count
Vapour phase
observation
Table 5
Variation in the [36Cl] surface count rate during the exposure of [36Cl]-
À1
rate ꢀcount min
)
labelled 1,1,1-trichlorotrifluoroethane ꢀ0.5 mmol) to b-AlF
at 295 K
3
ꢀ5.95 mmol)
2 3
CCl FCF addition
a
Heating for 22 h then cool 1329
Volatile material removed
[36Cl] Surface count rate
Major component:
b
Time ꢀh)
Vapour phase
observation
À1
CCl FCF
2
3
ꢀ
count min
)
Trace: CCl
CCl CClF
3 3
CF ,
b
1
1117
1054
3
2
b
Vapour removed
2
CCl
3
CF
3
Immediately
After 1 h
After ca. 6 h
695
651
660
Background
3
.5
Background
a
À1
a
À1
Initial count rate: 781 count min
By FTIR.
.
Count rate: 489 count min at 295 K.
By FTIR.
b
b

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
229
Table 7
Variation of the [36Cl] surface count rate during the exposure of [36Cl]-labelled 1,1,2-trichlorotrifluoroethane ꢀ0.4 mmol) to b-AlF
a
3
ꢀ5.95 mmol) at 643 K
[36Cl] Surface count rate ꢀcount min
À1
)
Vapour phase observation
Experimental step
CCl FCClF addition
2
2
Heating for 22 h then cool
Volatile material removed
2133
b
Major components: CCl
2
FCClF
CClF
2
, CCl
3
CF
3
, CCl
2 3
FCF
b
Minor components: CClF
2
, CClF CF
2 2
3
, CCl
3
CClF
2
, CCl
2 2
FCCl F
Immediately
After 1 h
After 3 h
982
780
633
After ca. 6 h
After pumping
306
Background
a
À1
Count rate: 635 count min at 295 K.
By FTIR.
b
it is plausible to propose that the adsorbed state of CF CCl
3
FTIR results obtained ꢀTables 7 and 8). Further reaction of
CCl CClF is also evident and a dismutation appears to be
an appropriate description in this case ꢀEq. ꢀ5)).
3
d‡
dÀ
on b-AlF is Cl3CC F2 Á Á Á F Á Á Á Al . Dismutation requires
3
3
2
adsorption of two CCl CF molecules at the neighbouring
3
3
sites followed by a concerted interchange of F from the aC
of one molecule to the bC of the other and Cl from bC to aC
in the reverse sense. From an operational point of view, it is
considered that the Lewis acidity of the surface site on b-
AlF at which CCl CF is adsorbed must be comparable
2
CCl2FCF3 ! CClF2CF3 ‡ CCl3CF3
CCl3CClF2 ! CCl2FCClF2 ‡ CCl3CCl2F
ꢀ4)
ꢀ5)
2
Further reaction of CCl CCl F would lead to C Cl , either
3
3
3
3
2
2
6
with that of a molecular strong Lewis acid such as antimony
penta¯uoride. For example, in a very recent report describ-
ing the SbF -catalysed alkylation of poly¯uoroole®ns by
via dismutation or by chlorination with other highly chlori-
nated chloro¯uoroethanes, and C Cl is formed from C Cl
6
2
4
2
by thermal decomposition.
5
hydro¯uorocarbons, it has been proposed that the ®rst step in
the reaction involves the formation of an activated complex,
The use of the term ``dismutation'', which we de®ne here
as a concerted ligand exchange reaction between two iden-
tical species without change in the oxidation state of the
central atoms involved, to describe the inter-conversions that
are characteristic of catalysed reactions involving chloro-
¯uoroethanes, has a long history. Although the term ``dis-
proportionation'' has often been used as a synonym, we
prefer to restrict its use to situations where changes in
oxidation states occur.
d‡
dÀ
for example, FC H Á Á Á F Á Á Á Sb F5. It is further proposed
2
that this reacts with the ole®n to generate a carbocation,
¯
uoroantimonate intermediate [13]. Interestingly, CF X,
3
X ˆ H, Br, Cl or I, apparently cannot be activated using
SbF but activation is possible using solid aluminiumꢀIII)
5
chloro¯uoride, AlCl0.2±0.1F2.8±2.9, as a catalyst [14]. A com-
parative study of solid aluminiumꢀIII) ¯uoride Lewis acids
with SbF is an area that would repay further study, parti-
Catalytic halogen inter-conversions among chloro¯uor-
oethanes were apparently ®rst proposed by Tatlow and
co-workers in 1974 to rationalise product distributions
that were the result of the ¯uorination of C Cl by HF
5
cularly since a theoretical basis for comparisons among
molecular Lewis acid ¯uorides has been established by
Christe et al. [15].
