Dynamic behaviour of chlorofluoroethanes at fluorinated chromia aerogels and fluorinated zinc(II) or magnesium(II) doped chromia aerogels

Article abstract of DOI:10.1016/S0022-1139(02)00340-8

The preparation and characterisation of two series of fluorinated chromia aerogel materials, lightly doped with zinc(II) or magnesium(II), are described. They behave as heterogeneous catalysts for transformations of 1,1,2-trichlorotrifluoroethane under HF

Full text of DOI:10.1016/S0022-1139(02)00340-8

Journal of Fluorine Chemistry 121 (2003) 83–92
Dynamic behaviour of chlorofluoroethanes at fluorinated
chromia aerogels and fluorinated zinc(II) or
magnesium(II) doped chromia aerogels
Hamid Bozorgzadeha, Erhard Kemnitza, Mahmood Nickkho-Amiryb,
c,*
Tomaz Skapin , John M. Winfield
b
ˇ
aInstitute of Inorganic Chemistry, Humboldt University, Brook-Taylor Str. 2, D-12489 Berlin, Germany
bDepartment of Chemistry, University of Glasgow, Glasgow G12 8QQ, UK
ˇ
Department of Inorganic Chemistry and Technology, ‘‘Jozef Stefan’’ Institute,
c
Jamova 39, SI-1000 Ljubljana, Slovenia
Received 7 October 2002; received in revised form 27 November 2002; accepted 6 December 2002
Abstract
The preparation and characterisation of two series of fluorinated chromia aerogel materials, lightly doped with zinc(II) or magnesium(II),
are described. They behave as heterogeneous catalysts for transformations of 1,1,2-trichlorotrifluoroethane under HF-free conditions and at
moderate temperatures. Product distributions depend critically on the contact time. When the latter is very long (static conditions) the surface
becomes chlorinated, notwithstanding its fluorinated nature. Rather surprisingly, in view of previous work, the catalytic behaviour of the
materials is almost identical to that of the undoped fluorinated chromia aerogel. The nature of reactions occurring under HF-free conditions
and low levels of doping achievable are discussed as possible factors in determining catalytic behaviour.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Fluorinated chromia aerogel; Zinc(II); Magnesium(II); CFC; Catalysis
1. Introduction
formation of the symmetric product, CClF2CClF2. It was
concluded that on aluminium-based catalysts isomerisation
was the dominant reaction, followed by the dismutation of
the asymmetric product, CCl3CF3, to give CCl2FCF3 and
CCl3CClF2 [2]. A [36Cl] radiotracer study demonstrated that
in these two reactions surface Al–Cl groups are not formed,
indicating that the processes involved are intramolecular [4].
In contrast, the formation of labile chlorine- and fluorine-
containing species on the surface of chromium-based cata-
lysts that were in contact with CFCs had been established
earlier by using [36Cl]- or [19F]-labelled compounds [5].
From previous studies it is also known that labile halogen
species are involved in intermolecular halogenation reactions
[5–7] that, depending on specific conditions, can be described
as concerted dismutation processes or as non-concerted
halogen exchange. Comparative studies clearly show that
on chromium-based catalysts these halogenation reactions
prevail over isomerisation even under HF-free conditions
[1,2]. As a result, chromium-based catalysts exhibit lower
selectivity for the asymmetric products. Chromium-based
materials can be therefore considered inferior catalysts for
the conversion of CCl2FCClF2 to CCl2FCF3 under HF-free
Recently we investigated the possibility of converting
CCl2FCClF2 (CFC-113), which was in the past largely
used as a solvent, to the alternative refrigerant, CH2FCF3
(HFC-134a), in a process where anhydrous hydrogen fluor-
ide is not used. The optimal reaction scheme for the planned
conversion consists of three steps:
CCl2FCClF2 ! CCl3CF3 ! CCl2FCF3 ! CH2FCF3
Work carried out to date has demonstrated that conversion of
CCl2FCClF2 under HF-free conditions is possible on alumi-
nium- or chromium-based catalysts, but the catalytic beha-
viour of the two groups of catalysts differs significantly [1,2].
High conversion with high selectivity for the two target
asymmetric intermediates, CCl3CF3 and CCl2FCF3, was
observed only when fluorinated g-alumina or b-AlF3 were
used as catalysts [2,3]. On similar chromium-based catalysts,
the conversion was considerably lower, with the preferential
* Corresponding author. Tel.: þ386-1-477-3557; fax: þ386-1-423-2125.
E-mail address: tomaz.skapin@ijs.si (T. Skapin).
0022-1139/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-1139(02)00340-8

84
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
conditions, especially when compared to similar aluminium-
based catalysts [2].
model [13], originally developed for mixed oxides and
extended recently to fluoride systems [14].
Doping chromium-based materials with divalent cations,
such as magnesium(II) or zinc(II), has often been used to
improve the behaviour of catalysts in reactions involving
CFCs, HCFCs and HFCs [8,9]. For example, it was demon-
strated that the introduction of small quantities of zinc(II) on
chromia promotes significantly its catalytic ability in the
fluorination of CH2ClCF3 with HF to give CH2FCF3 [9].
