Dynamic behaviour of 1,1,2-trichlorotrifluorethane on aluminium(III) chloride and related surfaces

Green Context

The safe destruction of CFCs is one of the major current issues in halocarbon chemistry. It seems likely that their production and consumption will be prohibited by the Montreal Protocol and related agreements. The remarkable volume of illegal traffic in these substances adds to the urgency to devise effective methods of destruction. Ideally their destruction should lead to other useful products and among the more likely are the more environmentally acceptable hydrofluorocarbons. This paper is concerned with one of the more important reactions that can be utilised in trying to achieve these goals—the isomerisation of 1,1,2-trichlorotrifluoroethane.

JHC

Summary

The isomerization of CCl₂FCClF₂ to CCl₃CF₃ is one of the steps in a possible scheme of reactions to convert CCl₂FCClF₂ to the more environmentally-acceptable refrigerant CF₃CH₂F. A reinvestigation of the isomerization in the presence of solid aluminium(III) chloride at ambient temperature indicates that the catalytically active sites for isomerization are generated in situ by the behaviour of aluminium(III) chloride as a chlorinating reagent towards CCl₂FCClF₂. Aluminium(III) chloride pretreated with CH₃CCl₃ exhibits very similar behaviour but γ-alumina chlorinated with CCl₄ shows little or no isomerization activity.

Introduction

One of the imperatives that underlie current research in the halocarbon field is the need to devise suitable methods for the destruction or conversion of compounds such as chloro- and bromo-fluorocarbons (CFCs and Halons) whose production and consumption may be prohibited by legislation following the Montreal Protocol and its successors. The methods selected should be capable of being operated on a reasonably large scale and are required urgently, not least as a discouragement to the illegal traffic in CFCs that has developed.¹ It is important that conversion processes do not result in the risk of toxic or ozone-depleting emissions and, ideally, they should lead to useful products. The literature relating to catalytic² or non-catalytic³ methods for the oxygenation of CFCs is now extensive, the products being CO₂ and hydrogen halides. Catalytic hydrogenolysis, for example the conversion of CCl₂F₂ to the useful, low-temperature refrigerant CH₂F₂, is also attractive and this reaction is receiving considerable attention.^4,5

Existing stocks of 1,1,2-trichlorotrifluoroethane held world-wide may be considerable, in view of its former use in a wide variety of industrial applications. In principle, the hydrogenolysis route is applicable here, Scheme 1. Steps i–iii of this scheme form part of the route used in a large scale process for the preparation of CFC-alternative refrigerant CF₃CH₂F (HFC-134a) from HF + C₂Cl₄/Cl₂⁶ and aspects of the catalytic fluorination⁷ and hydrogenolysis⁸ reactions have received fundamental study. As an alternative, CCl₂FCClF₂ could be converted to CCl₂FCF₃ and hence to CH₂FCF₃, via the isomerization CCl₂FCClF₂ → CCl₃CF₃ followed by fluorination of CCl₃CF₃ to CCl₂FCF₃, steps iv and v of Scheme 1.

Scheme 1 Possible routes from CCl₂FCClF₂ to CH₂FCF₃. (i) Fluorination with HF, fluorinated chromia catalyst, refs. 6 and 7. (ii) Isomerization, fluorinated chromia catalyst, refs. 6 and 7. (iii) Hydrogenolysis, Pd/C catalyst, ref. 8. (iv) Isomerization, see text. (v) Fluorination, see text.

Here, attention is focused on the conditions required to achieve facile isomerization of CCl₂FCClF₂. Mechanistically, this is an interesting reaction. We have previously shown that on heavily fluorinated chromia at 700 K the reaction occurs via an intramolecular process in contrast to other transformations, chlorination or fluorination, that are intermolecular.⁷