2
6
Further dismutations on b-AlF are possible. Dismutation
in the presence of an aluminium ¯uoride catalyst [16].
Subsequently, Kolditz et al. proposed dismutation reac-
tions combined with isomerisation, to account for product
3
of CCl FCF ꢀEq. ꢀ4)), although apparently less facile than
dismutation of CCl CF , does appear to occur, based on the
2
3
3
3
Table 8
Variation in the [36Cl] surface count rate during exposure of [36Cl]-labelled 1,1,1-trichlorotrifluoroethane ꢀ0.7 mmol) to b-AlF
a
3
ꢀ5.95 mmol) at 643 K
[36Cl] Surface count rate ꢀcount min
À1
)
Vapour phase observation
Experimental step
CCl CF addition
3
3
Heating for 22 h then cool
Volatile material removed
4058
b
Major component: CCl
2 3
FCF
b
Minor component: CCl
b
3
CF
3
2 3
Trace: CClF CF
Liquid formed on the counting cell surface
Immediately
After 1 h
After ca. 6 h
2034
1075
Background
a
À1
Count rate: 1123 count min at 295 K.
By FTIR.
b

230
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
distributions within the C Cl F , n ˆ 1±5, series that were
the gas phase counts after correction for background. Solid
samples ꢀ0.5 g) were introduced into the vessel contained in
a Pyrex boat connected to a Pyrex rod containing a soft iron
bar. The arrangement enabled the boat to be moved with the
aid of a magnet within the vessel after evacuation. Samples
were spread as thinly as possible in order to approach the
required criterion of an in®nitely thin solid layer. Cell and
solids were thoroughly degassed before a measured pressure
2
6Àn
n
observed over aluminium ¯uoride or chromia catalysts
17,18]. Marangoni et al. followed a similar approach to
account for the transformations of C Cl F isomers on
[
2
2 4
¯uorinated chromia [19,20]. The behaviour of C Cl F
2 3 3
isomers on a variety of supported chromia catalysts and
on AlF has been discussed also by Blanchard and co-
3
workers [21±23] in terms of isomerisation and dismutation.
They concluded that the acidity of the catalyst active site is
an important factor in determining the reaction pathway
followed [23]. However, dismutation reactions are not
3
6
of [ Cl]-C Cl Fn, was added via a calibrated gas-hand-
2
6Àn
ling manifold. The system was allowed to equilibrate and
counts from the two Geiger±M uÈ ller tubes recorded simul-
taneously. Counting times were chosen to enable substantial
appropriate in every case. The distributions of the isotopes,
[36
18
4
Cl] and [ F], in the individual components of chloro-
uoroethane mixtures obtained by ¯ow of CCl FCClF or
counts ꢀnormally 10 to minimise the counting errors) to be
¯
CClF CClF over [ Cl]- or [ F]-labelled ¯uorinated
accumulated. Volatile material was removed under static
vacuum and the process repeated using a different pressure.
Pressures of volatile components were in the range of 1300±
6700 Pa. At the conclusion of the determination of an
adsorption isotherm or of a reaction sequence, volatile
material was removed by distillation under static or dynamic
2
2
3
6
18
2
2
chromia catalysts are inconsistent with the occurrence
of concerted dismutation processes. In this instance, non-
concerted halogen exchange, F for Cl and Cl for F, reactions
that involve catalyst surface halide, are more appropriate
descriptions [10,24].
3
6
vacuum as appropriate and the count from [ Cl] material
retained on the solid determined.
3
6
The behaviour of [ Cl]-labelled chloro¯uoroethanes
towards b-AlF at elevated temperatures was studied in a
3
. Conclusions
3
similar fashion using a silica counting vessel equipped with
a heated zone remote from the counters, the solid being
moved into or out of the heated zone using a magnet. All
counts were recorded at ambient temperature.
The reaction of CCl FCClF on the surface of b-AlF
2
2
3
occurs via an intramolecular isomerisation process to give
CCl CF . Further reaction of CCl CF occurs via dismuta-
3
3
3
3
tion rather than by halogen exchange. This appears to be a
characteristic feature of b-AlF and presumably also of
3
6
4.2. Preparation of [ Cl]-labelled compounds
3
¯
group CCl CF that is required for dismutation implies that
uorinated alumina catalysts. The activation of the CF3
The conventional route to members of the chloro¯uor-
oethane series by reaction between C Cl , or a mixture of
3
3
the b-AlF surface has very strong Lewis acid sites.