This reaction is limited by the small equilibrium constant
and therefore requires highly active catalysts that are less
prone to coking. For the same reaction, introduction of MgF2
was claimed to promote activity, selectivity and durability of
CrF3-based catalysts [10]. However, doping AlF3Àx(OH)x
with magnesium(II) resulted in a progressive reduction of
Lewis acidity when the magnesium(II) content was
increased. A distinctive reduction in catalytic activity for
CCl2F2 dismutation was observed when the nominal Mg
content was greater that that required by the atom ratio
Mg:Al ¼ 1:9 [11]. In contrast, a synergic effect was observed
when CrF3Àx(OH)x was doped in a similar way. In this latter
case, an increase in Lewis acidity paralleled the increased
magnesium(II) content, the highest normalised catalytic
activity in the same test reaction was observed with the
atomic ratio Mg:Cr ¼ 1:3 [12]. These findings demonstrate
that significant modification of surface and catalytic proper-
ties can be achieved by doping. However, the real promo-
tional effect depends on many parameters, including, the
selected dopant–host pair, the dopant loading and dispersion,
and the nature of the reaction under investigation. Seldom can
it be anticipated. In some specific cases, the acidic properties
of fluorinated doped materials that may be relevant for
catalysis can be qualitatively predicted by using the Tanabe
In this paper we report the results of an investigation to
determine the effect of ZnII or MgII doping at low levels, on
the catalytic properties of fluorinated chromia aerogels with
respect to the behaviour of CCl2FCClF2.
2. Results and discussion
2.1. Pure and doped chromia aerogels
Modification of the general procedure to obtain pure
chromia aerogels by the chromium(VI) oxide/alcohol route
[15,16] allows the preparation of aerogels doped with alio-
valent cations. In this way two series of zinc(II) or magne-
sium(II) doped aerogels were prepared, denominated as Cr/
Zn or Cr/Mg, respectively. Some properties of the pure and
doped chromia aerogels used in this work are summarised in
Table 1. The main limitations to the procedure were imposed
by the solubility of the dopant compounds in the alcohol.
Two methods of dopant addition were therefore tested for
Cr/Zn aerogels (more details can be found in Section 4). For
the preparation of highly doped aerogels, the dopant com-
pound must be dissolved in bulk ethanol before an aqueous
solution of CrO3 is added (method B, Table 1). Using this
method, the highest value achievable for the atomic ratio
MII:Cr is ca. 26 ꢀ 10À3 and even so, loss of ZnII during the
latter stages of the preparation is significant (cf. later).
Higher doping levels could possibly be obtained by using
additional dopant deposition, for example by impregnation
as used in the earlier work [9], but this would be likely to
result in the destruction of the internal structures of the
Table 1
Some properties of zinc(II) or magnesium(II) doped chromia aerogels before and after fluorination
Sample (method)a
MII/Cr ratiob (ꢀ103)
BETc (m2 gÀ1
)
F-contentd (%)
Calculated
0
Foundc
S
VH
329
F
S
VH
F
Cr
–
–
–
482
271
12.7
Zinc
Cr/Zn-1 (A)
Cr/Zn-2 (A)
Cr/Zn-3 (B)
Cr/Zn-4 (A)
Cr/Zn-5 (B)
Cr/Zn-6 (B)
1.29
2.61
2.60
6.49
6.49
25.99
0.74
1.01
1.56
4.46
3.92
18.94
1.36
0.82
2.00
1.46
4.87
4.44
21.54
490
489
486
492
504
535
330
335
385
390
389
392
272
293
327
369
296
350
11.3
12.0
12.7
14.6
13.1
12.9
1.77
2.05
4.62
3.67
18.49
Magnesium
Cr/Mg-1 (B)
Cr/Mg-2 (B)
Cr/Mg-3 (B)
2.53
6.49
–
–
–
–
–
–
–
–
–
495
500
518
338
361
328
318
317
336
12.7
12.9
12.9
25.93
a Method of MII-acetate (M ¼ Zn or Mg) addition (see Section 4).
b MII/Cr atom ratio; determined only for the zinc(II) doped series.
c Stages of preparation: starting aerogel after autoclave treatment (S), vacuum heated (VH) and fluorinated (F).
d Total F-content in fluorinated samples.

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
85
aerogels. Pure chromia aerogel, denominated as Cr, was
used as a reference material.
the samples, indicating that part of the zinc is lost during
preparation. A more definitive comparison involving the
various preparation stages is not justified due to the sig-
nificant scattering in the data. However, the results clearly
show that the majority of zinc is lost during supercritical
drying and that the fluorination step had practically no effect
on the final zinc content. The preparation route used here
enabled the preparation of fluorinated Cr/Zn aerogels in
which a considerable portion of the original zinc is retained.
The scattering of results observed indicates that dopant
distribution throughout the batch of the aerogel may not
be uniform. A more precise determination of zinc distribu-
tion was not possible due to the poor mechanical character-
istics of the present samples.