It has been known for many years that CCl₂FCClF₂ is isomerized extensively to CCl₃CF₃ when it is refluxed in the presence of aluminium(III) chloride⁹ and this has been the basis for a large scale batch process to prepare CCl₃CF₃ as a synthetic intermediate. Extensive chlorination of CCl₂FCClF₂ also occurs, the stated composition obtained from the reaction under laboratory conditions being CCl₃CF₃ (50%), CCl₃CClF₂ (40%), C₂Cl₆ (5%) and unchanged CCl₂FCClF₂ (5%). When [³⁶Cl]-labelled aluminium(III) chloride was used, the radiolabel was detected only in CCl₃CClF₂ and C₂Cl₆, suggesting that in this case the isomerization is also intramolecular.⁹ This observation was the starting point for the investigation reported here, undertaken at room temperature under vapour–solid heterogeneous conditions. The original observations have been confirmed but, although isomerization and chlorination are nominally separate processes, the latter is required to produce the catalytically active sites for isomerization. These are believed to be related closely to the recently reported strong Lewis acid aluminium chlorofluoride.¹⁰

Results

The behaviour of CCl₂FCClF₂ vapour over resublimed solid aluminium(III) chloride, aluminium chloride sublimed directly onto calcined γ-alumina and γ-aluminina previously chlorinated with carbon tetrachloride has been studied at room temperature under static, anhydrous conditions. In some instances solids were pretreated with 1,1,1-trichloroethane to determine the effect of a supported organic layer^11,12 on the subsequent behaviour of CCl₂FCClF₂. The components of product mixtures were identified by ¹⁹F NMR spectroscopy, GC and GCMS and, although the isomerization of CCl₂FCClF₂ under the conditions used was slow, its progress could be followed conveniently by quantitative IR spectroscopic measurements made over 24 h periods. Uptakes of CCl₂FCClF₂ and related CFCs by the solids were determined by mass balance measurements and changes occurring at the surface during series of catalytic reactions by direct Geiger-Müller monitoring¹³ in experiments where [³⁶Cl]-labelled CCl₂FCClF₂ or aluminium(III) chloride were used. The key observations made, which must be taken into account in formulating a model to explain the dynamic behaviour of CCl₂FCClF₂, are summarised below.

Mass balance and spectroscopic studies

Solid aluminium(III) chloride and aluminium(III) chloride sublimed directly onto calcined γ-alumina are active catalysts for CCl₂FCClF₂ isomerization at room temperature, although their activity is reduced markedly by the presence of even trace quantities of water vapour. In contrast, γ-alumina chlorinated with CCl₄ shows little or no catalytic ability even though it, like solid aluminium(III) chloride, is a strong Lewis acid capable of room temperature dehydrochlorination of CH₃CCl₃ and subsequent oligomerization of the CH₂[double bond, length half m-dash]CCl₂ produced.^11,12 Exposure of solid aluminium(III) chloride to CCl₂FCClF₂ vapour results in an immediate change in colour, colourless → pale yellow, and the retention of a substantial fraction of organic material which is removed only with difficulty by pumping at room temperature. After 24 h exposure, the major volatile product is CCl₃CF₃; minor products are the chlorinated species CCl₃CClF₂, CCl₃CCl₂F and C₂Cl₆ and traces of the fluorinated product, CCl₂FCF₃, are also present. However the symmetric isomers, CCl₂FCCl₂F and CClF₂CClF₂, were never observed in product mixtures. Analysis of the retained organic material indicated that C₂Cl₆ and CCl₃CClF₂ were present. Both compounds have very small vapour pressures at room temperature (C₂Cl₆, mp 463–468 K; CCl₃CClF₂, mp 314, bp 364 K) and their incomplete removal by static sublimation in vacuo does not necessarily indicate any significant interaction involving surface Al^III centres.

A similar pattern of behaviour was observed when samples of aluminium(III) chloride, were repeatedly exposed to aliquots of CCl₂FCClF₂ vapour (up to five) for 24 h periods. The results of two such experiments, including the experimental conditions used, are given in Table 1. In the first experiment complete consumption of CCl₂FCClF₂ was not achieved, although formation of its isomer, CCl₃CF₃, was always significant. In the second experiment, the highest conversions to CCl₃CF₃ were associated with prior addition of a CCl₂FCClF₂ aliquot from which the formation of highly chlorinated organics had been significant. Most importantly, there was no indication of a chlorine-to-fluorine mass balance, among the organic products, hence a dismutation of CCl₂FCClF₂ was unlikely. Analogous experiments using aluminium(III) chloride pretreated with CH₃CCl₃ led to very similar results (Table 2). There is no evidence that the highly unsaturated hydrochlorocarbon oligomers that are the result of the pretreatment,¹¹ inhibit the isomerization of CCl₂FCClF₂. The purple colour of the oligomer–aluminium(III) chloride surface was discharged during the course of an experiment, the solid becoming grey, and the proportion of highly chlorinated C₂ species in the volatile product mixture was somewhat higher (Table 2) than was the case using aluminium(III) chloride (Table 1). Interestingly, pretreatment of aluminium(III) chloride with aliquots of CCl₂FCClF₂, had no effect on its ability to catalyse CH₃CCl₃ dehydrochlorination.