3
2
6
C2Cl4 ‡ Cl2, and anhydrous HF, catalysed by ¯uorinated
chromia under ¯ow conditions [25], was modi®ed for use
under static conditions in order to facilitate the use of
isotopically labelled reagents.
4
. Experimental
4
.1. Experiments with chlorine-36
All operations were carried out using a Monel vacuum
system or an N -®lled glove box. The reaction vessel was a
2
3
Monel pressure vessel ꢀ90 cm ). A sample of commercial,
The behaviour of b-aluminiumꢀIII) ¯uoride towards
[36
Cl]-labelled chloro¯uoroethanes was examined using a
Geiger±M uÈ ller direct monitoring method, developed in
amorphous chromia ꢀ1.0 g), previously calcined at 523 K in
vacuo for 24 h, was pre-¯uorinated with three aliquots
ꢀ8.4 mmol) of anhydrous HF ꢀFluorochem, 99%) at 523 K
for 24 h, unused HF being removed before admission of the
next aliquot. The vessel containing the catalyst was charged
with C Cl ꢀ1.0 g; 4.27 mmol) in the glove box and then,
1
4
Glasgow for [ C] adsorption measurements [4] and subse-
quently used to study heterogeneous systems involving
¯
À
36
35
uorides where the b emitting isotopes, [ Cl] and [ S],
were employed as probes [5±9]. An evacuable Pyrex count-
ing vessel with a gas handling facility was used for mea-
surements at ambient temperature. Two Geiger±M uÈ ller
counters were positioned within the vessel to enable
2
6
after connection to the vacuum line, with suf®cient HF to
give the required molar ratio. This was varied in different
experiments using C Cl :HF molar ratios in the range 1:2 to
2
6
[36
Cl] activity from the vapour phase and from the vapour
1:4. The mixture was allowed to react using temperatures in
the range 423±573 K. Reaction times were in the range 4±
72 h. At the end of the reaction period, the vessel was
allowed to cool to room temperature and the volatile con-
tents distilled onto NaF, contained in a Monel or stainless
steel vessel. The mixture was allowed to contact NaF for
suf®cient time for unchanged HF to be sorbed, then the
volatile material was transferred to an evacuated Pyrex
À
36
plus surface ꢀdue to self-absorption of the b emitter [ Cl],
activity from the bulk solid was not detected) to be mon-
itored concurrently. The counting tubes were inter-cali-
3
6
brated using H Cl at several pressures, counts being
recorded simultaneously on two scalers. These enabled
[36
Cl] counts from the surface of a solid placed below
one of the counters to be determined by subtraction of

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
231
ampoule for analysis by FTIR spectroscopy, 19F NMR
spectroscopy and GC.
meric mixtures were present, since, using GC, base line
separation of C Cl F and C Cl F isomers was dif®cult to
2
3
3
2
2 4
The composition of the product depended on the reaction
variables in a complex fashion. For example, with an
HF:C Cl molar ratio of 4:1 at 673 K, the product mixture
achieve. Less volatile organic products were removed by
heating the reaction vessel while pumping through a cooled
U-trap. The components present were identi®ed by 19F NMR
and GC±MS ꢀHP 5971, quadrupole detector interfaced with
an HP 5890 series II chromatograph).
2
6
contained predominantly di¯uorinated species. Similar
results were obtained with a 3:1 molar ratio at 573 K
for 5 h or at 423 K for 27 h. The most useful reaction
conditions were HF:C Cl ˆ 3:1, temperature 548±573 K
The effect of successive additions ꢀup to six) of
CCl FCClF to a sample of b-AlF or ¯uorinated g-alumina
2
6
2
2
3
and time 48 h. The major components of product mixtures
were CCl FCClF , CCl CF , CClF CClF and CCl FCF .
Reactions using C Cl :Cl ꢀ1:1 molar ratio) gave similar
was examined in a similar fashion, volatile products being
analysed by GC. In these experiments, the less volatile
material was not removed between each addition but was
examined at the end of a reaction sequence.
2
2
3
3
2
2
2
3
2
4
2
mixtures. Although the proportions of each component
varied somewhat from reaction to reaction, the ranges,
CCl FCClF ꢀ19±40 mol%), CCl CF ꢀ11±35 mol%),
4.4. Catalyst preparation
2
2
3
3
CClF CClF ꢀ18±30 mol%) and CCl FCF ꢀ8±27 mol%)
3
2
2
2
were de®ned suf®ciently for the experiments in which they
were used. Minor components containing two or less F
atoms were readily removed by distillation in vacuo.