Photoacoustic FTIR spectroscopy of adsorbed pyridine
(PAS-py) is a valuable tool to determine the nature of surface
acidic sites in different fluorinated materials based on alu-
mina or chromia [19]. PAS-py spectra of representative pure
and doped chromia aerogels are shown in Fig. 1. For all
aerogels, remarkably similar spectra were recorded that
indicate the presence of Lewis acid sites. Introduction of
Comparisons among the data in Table 1 show that neither
the identity of the dopant, its quantity, nor the method of
dopant introduction have any decisive effects on the proper-
ties determined at the dopant levels employed. Chromia
precursors prepared in the autoclave (preparation stage S)
are completely amorphous to X-rays and have extraordina-
rily high surface areas in the range of 500 m2 gÀ1. An
important feature of the autoclave-prepared doped chromias
is a large amount of organic residue, identified by photo-
acoustic FTIR spectroscopy (PAS) as acetate species with
characteristic bands at 1350, 1444 and 1547 cmÀ1. Evacua-
tion at 723 K (preparation stage VH) eliminates a consider-
able part of the organic residue with only partial reduction of
the surface area. Further reduction of the surface area is
observed after fluorination with CHF3 at 623 K (preparation
stage F). The extent of the fluorination is very similar for all
aerogels. Fluorinated aerogels remain amorphous and retain
relatively high surface areas. This indicates that the bulk
structure of the oxide aerogel precursors is largely conserved
throughout the preparation sequences. In this respect the
current materials resemble closely the behaviour of pure
chromia aerogels reported earlier [2,16] and differ signifi-
cantly from similarly treated alumina-based aerogels [17].
Bulk transformation to crystalline aluminium fluoride
occurs in the latter case, yielding materials with consider-
ably reduced surface areas.
Comparison within each series of doped aerogels, shows
that surface areas of highly doped materials are the highest,
as evidenced by the two highly doped samples, Cr/Zn-6 and
Cr/Mg-3. When conventional impregnation techniques are
used for the preparation of doped catalysts, the opposite
trend is usually observed due to the pore clogging and/or
filling processes induced by the dopant species. Such pro-
cesses, resulting in the loss of internal surface area, seem to
be less important when doped catalysts in the form of
aerogels are prepared. In the technique used here, both
constituents are homogeneously mixed in the liquid phase
from which the solid aerogel structure is formed during the
autoclave treatment under supercritical conditions. Even if
pore clogging and/or filling processes did occur, they are of
less importance in view of the very open structure of the
solid aerogel network. The use of acetates as the dopant
source has an additional beneficial effect. It was previously
found that ethanol-derived chromia aerogels, in which the
organic residue consisted mainly of thermally stable acetate
species, had relatively high mechanical stability and could
be prepared in large pieces or monoliths [18]. It is likely that
the additional acetate, present in the dopant precursors,
contributed further to consolidate the solid structure of
doped aerogels.
Fig. 1. Photoacoustic spectra of adsorbed pyridine (PAS-py) for CHF3-
treated pure Cr (A), doped Cr/Zn-6 (B), and Cr/Mg-3 (C) chromia
aerogels. Pyridine adsorption at 423 K; prior to pyridine adsorption the
samples were pre-treated in a flow of nitrogen at 423 K (upper bold trace)
or 873 K (bottom trace). L: bands of pyridine bonded to Lewis acid sites,
B: possible position of bands for pyridine bonded to Brønsted acid sites (if
present).
Chemical analyses of the complete zinc doped series were
performed to follow the behaviour of zinc in each prepara-
tion step (Table 1). The Zn/Cr atom ratios determined are
considerably lower than those calculated for the majority of

86
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
dopant species at the levels used here has therefore not
modified the Lewis acidity of the chromia surface. Very
similar surface characteristics are retained also after thermal
treatment up to 873 K, as shown by the corresponding lower
traces in Fig. 1. Disappearance of some minor bands in the
1500–1600 cmÀ1 region can be attributed to the elimination
of the last traces of organic residue retained after the
evacuation and fluorination steps. A slight shift of approxi-
mately 1 cmÀ1 to lower wavenumber could possibly be
interpreted as a slight reduction in the strength of Lewis
acid sites, probably as a consequence of a partial loss of
fluorine in the thermolysis reactions that take place at higher
temperatures (see later). The remarkable thermal stability
under non-oxidising conditions, previously reported for
fluorinated pure chromia aerogels [16], is confirmed also
in the present study for the doped materials.
extensively pre-fluorinated chromia, are likely to occur to
some extent, even at the intermediate temperature (623 K)
that is relevant for the catalytic work reported here. Under
these conditions, however, such hydrolytic modifications of
fluorinated surfaces seem to have a negligible effect on the
Lewis acidity of surface chromium sites because the inter-
mediate hydroxylation (Cr–F ! Cr–OH) is compensated for
by the dehydroxylation (2Cr–OH ! Cr–O–Cr þ H2O) a
process that creates new Lewis acid surface sites. Since these
processes involve just a small fraction of the surface, they do
not result in any measurable change in the overall acidity. As
ZnF2 is highly hygroscopic, these transformations will be
facile also at zinc(II) sites.