Table 1 Mass balance data from multiple additions of CCl₂FCClF₂ to freshly sublimed aluminium(III) chloride^a

Retained	Volatile material^c (w/w%)
Addition	material^b
number	(w/w%)	CCl3CF3	CCl2FCClF2
a Reaction conditions: room temperature, 24 h, aluminium(III) chloride (expt. 1, 7.3 mmol, 0.976 g) or (expt. 2, 12.8 mmol, 1.710 g) sublimed directly into the reaction flask in vacuo, CCl2FCClF2, aliquots in the range 4–7 mmol.b Defined as (CCl2FCClF2 − volatile products) g × 100/(CCl2FCClF2) gc Defined as (product) g × 100/(CCl2FCClF2) g, individual components being determined using ^19F NMR. Other products observed after each addition were (w/w%): CCl2FCF3 ⩽ 1, CCl3CClF2 ⩽ 3, CCl3CCl2F trace; a build-up of C2Cl6 (identified by GCMS) on the solid was observed in each expt.
Expt.	1	25	69	trace
1	2	0	35	62
3	5	32	62
4	3	57	37
Expt.	1	70	29	0
2	2	85	13	trace
3	10	84	0
4	6	89	1
5	48	48	2

---

Table 2 Mass balance data from multiple additions of CCl₂FCClF₂ to freshly sublimed aluminium(III) chloride pretreated with CH₃CCl₃^a

Retained	Volatile material^c (w/w%)
Addition	material^b
number	(w/w%)	CCl3CF3	Cl2FCClF2	CCl3CClF2
a Reaction conditions: room temperature, 24 h, aluminium(III) chloride (sublimed directly into the reaction flask in vacuo then pretreated with CH3CCl3 vapour for 1 h at room temperature; quantities of aluminium(III) chloride and CH3CCl3 were respectively: 0.01 and 0.03 mol (expt. 1), 0.49 and 0.06 mmol (expt. 2) and 8.7 and 3.9 mmol (expt. 3).b See note b in Table 1.c Defined as in Table 1. Product analyses by ^19F NMR (expts.1 and 3) or GC/GCMS (expt. 2). Other products observed after each addition were CCl2FCF3 ⩽ 2, CCl3CCl2F ⩽ 1, C2Cl6 and C2Cl4 (both trace).
Expt. 1	1	45	24	6	6
2	4	37	53	1
3	3	36	54	3
4	82	15	trace	trace
5	95	0	0	0
6	56	43	0	0
7	98	0	0	0
Expt. 2	1	73	1	0	25
2	25	67	trace	3
3	38	59	trace	1
4	30	55	0	9
Expt. 3	1	25	72	0	trace
2	38	57	trace	3
3	37	55	1	2
4	75	17	trace	5

---

The decrease in CCl₂FCClF₂ in the vapour phase above solid aluminium(III) chloride, aluminium(III) chloride pretreated with one or more aliquots of CCl₂FCClF₂ or CH₃CCl₃ or γ-alumina-supported aluminium(III) chloride pretreated with CCl₂FCClF₂, were inversely related to the formation of CCl₃CF₃ in the vapour phase as demonstrated by IR spectroscopic measurements over 24 h periods. In all cases three isosbestic points were observed, indicating a direct relationship between consumption of CCl₂FCClF₂ and formation of CCl₃CF₃. However, neither of these processes followed simple first or second order kinetics under the conditions used.