Mixtures labelled with chlorine-36 were prepared in an
identical fashion from 36ClCl, which was prepared by the
g-AluminiumꢀIII) ¯uoride was prepared by the tempera-
ture programmed dehydration [3] of AlF Á3H O ꢀ5.0 g,
3
2
À1
3
Aldrich, purity 97%) under He ꢀ30 cm min ), heating
À1
À1
from room temperature to 493 K at 5 K min , held at
493 K for 1 h, heating to 723 K at 10 K min , held at
723 K for 2 h, ®nally allowed to cool to room temperature
3
6
À
oxidation of [ Cl] anion ꢀAmersham International,
9
250 kBq diluted with concentrated aqueous HCl), by
À
under He ¯ow. The b-AlF so formed was transferred in a
3
[
and H O were removed by trap to trap distillation over solid
MnO4] in aqueous acidic conditions [26,27]. Trace HCl
sealed vessel to a N glove box ꢀH O ca. 1 ppm) where
subsequent manipulations were performed. Its identity was
con®rmed by XRD. Samples prepared in Berlin or Glasgow
2
2
2
KMnO then P O and the product, contained in a stainless
4
2 5
2
showed identical behaviour. Its BET area was 26.3 m g
À1
steel vessel, was stored over P O in vacuo.
2
.
5
Individually labelled chloro¯uoroethanes were prepared
according to the previously reported method [27] by photo-
chemical ꢀmedium pressure Hg lamp) irradiation at room
temperature. Mixtures of the appropriate hydro¯uoroethane
or hydrochloro¯uoroethane precursor and 36ClCl were irra-
diated for 24 h in a silica ampoule ®tted with a Pyrex/PTFE
stopcock ꢀJ. Young). The conditions required for complete
Fluorinated g-alumina was prepared under static condi-
tions in a Monel metal pressure vessel ꢀHoke, 90 cm )
3
attached to a Monel vacuum line [28,29]. Typically, g-
2
À1
alumina ꢀDegussa, BET area ˆ 110 m g , 1.5 g) pre-
viously caked, sieved to produce 500±1000 mm particles
and calcined in vacuo for 8 h at 523 K, was allowed to react
with SF ꢀ9.0 mmol, 99%, Fluorochem) for 2 h, nominally at
4
reaction were established using inactive Cl ; no impurities
room temperature. Volatile products, a mixture of OSF and
2
2
were detected using FTIR or 19F NMR spectroscopy after
fractionation of the product mixtures in vacuo. Reaction
stoichiometries were for CCl FCClF :CH FCHF :Cl ꢀ1:3);
SO , whose components were identi®ed by FTIR spectro-
2
scopy, were removed by distillation and the process repeated
twice. The product, an off-white solid, was transferred to and
handled subsequently in a glove box. It could be stored for
short periods in FEP; storage in Pyrex led to etching,
indicating that HF was lost slowly from the solid. For this
reason, smaller quantities ꢀ0.5 g) were prepared for use in
situ, with the appropriate adjustment of the quantity of SF4.
Single point determinations of BET area ꢀCoulter SA 3100
2
2
2
2
2
for CCl CF :CHCl CF :Cl ꢀ1:1); for CCl FCF :CH FCF :
3
3
3
2
3
2
2
3
2
Cl ꢀ1:2).
2
4
.3. Reactions between 1,1,2-trichlorotrifluoroethane
and b-AlF or fluorinated g-alumina under
3
static conditions
2
À1
instrument) gave values in the range 80±90 m g . The
imprecision was possibly a result of the corrosive nature of
the material. Fluorine content was not determined directly
1,1,2-Trichlorotri¯uoroethane ꢀca. 0.5 g), previously
Ê
degassed and dried by storage over 4 A molecular sieves,
18
was added to b-AlF or ¯uorinated g-alumina ꢀ0.5 g in each
but a value of ca. 22% was inferred from a previous [ F]
study of the ¯uorination carried out under very similar
conditions [28].
3
3
case) contained in a stainless steel pressure vessel ꢀ75 cm ,
Hoke) and attached to a Pyrex vacuum line. Mixtures were
heated at 613±643 K for 24±72 h. Volatile products were
removed by distillation in vacuo at room temperature,
weighed, and analysed by FTIR ꢀNicolet Impact 410),
GC ꢀVarian 3400, FID, Chrompak 30 m capillary column)
and 19F NMR spectroscopy ꢀBruker WP200 SY). The latter
was particularly useful for quantitative analyses when iso-
Acknowledgements
We thank the European Union for support of this work
under Contract No. ENV4-CT97-0601.

232
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 112 ꢀ2001) 225±232
[
16] M. Vecchio, G. Groppelli, J.C. Tatlow, J. Fluorine Chem. 4 ꢀ1974)
17±139.