A change in Lewis acidity can therefore be detected using
PAS-py measurements, only from samples treated under
more drastic conditions (873 K). However, hydration/hydro-
xylation of the fluorinated chromia aerogels does not result
in the formation of Brønsted acidic sites that can be detected
by pyridine adsorption at 423 K (Fig. 1). Using the same
method, very low or negligible Brønsted acidity was pre-
viously found also for fluorinated conventional chromia
[19], iron(III) or magnesium(II) doped chromium hydroxy-
fluorides [12] and mixed MgF2/CrF3 catalysts [22]. Simi-
larly, a very small surface coverage of ammonia bound to
Brønsted acid sites was determined by using FTIR spectro-
scopy in an earlier study of reduced amorphous chromia
[23]. On basis of the Tanabe model [13,14], incorporation of
a metal(II) species in the Cr2O3 or CrF3 lattice should result
in the creation of Brønsted sites. However, since this model
does not allow any conclusion to be drawn regarding the
strength of the acid sites, it seems that they are very weak
and can not be detected by PAS-py. Such sites can be rather
regarded as weakly basic in nature, thus explaining their
behaviour in contact with H36Cl (mentioned later).
By using a radio-labelled probe molecule, H36Cl, it was
possible to determine the role of water on surface properties
of differently fluorinated aluminium-based materials [24]. A
similar approach has been used in the present study to
examine the behaviour of fluorinated chromia aerogels
towards H36Cl at room temperature. In Table 2 the results
for fluorinated pure and zinc(II) doped chromia aerogels are
compared with those obtained with some aluminium-based
materials. The most interesting feature is the extremely high
proportion of H36Cl that was retained on the surface of
fluorinated chromia aerogel and its zinc(II) doped derivative.
The fractions of [36Cl] retained are far greater than was the
case for b-AlF3 or fluorinated g-alumina (Table 2 and [24])
and are comparable with, or greater than, that retained by
calcined g-alumina under identical conditions. It has been
shown previously [25] that the latter material is chlorinated
by HCl under these conditions, with surface hydroxyl groups
being replaced by Al–Cl bonds. It appears that the situation is
similar for the chromia aerogels and, despite the fact that their
surfaces have been extensively pre-fluorinated, they contain
sufficient Cr–OH groups that are basic enough to interact
with HCl, for example to replace Cr–OH by Cr–Cl bonds.
PAS spectra (not presented) show broad features in the
950, 1630 and 3400 cmÀ1 regions, indicating that some
hydration/hydroxylation of fluorinated chromia aerogels
always occurs, even if the samples are in contact with
ambient air for very short periods of time during manipula-
tion. Dissociative water adsorption leading to the formation
of surface hydroxyl groups, and non-dissociative water
adsorption, evidenced by the presence of molecular water,
are both well documented for crystalline and amorphous
chromia [20,21]. Similar behaviour towards water is appar-
ently exhibited also by fluorinated chromia aerogels. It is
reasonable to assume that water adsorption on these materi-
als is additionally promoted by their very high surface area
and due to the increased Lewis acidity induced by fluorina-
tion. In temperature programmed desorption (TPD) inves-
tigations, performed as an extension of previous ammonia
TPD studies [16], practically continuous desorption of water
from such aerogels, preconditioned at 373 K, was detected
by FTIR over the whole range, 373–873 K, investigated. A
distinctive maximum in water desorption at approximately
433 K was always observed. This peak indicated the deso-
rption of physisorbed or weakly bonded water. Desorption at
higher temperatures could be associated with the condensa-
tion of surface hydroxyl groups. Desorption of water from
fluorinated chromia aerogels resembles therefore that pre-
viously observed for amorphous chromia [21]. Desorption of
HF commenced at 493–533 K, depending on the pre-treat-
ment procedure used. As expected, the onset of HF deso-
rption from evacuated samples was always shifted to higher
temperatures. Corresponding initial losses of fluorine on
heating up to 873 K ranged from 28 to 35%. In addition, loss
of fluorine was almost complete when fluorinated aerogels
were heated under ambient air atmosphere at 723 K. It was
concluded that HF, detected in TPD experiments, originated
from desorption of physisorbed HF and from hydrolysis of
metal fluoride species affecting the fluoride phase, the latter
process becoming increasingly important at higher tempera-
tures. It appears that when contact with water cannot be
completely excluded, formation of surface hydroxyls, Cr–
OH, and a partial replacement of Cr–F bonds by Cr–O in

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
87
Table 2
of CCl2FCClF2 were studied under dynamic flow condi-
tions. Results are collected in Table 3; distributions of the
two major products for both series of doped catalysts are
compared in Fig. 2. Comparisons among the results show
that there were no major differences between the catalytic
behaviour of pure and doped chromia aerogels. Similarly,
there is apparently no effect arising from the extent of dopant
loading. Conversion of CCl2FCClF2 is in the range 50–59%;
the two major products are CClF2CClF2 and CCl3CClF2 and
the extent of isomerisation to CCl3CF3 is very low. Halogen
balances for the identified volatile products always show a
partial loss of chlorine. In general, the behaviour of the
present catalysts resembles very much that exhibited by
chromia aerogels treated in a similar way and used in our
previous study [2]. Formation of the two main products,
CClF2CClF2 and CCl3CClF2, is consistent with the pre-
viously proposed intermolecular models [2,5,6], in which
labile halogen species, denoted as ‘‘F’’ or ‘‘Cl’’, on the
surface of chromia-based catalysts are involved (Eqs. (1) and
(2)). Depending on whether these reactions proceed in a
concerted or non-concerted manner, they can be described
either as dismutation or as halogen exchange. Based on the
present data, a clear distinction between both options is not
possible:
Surface [36Cl] count rates determined from g-alumina, CHF3-fluorinated
chromia aerogel and CHF3-fluorinated, zinc(II) doped chromia aerogel in
the presence of H36Cl at room temperature and after its removal
Surface [36Cl] count ratea (count minÀ1
)
Activity of [36Cl]
retainedb (%)
In the presence
of H36Cl
After removal
of H36Clc
Fluorinated Cr aerogel
5892
5868d
4994
3623
79
62
Fluorinated Cr/Zn-2 aerogel
4954
5413
4821
4398
97
81
g-Alumina
4510
2956
3153
2943
66
71
65
4440
4519
b-AlF3 or fluorinated g-aluminae
7–11
a Saturation values.