The isomerizations of both CCl₂FCClF₂ and CClF₂CClF₂ are catalysed by fluorinated chromia at ca. 700 K under flow conditions,⁷ but there was no evidence that isomerization of CClF₂CClF₂ occurred at room temperature in the presence of aluminium(III) chloride. No interaction was observed between solid aluminium chloride and CClF₂CClF₂ or n-C₆F₁₄ (which has physical properties similar to CCl₂FCClF₂) but a mixture of solid CCl₃CClF₂ and aluminium(III) chloride did react slowly to give a mixture of CCl₃CF₃, CCl₃CCl₂F, CCl₂FCF₃ and CCl₂FCClF₂. It appears therefore that only chlorofluoroethanes containing CCl_xF_3−x, x = 2 or 3, groups interact with aluminium(III) chloride.

[³⁶Cl]-Radiotracer studies

The extent to which chemical reactions occurred at the surface of aluminium(III) chloride was highly dependent on the degree of hydration/hydroxylation of the surface. For example, the effect of trace H₂O, adventitious or deliberately added, on [³⁶Cl] exchange between H³⁶Cl vapour and resublimed aluminium(III) chloride samples is shown in Table 3. The extent to which exchange was observed increased markedly when H₂O vapour was deliberately added to the surface prior to H³⁶Cl and was usually measurable when the reagents had been manipulated in a Pyrex vacuum system. It was demonstrated many years ago that no observable [³⁶Cl] exchange occurred between H³⁶Cl and solid aluminium(III) chloride when the materials were rigourously purified.¹⁴

Table 3 [³⁶Cl] Exchange between solid resublimed aluminium(III) chloride and H³⁶Cl at room temperature

Pretreatment of reagents	Fraction of [^36Cl] activity exchanged^a
a Defined as So − St/So − S∞, where So and St are the [^36Cl] specific count rates [counts s^−1 (mg AgCl)^−1] of H^36Cl before and after exposure to aluminium(III) chloride; S∞ is the specific count rate calculated on the basis of complete exchange.
H^36Cl distilled directly from P4O10	0.27, 0.24, 0
None	0.13, 0.36, 0.55, 0.59
AlCl3 exposed to H2O vapour	0.60, 0.78, 0.83

---

In contrast, exposure of solid aluminium(III) chloride to [³⁶Cl]-CCl₂FCClF₂ led to the immediate observation of [³⁶C]-activity on the solid, the count rate due to [³⁶Cl]-CCl₂FCClF₂ vapour decreasing by 41% over 25 min in agreement with mass balance measurements that indicated 42% retention by the solid. Subsequent addition of aliquots of non-radioactive CCl₂FCClF₂ had little effect on the surface [³⁶Cl] count rate. [³⁶Cl]-Activity in the vapour phase was too small to quantify precisely but IR spectroscopy showed that CCl₃CF₃ was the major component. A very slow desorption process occurred over a period of weeks in the counting vessel resulting in a small increase in surface count rate, attributable to a reduction in the β⁻ self absorption effect from the [³⁶Cl]-species on the solid. This was accompanied by the observation of C₂Cl₆ on the walls of the vessel.

Consistent with these observations, exposure of [³⁶Cl]-labelled solid aluminium(III) chloride to CCl₂FCClF₂ did not result in the incorporation of [³⁶Cl] into CCl₃CF₃, formed as a result of isomerization. It is possible that [³⁶Cl] was incorporated to a small extent into the minor, volatile components of the reaction mixture, CF₃CCl₂F and CCl₂FCClF₂, but this could not be established definitively.

Exposure of solid aluminium(III) chloride to [³⁶Cl]-CH₃CCl₃ at room temperature led to a rapid build-up of [³⁶Cl]-activity on the surface of the solid as the purple colour developed. The surface count rate was unaffected by removal of volatile products and fractionation of the latter indicated that H³⁶Cl was formed. Addition of inactive CH₃CCl₃ to the [³⁶Cl]-labelled purple solid had no effect on the surface count rate, although a small quantity of H³⁶Cl was detected in the vapour. The incorporation of [³⁶Cl] onto the surface when [³⁶Cl]-CH₃CCl₃ vapour was admitted to an unlabelled purple solid was minimal, indicating that the reaction depends on the availability of aluminium(III) chloride surface sites. Consistent with this view, no [³⁶Cl]-surface count rate and no colour change were observed when aluminium(III) chloride which had been pretreated with H₂O vapour was exposed to [³⁶Cl]-CH₃CCl₃. The reaction was not inhibited if H₂O vapour and [³⁶Cl]-CH₃CCl₃ were added concurrently. These observations, taken together, suggest that H₂O can block the adsorption of CH₃CCl₃ on aluminium(III) chloride but that it does not compete effectively with CH₃CCl₃, providing the latter is in large excess. The purple layer was hydrolytically unstable. Admission of H₂O vapour to the [³⁶Cl]-labelled purple solid resulted in a steady decrease in the surface count rate together with the colour changes, purple → brown → off-white. Some [³⁶Cl] was still retained however.