References
1
[
[
[
[
[
[
[
1] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M.
Winfield, J. Fluorine Chem. 107 ꢀ2001) 45±52.
[17] L. Kolditz, G. Kauschka, W. Schmidt, Z. Anorg. Allg. Chem. 434
ꢀ1977) 41±54.
2] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M.
Winfield, J. Fluorine Chem. 110 ꢀ2001) 181±189.
[18] L. Kolditz, U. Calov, G. Kauschka, W. Schmidt, Z. Anorg. Allg.
Chem. 434 ꢀ1977) 55±62.
3] A. Le Bail, C. Jacoboni, M. Leblanc, R. De Pape, H. Duroy, J.L.
Fourquet, J. Solid State Chem. 77 ꢀ1988) 96±101.
[19] L. Marangoni, C. Gervasutti, L. Conte, J. Fluorine Chem. 19 ꢀ1981/
82) 21±34.
4] A.S. Al-Ammar, G. Webb, J. Chem. Soc., Faraday Trans. 174 ꢀ1978)
[20] L. Marangoni, D. Carmello, R. Passerini, Chim. Ind. ꢀMilan) 67
ꢀ1985) 467±480.
1
95±205.
5] K.W. Dixon, J.M. Winfield, J. Chem. Soc., Dalton Trans. ꢀ1989)
37±942.
6] T. Baird, A. Bendada, G. Webb, J.M. Winfield, J. Mater. Chem. 1
1991) 1071±1077.
7] T. Baird, A. Bendada, G. Webb, J.M. Winfield, J. Fluorine Chem. 66
1994) 117±122.
[21] D. Bechadergue, M. Blanchard, P. Canesson, Appl. Catal. 20 ꢀ1986)
179±187.
9
[22] D. Bechadergue, M. Blanchard, P. Canesson, in: Heterogeneous
Catalysis and Fine Chemicals, in: M. Guisnet, et al. ꢀEds.), Elsevier,
Amsterdam, 1988, pp. 257±264.
ꢀ
ꢀ
[23] M. Blanchard, L. Wendlinger, P. Canesson, Appl. Catal. 59 ꢀ1990)
123±128.
[
[
8] G.A. Kolta, G. Webb, J.M. Winfield, Appl. Catal. 2 ꢀ1982) 257±268.
9] T. Baird, A. Bendada, M. Selougha, G. Webb, J.M. Winfield, J.
Fluorine Chem. 69 ꢀ1994) 109±113.
[24] L. Rowley, G. Webb, J.M. Winfield, A. McCulloch, Appl. Catal. 52
ꢀ1989) 69±80.
[
[
[
[
10] L. Rowley, J. Thomson, G. Webb, J.M. Winfield, A. McCulloch,
Appl. Catal. A: Gen. 79 ꢀ1991) 89±103.
11] W.T. Miller Jr, E.W. Fager, P.H. Griswald, J. Am. Chem. Soc. 72
[25] B. Kemnitz, J.M. Winfield, in: Advanced Inorganic Fluorides, in: T.
Ï
Nakajima, B. Zemva, A. Tressaud ꢀEds.), Elsevier, Amsterdam, 2000,
pp. 367±401 ꢀChapter 12).
ꢀ1950) 705±707.
[26] G. Brauer, Handbook of Preparative Inorganic Chemistry, 2nd
Edition, Vol. 1, Academic Press, New York, 1965, pp. 272±274.
[27] L. Rowley, G. Webb, J.M. Winfield, J. Fluorine Chem. 38 ꢀ1988)
115±117.
12] D.G. McBeth, M.M. McGeough, G. Webb, J.M. Winfield, A.
McCulloch, N. Winterton, Green Chem. 2 ꢀ2000) 15±20.
13] G.G. Belen'kii, V.A. Petrov, P.R. Resnick, J. Fluorine Chem. 108
ꢀ
2001) 15±20.
[28] A. Bendada, G. Webb, J.M. Winfield, Eur. J. Solid State Inorg. Chem.
33 ꢀ1996) 907±916.
[
[
14] V.A. Petrov, C.G. Krespan, J. Fluorine Chem. 102 ꢀ2000) 199±204.
15] K.O. Christe, D.A. Dixon, D. McLemore, W.W. Wilson, J.A. Sheehy,
J.A. Boatz, J. Fluorine Chem. 101 ꢀ2000) 151±153.
[29] J. Thomson, G. Webb, J.M. Winfield, D. Bonniface, C. Shortman, N.
Winterton, Appl. Catal. A: Gen. 97 ꢀ1993) 67±76.