b Estimated relative error ꢁ10%.
c Removed under static vacuum.
d In contact for 1 h before removal of H36Cl.
e Range of values reported for fluorinated aluminium-based materials
taken from [24].
CCl2FCClF2 þ ‘‘F’’ ! CClF2CClF2 þ ‘‘Cl’’
CCl2FCClF2 þ ‘‘Cl’’ ! CCl3CClF2 þ ‘‘F’’
(1)
(2)
Based on PAS-py results discussed earlier, it is reasonable to
conclude that non-acidic hydroxyl groups, capable of inter-
acting with H36Cl at room temperature, may originate from
processes that follow the surface rehydration of fluorinated
chromia aerogels.
Formation of other minor products can be explained by
similar fluorination or chlorination sequences. Although
competitive catalytic reactions, such as isomerisation, prob-
ably also occur, they are of lesser importance under the
conditions used in flow experiments. The final result of all
these transformations is the formation of chlorine- and
fluorine-rich by-products. The formation of the highly fluori-
nated product, CClF2CF3, which may be formed by fluorina-
tion of either CCl2FCF3 or CClF2CClF2 (Eqs. (3) and (4)), is
2.2. Catalytic behaviour
2.2.1. Steady flow conditions
The influences of the dopants at different concentrations
on the catalytic activity and selectivity in HF-free conversion
Table 3
Reaction of CCl2FCClF2 (CFC-113) under steady flow conditions at 623 K and contact time 1 s
Catalyst
CFCa products (mol%)
F:Cl ratio
3.03:2.91
1110
1.00
110
111
112
112a
113
113a
0.69
114 þ 114ab
115
Cr
0.22
1.32
2.00
17.11
47.11
29.67
0.21
Cr/Zn-1
Cr/Zn-2
Cr/Zn-3
Cr/Zn-5
Cr/Zn-6
1.32
3.11
0.90
2.09
0.84
0.19
0.15
0.11
0.17
0.12
1.51
1.38
1.13
1.59
1.16
1.89
1.93
1.73
2.02
1.74
19.47
16.68
18.70
19.08
18.75
43.35
41.26
50.01
41.84
49.36
1.80
2.14
0.68
2.15
1.14
30.24
33.06
26.60
30.79
26.74
0.24
0.30
0.14
0.26
0.14
3.02:2.96
3.03:2.91
3.01:2.97
3.00:2.97
3.01:2.95
Cr/Mg-1
Cr/Mg-2
Cr/Mg-3
2.09
2.05
2.16
0.14
0.09
0.12
1.47
1.11
1.51
2.14
1.81
2.06
18.37
17.31
18.42
43.31
46.46
44.07
1.82
1.32
1.72
30.44
29.62
29.73
0.22
0.23
0.21
3.01:2.95
3.02:2.94
3.00:2.96
a Key for compounds: 1110, CCl2¼CCl2; 110, CCl3CCl3; 111, CCl3CCl2F; 112, CCl2FCCl2F; 112a, CCl3CClF2; 113, CCl2FCClF2; 113a, CCl3CF3; 114,
CClF2CClF2; 114a, CCl2FCF3; 115, CClF2CF3.
b The two isomers were not separated by GC, the sums of both are given. GC/MS and 19F NMR indicate that 114 prevails, typical molar ratios were 114/
114a ꢂ 6:25.

88
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
Fig. 2. Distribution of the main products, CCl2FCF3 (CFC-114a), CClF2CClF2 (CFC-114) and CCl3CClF2 (CFC-112a), and unconverted CCl2FCClF2 (CFC-
113) obtained under steady flow conditions at 623 K and contact time of 1 s for the fluorinated zinc(II) (Cr/Zn) and magnesium(II) (Cr/Mg) doped chromia
aerogel series in comparison with the fluorinated undoped chromia aerogel (Cr).
evidently limited under HF-free steady flow conditions used
here:
2.2.2. Behaviour of surface halogen species
A series of catalytic tests under static conditions was
performed for representative pure and zinc(II) doped chro-
mia aerogel catalysts in order to obtain supplementary
information on the course of the HF-free conversions men-
tioned before. In contrast to the flow experiments, the results
obtained under static conditions, involving much longer
contact times, represent a closer approach to the equilibrium
situation where the contributions of different competitive
reactions may become significant or predominant. The
behaviour of CCl2FCClF2 and CCl3CF3 under static condi-
tions is summarised in Table 4. In all cases a distinctive
deficiency in the chlorine balance is observed. This is most
likely to be the result of the formation of completely
chlorinated products, CCl3CCl3 and CCl2¼CCl2, which
were not detected by the 19F NMR analytical method used.