Admitting successive aliquots of CCl₂FCClF₂ to aluminium(III) chloride pretreated with [³⁶Cl]-CH₃CCl₃ as described above, resulted in behaviour which was similar to that observed for aluminium(III) chloride pretreated with [³⁶Cl]-CCl₂FCClF₂. Significant retention was normally observed and the major components in the vapour above the solid were CCl₂FCClF₂ and CCl₃CF₃, although the extent of conversion to the latter was generally less than observed with unlabelled CH₃CCl₃ (Table 2). [³⁶Cl]-Activity in the vapour was too small to quantify but admission of a CCl₂FCClF₂ aliquot did have an affect on the [³⁶Cl] surface count rate. An immediate change was observed when CCl₂FCClF₂ was admitted (increase or decrease observed in different experiments) but thereafter, the surface count rate remained constant over 24 h at its new value. The behaviour indicated that although there was no evidence for [³⁶Cl] incorporation during the isomerization, CCl₂FCClF₂ → CCl₃CF₃, changes to the nature of the organic material coating aluminium(III) chloride did occur but in an unpredictable fashion.

The behaviour of CCl₄-chlorinated γ-alumina

γ-Alumina chlorinated with CCl₄ is a solid Lewis acid which like aluminium(III) chloride is capable of catalysing the room temperature dehydrochlorination of CH₃CCl₃ and the oligomerization of CH₂[double bond, length half m-dash]CCl₂ so formed.¹² However mass balance and spectroscopic data obtained under identical conditions to those in Table 1 indicated that its activity for room temperature isomerization of CCl₂FCClF₂ was very low. Retention of organic material was observed in all reactions, up to 80% for extended exposure times. Samples of chlorinated γ-alumina that had been pretreated with CCl₂FCClF₂ showed no dehydrochlorination activity towards CH₃CCl₃, indicating that the Lewis acid sites required for dehydrochlorination had been blocked by the pretreatment.

Discussion

The elements of the model proposed to describe the dynamic behaviour of CCl₂FCClF₂ in the presence of solid aluminium(III) chloride at room temperature can be represented by the following:

(a) Adsorption (equation 1)

CCl₂FCClF₂ (g) → CCl₂FCClF₂ (ad) (1)

The experiments using [³⁶Cl]-labelled CCl₂FCClF₂ indicate that adsorption is rapid, that organic material is retained strongly at the surface and that desorption is slow. These conclusions are supported by mass balance data; analyses of product mixtures indicate that chlorination, isomerization and fluorination (albeit to a very small extent) occur.

(b) Chlorination (equations 2, 3 and 4)

CCl₂FCClF₂ (ad) + AlCl₃ (s) → CCl₃CClF₂ (ad) + ‘AlCl₂F’ (s) (2)

CCl₃CClF₂ + AlCl₃ (s) → CCl₃CCl₂F + ‘AlCl₂F’ (s) (3)

and

CCl₃CCl₂F + AlCl₃ (s) → C₂Cl₆ (s) + ‘AlCl₂F’ (s) (4)

Product analyses indicate that equation (2) represents an important reaction in the system. The reactions represented by equations (3) and (4), although less important, do occur as shown by the behaviour of solid CCl₃ClF₂ towards solid aluminium(III) chloride. The chlorination reactions lead to a build-up of organic material on the inorganic surface but they do not result in the rapid blocking of active sites for the isomerization of CCl₂FCClF₂.