In addition, chemical analyses of used catalysts, discussed
later, indicated that deposition of chlorine, both organic and
inorganic, was substantial. The data given in Table 4 are
representative therefore for the volatile fraction in which
highly fluorinated products prevail and do not accurately
CCl2FCF3 þ ‘‘F’’ ! CClF2CF3 þ ‘‘Cl’’
CClF2CClF2 þ ‘‘F’’ ! CClF2CF3 þ ‘‘Cl’’
(3)
(4)
It appears that under these conditions the reactions leading to
chlorine-rich products complete more easily, presumably
due to kinetic effects. The observed trends reflect well the
large differences in the reactivity of different members of the
C2Cl6ÀnFn series. It is well known that the reactivity of these
molecules increases rapidly with increasing number of
chlorine atoms. As a consequence, completely chlorinated
products are readily formed even with the relatively short
contact times used in flow experiments. The unsaturated
compound, CCl2¼CCl2, is formed in the final step by the
dechlorination of CCl3CCl3, according to Eq. (5). Formation
of CCl2¼CCl2 with the liberation of chlorine that was not
detectable by GC, accounts for the deficit in chlorine
balances observed in all catalytic tests (Table 3):
CCl3CCl3 ! CCl2¼CCl2 þ Cl2
(5)
Table 4
Reaction of CCl2FCClF2 (CFC-113) or CCl3CF3 (CFC-113a) with CHF3-fluorinated chromia aerogel or zinc(II) doped chromia aerogel at 623 K under static
conditions
Catalysta
Test no.a CFCc reactant
Duration (h)
CFCc products in the volatile fraction (mol%)b
F:Cl ratio
111
112
112a
4.67
113
113a
114
114a
115
4.97
Cr
I
113
24
0.41
4.33
7.73
14.39
45.10
18.40
3.64:2.36
Cr/Zn-2
II
113
113
113
113a
24
72
72
72
4.85
7.12
2.70
6.40
8.10
3.23
0.60
6.65
8.60
7.34
14.1
4.12
8.27
1.07
1.60
1.58
32.21
21.37
23.60
46.60
13.10
6.04
5.40
1.30
14.67
21.70
17.00
16.40
10.20
32.13
35.00
16.95
3.22:2.68
3.67:2.33
3.72:2.28
3.28:2.72
III
V
IV
a Numbers correspond to tests given in Table 5.
b Determined by 19F NMR spectroscopy from the combination of volatile and less volatile fractions, only fluorinated products were detected.
c For CFC key, see footnote in Table 3.

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
89
reflect their true equilibrium concentrations. As with the
flow experiments, the behaviour of the zinc(II) doped chro-
mia towards CCl2FCClF2 is very similar to that of the
analogous undoped material. With very long contact times,
used in static experiments, high conversions of CCl2FCClF2
are observed with the formation of significant quantities of
the highly fluorinated product, CClF2CF3. This compound is
very likely to be formed by the fluorination steps given in
Eqs. (3) and (4). There is a direct correlation between the
contact time and the amount of CClF2CF3 formed. Under the
HF-free conditions used here these fluorination steps seem to
be very slow and have high activation energies. Accordingly,
very low conversion to CClF2CF3 is observed with the much
shorter contact times used in flow experiments.
Catalytic tests in which the asymmetric isomer, CCl3CF3,
was used as the starting substance, indicates that even under
the extreme conditions used in static experiments, the
reactivity of this compound is much lower than that observed
for the more symmetric isomer, CCl2FCClF2. This finding
correlates well with the comparative investigations per-
formed under flow conditions on conventional chromia
catalysts, pre-treated with CCl2F2 and used in a previous
study [2]. Under flow conditions, conversion of CCl2FCClF2
was several times higher than that observed for CCl3CF3.
The main highly fluorinated product was CClF2CClF2 when
starting with CCl2FCClF2, while the asymmetric CCl2FCF3
prevailed in the case of CCl3CF3. The difference reflects the
kinetic and thermodynamic stability of the CF3– group.
Different halogen-containing species, present on the sur-
face of fluorinated chromia, are recognised to be the key
factors determining the catalytic behaviour of these materi-
als [5,7,9]. Catalytic work performed within this study gives
additional evidence in this respect. To obtain a deeper insight
in the behaviour of an individual halogen, chemical analyses
of some pure or zinc(II) doped chromia aerogels before
and after catalytic tests under static conditions were per-
formed. Results are summarised in Table 5. Comparison
of initial and final fluorine contents demonstrated that the
level of fluorine remained practically constant, within the
experimental error, for all chromia aerogels investigated.
Deviations from this general behaviour, most evident in the
case of catalytic test III (Table 5) could be ascribed primarily
to the possible differences between aliquots of the same
catalyst used for each test. As already mentioned, scattering
of results from the determination of the metals ratio suggests
that the homogeneity of catalysts might be low, even if they
originated from the same batch of aerogel. Nevertheless, it
can be concluded that fluorine, initially present in the pre-
fluorinated chromia, has not contributed significantly to the
fluorine balance among the organic products. Bulk fluorine
is evidently not involved in catalytic halogenation reactions
that are believed to be the main processes taking place under
the HF-free conditions investigated here. This conclusion is
supported by a previous investigation [7] where, using [18F]
as a radiotracer, three types of fluorine-containing species
were identified in fluorinated chromia catalysts. Only the
labile fluorine species, representing a minor fraction of the
total fluorine present, was catalytically active.
Incorporation of chlorine in used catalysts was remark-
ably high, irrespective of the type of catalyst or chlorofluor-
oethane used as the starting substance. Longer contact times
had apparently no effect on the levels of chlorine incorpo-
rated, indicating that its deposition is relatively fast. In
previous investigations [2] deposition of organic residue
was always observed after prolonged catalytic tests under
HF-free static conditions. Identification of CCl3CCl3 and
CCl2¼CCl2 suggested that the organic residue consisted
mainly of fully chlorinated and less volatile compounds. In
order to differentiate among possible forms of chlorine
present in used catalysts, solvent extractions on the solids
were performed. Two different solvents, chloroform and
diethyl ether, were tested and both gave comparable results
(Table 5). Slightly better extraction results were usually
obtained with chloroform. This may be correlated with its
better solvent properties for chlorinated hydrocarbons. It
should be also pointed out that the use of the highly
Table 5
Chemical analyses of fluorinated chromia aerogels used in HF-free catalytic reactions under static conditions at 623 K
Catalyst
Test no.
CFC reactant
Duration (h)
Initial FÀ (%)a
After catalytic test FÀ (%)a
Extracted
ClÀ (%)b
Extracted
CHCl3
7.0a
CHCl3
11.9
Ether
Ether
8.7
Cr
I
113
20
12.2
11.8
12.2
13.0
Cr/Zn-2
II
113
113
113a
113
113
24
72
72
72
72
11.5
11.5
11.5
11.5
12.0
9.8
18.4
8.2
8.9
18.9
7.9
10.8
17.1
14.7
12.6
8.0
6.2a
10.7a
9.5a
7.2
III
IV
V
11.0
13.7
10.7
13.3
9.8
7.5
VI
6.1
Cr/Zn-3
VII
113
72
12.7
16.1
16.9
12.0
7.6
a Sample decomposition in alkaline melt.
b Sample pyrohydrolysis.

90
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
chlorinated solvent, chloroform, has, evidently, not inter-
fered with chlorine determinations. In the analytical data
obtained from catalysts after extraction, a considerable
reduction in chlorine content was observed while there were
no distinctive changes in fluorine values. The latter findings
confirmed that the fluorine content of the organic residue
was negligible. The presence of extractable chlorine in used
catalysts that had been treated in vacuo for 2 h at 423 K
provided further evidence that the deposit consisted of high
boiling fully chlorinated products. Chlorine found in
extracted catalysts was still significant and indicated the
part of the total chlorine that is strongly retained by chromia,
very probably bound directly to Cr. Current investigations
gave clear evidence that substantial incorporation of
chlorine is possible, even in cases when chromia has
been extensively pre-fluorinated. These findings comple-
ment therefore very well the work with H36Cl, described
earlier.
nationandchlorination cycles exemplified by Eqs. (1) and (2).
Present catalytic tests performed under static conditions show
that labile fluorine species gradually form highly fluorinated
compounds. Although these processes are slow, the com-
pounds formed are stable and highly volatile. Incorporation of
fluorine in these compounds can be therefore regarded as a
sink for the active fluorine species. On the other hand, tests
performed under static and flow conditions clearly demon-
strate that the labile chlorine species easily form less stable
andlessvolatilecompoundsthat aredepositedonchromiaand
may, finally, decompose to yield Cl2. The partially fluorinated
chromia aerogels used in static experiments are therefore
subjected to strong chlorination conditions. As a result,
additional halogenation takes place, converting some of the
oxide, unconverted in the pre-fluorination step, to chloride.
Formation of significant quantities of CCl2¼CCl2 in flow
experiments suggests that additional chlorination of pre-
fluorinated chromia aerogels is likely to occur also under
these conditions. This aspect may become relevant in situa-
tions when the time on-stream is long.
Based on the results discussed so far, the following
rationalisation for the behaviour of halogens on fluorinated
chromia aerogel surfaces can be made. First it must be pointed
outthatthebehaviourofdoped andundoped chromia aerogels
observed in the present study resembles very much that of
pure chromia aerogels reported here and earlier [2]. Introduc-
tion of dopants at the levels used in the present study has
therefore a negligible effect on the catalytic processes under
discussion. Fluorination of all chromia aerogels yielded
materials with remarkably similar characteristics. Only par-
tial fluorination is indicated in all cases. It is assumed that
fluorination of oxide aerogel precursors proceeds predomi-
nantly on their surfaces, by the gradual replacement of surface
Cr–OH groups and Cr–O–Cr clusters with Cr–F bonds. The
interior of the chromia particles consists mainly of uncon-
verted oxide. Asdemonstrated, the highlyfluorinatedchromia
surfaces formed in this way, are very susceptible to water
adsorption and, consequently, easily loose some fluorine.