Chlorination of C–F bonds by aluminium(III) chloride is well documented^9,10,15–17 although such reactions are usually performed above room temperature. The formation of an uncharacterized material, AlF_xCl_3−x, active for isomerization and dismutation of fluorohalocarbons, was reported¹⁵ during a re-examination of the reaction of refluxing CCl₂FCClF₂ with aluminium(III) chloride. More recently, the very active Lewis acid catalyst, amorphous AlF_2.8–2.9Cl_0.2–0.1, has been described.¹⁰ It is prepared from aluminium(III) chloride and CCl₃F in a reaction moderated by CCl₄. Calculated fluoride ion affinities of monomolecular aluminium(III) halides, AlCl_3−nF_n, n = 0–3, are high and from these it has been inferred that amorphous aluminium(III) chlorofluoride has surface sites having similar very high fluoride affinites.¹⁰ The material described in equations (2)–(4) as ‘AlCl₂F’, although not characterised, may be similar.

(c) Isomerization (equation 5)

CCl₂FCClF₂ (g) → CCl₂FCClF₂ (ad) → CCl₃CF₃ (g) (5)

It is proposed that isomerization occurs at an ‘AlCl₂F’ site, i.e. a surface Al^III in a disordered F/Cl environment, rather than at AlCl₃ which is therefore the catalyst precursor rather than the catalyst itself. Isomerization via a long-lived surface (or other) intermediate is inconsistent with the IR results and the absence of substantial [³⁶Cl] incorporation into CCl₃CF₃ in the presence of labelled aluminium(III) chloride is less easily visualized if an intermediate of the type, CClF₂CClFCl - - - AlCl₃, was involved. Therefore, it is reasonable to suggest that CCl₂FCClF₂ adsorbed at AlCl₃ results exclusively in chlorination while isomerization occurs exclusively at ‘AlCl₂F’ sites.

Consistent with this view is the inhibition of isomerization activity by trace H₂O, in agreement with an earlier observation,¹⁵ as similar behaviour has been reported for amorphous aluminium(III) chlorofluoride.¹⁰

(d) Fluorination

Although fluorination does occur, it is relatively unimportant and was not detected in the initial study.⁹ Conversion of a C–Cl bond to C–F by Al–F is thermodynamically unfavourable (unlike the analogous Cr^III/F/Cl system where the energetics are more balanced) and therefore extensive fluorination is not to be expected.

Conclusions

The most obvious way of modifying the catalytic properties of the archetypal solid Lewis acid, aluminium(III) chloride is by treatment with H₂O vapour even at a trace level. Experimentally this can be demonstrated by a change in the [³⁶Cl] exchange behaviour of H³⁶Cl toward the solid (Table 3) and can result in complete inhibition of activity towards the room temperature dehydrochlorination of CH₃CCl₃. It is also likely to be a factor in explaining the degree of irreproducibility observed in the behaviour of CCl₂FCClF₂ towards aluminium(III) chloride samples (Table 1). More subtle is the modification of the surface that results from chlorination of CCl₂FClF₂ by aluminium(III) chloride which, we have argued, is required for the catalytic isomerization of CCl₂FCClF₂ to CCl₃CF₃ to occur. In this situation aluminium(III) chloride is the catalyst precursor rather than the catalyst and we speculate that the active site for isomerization is a surface Al^III atom in a disordered chlorofluoride environment. In view of the calculated F⁻ ion affinities of molecular binary and mixed aluminium(III) halides,¹⁰ it is a moot point whether surface Lewis acidity is promoted by the fluorination process.

Somewhat surprisingly, the organic layer, comprising unsaturated oligomers derived from CH₂[double bond, length half m-dash]CCl₂ which coats aluminium(III) chloride as a result of its exposure to CH₃CCl₃, has little or no effect on the CCl₂FCClF₂ isomerization reaction. The active sites required are still accessible, presumably due to the quasi-liquid nature¹¹ of the surface layer and deactivation by adventitious H₂O is still possible.

We have previously used the dehydrochlorination behaviour of CH₃CCl₃ at an inorganic solid as an operational probe for the Lewis acidity of its surface.¹² Using this criterion is was expected that the γ-alumina chlorinated using CCl₄ would be active in catalysing the isomerization of CCl₂FCClF₂ at room temperature. This material would be an attractive alternative to aluminium(III) chloride particularly for large scale use, since it is hydrolytically less sensitive and has a greater surface area. However its isomerization activity is negligible although organic material, which blocks the sites at which CH₃CCl₃ dehydrochlorination occurs, is retained by the solid. The behaviour of CCl₂FCClF₂ in the presence of this material differs from that exhibited in the presence of aluminium(III) chloride, either bulk or sublimed onto calcined γ-alumina. A possible reason is that the chlorinating ability of chlorinated γ-alumina is inferior to that of aluminium(III) chloride, presumably for kinetic rather than for thermodynamic reasons. We conclude that the intrinsic Lewis acid strength of surface sites in heterogeneous catalyses involving halocarbon compounds is only one of the factors that must be considered.

Experimental

Standard vacuum and glove box techniques were used throughout. Mass balance experiments were carried out using an evacuable, thin Pyrex bulb equipped with a side arm and designed to enable aluminium(III) to be transferred directly by sublimation in vacuo. Solid transfers were also carried out in a glove box but such samples had lower activity for isomerization. [³⁶Cl]-Radiotracer experiments were performed in a Pyrex counting cell equipped with two intercalibrated Geiger Müller counters, a moveable Pyrex boat which could be positioned under either counter, an ampoule from which a solid could be dropped directly into the boat in vacuo and a gas handling system. The principles that underlie the technique and the application of the method for the study of [³⁶Cl]-halocarbons at inorganic surfaces have been described elsewhere.^13,18 Studies of the isomerization of CCl₂FCClF₂ to CCl₃CF₃ by IR spectroscopy were carried out in a Pyrex cell fitted with an evacuable ampoule from which solid was dropped into a depression at the bottom of the cell. Windows were either KBr or AgCl. Spectral changes were monitored over 24 h, the cell holder ensuring that positioning of the cell in the spectrometer was reproducible. Instrumentation was: IR, PE 983; ¹⁹F NMR, Bruker WP200 or AM200; GC, PE Sigma with FID and PE 8410 with hot-wire detection, 2 m packed column, calibrations being determined experimentally for all compounds; CGMS, VG 7070F.

Aluminium(III) chloride was sublimed before use (several times if required) as previously described.¹¹ Calcined γ-alumina (Degussa) was chlorinated with CCl₄ at 503 K for 6 h as described elsewhere.¹⁸ Samples of resublimed aluminium(III) chloride and chlorinated γ-alumina were treated with CH₃CCl₃ vapour at room temperature in the mass balance or radiochemical counting cells as appropriate. Both solids behaved as previously reported.^11,18

[³⁶Cl]-Labelled compounds were prepared as follows: CCl₂FCClF₂ by photolysis (medium pressure Hg lamp, silica vessel) of a 3∶1 mixture of Cl₂ and CH₂FCHF₂, CCl₄ by the thermal reaction of a 1∶1 mixture of Cl₂ and CHCl₃,¹⁹ CH₃CCl₃ by the iron(III) chloride-catalysed addition of HCl to CH₂[double bond, length half m-dash]CCl₂,¹⁸ HCl by the reaction of Cl⁻ with conc. H₂SO₄²⁰ and Cl₂ by KMnO₄ oxidation of HCl.²¹ Solid aluminium(III) chloride was labelled with [³⁶Cl] by exchange with CCl₄ at 60 °C for 24 h.²²

Acknowledgements

We thank the EPSRC and ICI for support of this work through CASE awards to D. G. McB. and M. M. McG.

References

1. See e.g., Chem. Br., Dec. 1998, p. 13; Chem. Eng. News, 1998, Sept. 21st, p. 41..
2. See e.g., ; S. Imamura, H. Shimizu, T. Haga, S. Tsuji, K. Utani and M. Watanabe, Ind. Eng. Chem. Res., 1993, 32, 3146; G. M. Bickle, T. Suzuki and Y. Mitarai, Appl. Catal. B, 1994, 4, 141; E. Kemnitz, A. Kohne and E. Lieske, J. Fluorine Chem., 1997, 81, 197; C. F. Ng, S. Shan and S. Y. Lai, Appl. Catal. B, 1998, 16, 209; S. Tashiro, H. Nagata, M. Kishida, K. Mizuno and K. Wakabayashi, Catal. Today, 1998, 45, 185.
3. See e.g., ; H. J. P. de Lijser, R. Louw and P. Mulder, J. Chem. Soc., Perkin Trans. 2, 1994, 135; J. Burdeniuc and R. H. Crabtree, Science, 1996, 271, 340; N. Sano, H. Tamon and M. Okazaki, Ind. Eng. Chem. Res., 1998, 37, 1428.
4. B. Coq, F. Figuéras, S. Hub and D. Tournigant, J. Phys. Chem., 1995, 99, 11159.
5. B. S. Ahn, S. C. Lee, D. J. Moon and B. G. Lee, J. Mol. Catal. A, 1996, 106, 83; E. J. A. X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum and J. A. Moulijn, Appl. Catal. A, 1997, 155, 59; A. Wiersma, E. J. A. X. van de Sandt, M. A. den Hollander, H. van Bekkum, M. Makkee and J. A. Moulijn, J. Catal., 1998, 177, 29; E. J. A. X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum and J. A. Moulijn, Appl. Catal. A, 1998, 173, 161; M. Makkee, E. J. A. X. van de Sandt, A. Wiersma and J. A. Moulijn, J. Mol. Catal. A, 1998, 134, 191; A. Malinowski, W. Juszczyk, M. Bonarowska, J. Pielaszek and Z. Karpinski, J. Catal., 1998, 177, 153; W. Juszczyk, A. Malinowski and Z. Karpinski, Appl. Catal. A, 1998, 166, 311.
6. G. Webb and J. M. Winfield, Chemistry of Waste Minimisation, ed. J. H. Clark, Blackie Academic and Professional, London, 1995, pp. 225–227..
7. L. Rowley, J. Thomson, G. Webb, J. M. Winfield and A. McCulloch, Appl. Catal. A, 1991, 79, 89; and references therein..
8. C. Gervasutti, L. Marangoni and W. Marra, J. Fluorine Chem., 1981, 19, 1; Z. Karpinski, K. Early and J. L. d’Itri, J. Catal., 1996, 164, 378; F. H. Ribeiro, C. A. Gerken, G. Rupprechter, G. A. Somorjai, C. S. Kellner, G. W. Coulston, L. E. Manzer and L. Abrams, J. Catal., 1998, 176, 352.
9. W. T. Miller Jr., E. W. Fager and P. H. Griswald, J. Am. Chem. Soc., 1950, 72, 705.
10. C. G. Krespan and D. A. Dixon, J. Fluorine Chem., 1996, 77, 117; V. A. Petrov, C. G. Krespan and B. F. Smart, J. Fluorine Chem., 1996, 77, 139.
11. D. G. McBeth, J. M. Winfield, B. W. Cook and N. Winterton, J. Chem. Soc., Dalton Trans., 1990, 671.
12. J. Thomson, G. Webb and J. M. Winfield, J. Mol. Catal., 1991, 68, 347; J. Thomson, G. Webb, J. M. Winfield, D. Bonniface, C. Shortman and N. Winterton, Appl. Catal. A, 1993, 97, 67.
13. J. U. Reid, S. J. Thomson and G. Webb, J. Catal., 1973, 29, 421; A. S. Al-Ammar and G. Webb, J. Chem. Soc., Faraday Trans. 1, 1978, 74, 195.
14. R. L. Richardson and S. W. Benson, J. Am. Chem. Soc., 1951, 73, 5096.
15. K. Okuhara, J. Org. Chem., 1978, 43, 2745.
16. N. Ishikawa, H. Iwakiri, K. Edamura and S. Kubota, Bull. Chem. Soc. Jpn., 1981, 54, 832.
17. K. C. Eapen, K. J. Eisentraut, M. T. Ryan and C. Tamborski, J. Fluorine Chem., 1986, 31, 405.
18. J. Thomson, G. Webb and J. M. Winfield, J. Mol. Catal., 1991, 67, 117.
19. H. A. Taylor and W. E. Hanson, J. Chem. Phys., 1939, 7, 48.
20. R. N. Maxson, Inorg. Synth., 1939, 1, 147.
21. G. Brauer, Handbook of Preparative Inorganic Chemistry, Academic Press, London, 2nd edn., 1965, vol. 1, p. 272..
22. C. H. Wallace and J. E. Willard, J. Am. Chem. Soc., 1950, 72, 5275.

---

Footnote

† Present address: The Leverhulme Centre for Innovative Catalysis, Dept. of Chemistry, The University of Liverpool, Liverpool L69 3BX, UK.

---