Depending on the specific conditions, new Cr–O–Cr clusters
or Cr–OH groups in a distorted fluoride environment are
formed. These species represent the active sites that may
interact with reactive halogen species. High retention of
H36Cl after adsorption at room temperature is clear evidence
for such strong interaction. In catalytic reactions under
HF-free conditions, the reactive halogen species originate
from CCl2FCClF2 and its products in reactions similar to
those given in Eqs. (1) and (2). Due to the small differences in
bond energies, the formation of both Cr–F and Cr–Cl bonds is
thermodynamically possible [2]. This was clearly demon-
strated by photoelectron and Auger electron spectroscopy in
the case of conventional chromia where both halogens were
detected on the surface after treatment with different HCFCs
or CFCs [26]. The situation encountered on fluorinated chro-
mia differs therefore distinctively from that observed for
fluorinated alumina materials where the formation of Al–F
over Al–Cl bonds is strongly favoured [2,4]. According to
previous studies [5,7], some of the halogen species formed on
chromia may be labile and, as such, can be involved in fluori-
2.2.3. Role of the dopant
Catalytic behaviour of zinc(II) or magnesium(II) doped
chromia aerogels in contact with CCl2FCClF2, under flow or
static conditions, was unexpectedly similar to the corre-
sponding undoped materials. This situation parallels well the
similarity in other properties determined for these materials.
It appears that the introduction of an aliovalent cation at
levels used here has no effect on catalytic reactions studied.
Dopants do not promote the lability of surface halides
involved in intermolecular halogenations that are believed
to be the dominant reaction mechanisms when CCl2FCClF2
is contacted with chromia catalysts under HF-free condi-
tions. Isomerisation, the most plausible competitive reac-
tion, especially when long contact times are used, is also not
influenced by doping. Lack of any observable effect induced
by doping can be explained both by the nature of catalytic
reactions investigated here and by the method used for
doping. Earlier it was proposed [9] that the promotional
role of zinc(II) in the catalytic fluorination of CH2ClCF3 by
HF that leads to CH2FCF3, is due to the formation of ZnF2
‘islands’ on the surface and their consequent effect on the
adsorbed state of HF. The current findings are not incon-
sistent with this assumption, since the reactions studied here
do not involve HF as a reagent. The nature of the adsorbed
state of CCl2FCClF2 and its subsequent reactions with sur-
face labile halide species (cf. earlier) are the important
factors here. Adsorption of CCl2FCClF2 is likely, following
earlier work [5], to occur at coordinatively unsaturated CrIII
sites that have a disordered O/F environment.
The degree of inhomogeneity in the distribution of dopant
observed for some samples and the intrinsic surface levels
of the dopant are also relevant. In the study referred to [9],
the best results were obtained with a low level of ZnII but
this was incorporated by aqueous impregnation of conven-
tionally prepared chromias. Using this method, ZnII may

H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
91
react with surface/subsurface layers only, thus the dopant
concentration on the surface could be higher than in the
present work, where both bulk and surface are doped.
For chromia-based catalysts doped with MgII obtained by
co-precipitation [12], the best catalytic activities have been
obtained using samples having significantly higher levels of
the dopant (ca. 25%) than was possible here, as discussed
earlier (Section 2.1). From the present work it appears that if
high levels of doping are required, these are not compatible
with the use of aerogel materials.
quantity of organic contaminants. This procedure preserved
substantially the morphology of the precursor. Fluorination
of conditioned oxidic aerogels (4–5 g) was performed under
flow conditions at 623 K with trifluoromethane, CHF3
(Matheson, minimum 98%; 20 g hÀ1) for 4 h. Other details
for the preparation of chromia aerogels starting from CrO3
and alcohols are given elsewhere [15,16,18].
4.2. Catalyst characterisation
All solids were examined by X-ray powder diffraction.
Specific surface areas were determined by the BET method
using nitrogen as the adsorbate. Temperature programmed
desorption and infrared photoacoustic spectroscopy of che-
misorbed pyridine (PAS-py) were performed according to
the procedures reported earlier [19]. The framework vibra-
tions of the solids were used as an internal reference to
obtain the normalised PAS-py spectra. When transferring the
sample to the photoacoustic cell, a short contact with
ambient air could not be prevented, due to the technical
characteristics of the cell used.
Chromium and zinc contents for the zinc(II) doped series
were determined by redox titration or inductively coupled
plasma atomic emission spectroscopy (ICP-AES), respec-
tively, after decomposition in carbonate (alkaline) melts.
Total fluorine and chlorine contents were determined with
the respective ion-selective electrodes following the total
decomposition in carbonate (alkaline) melts or by pyrohy-
drolysis. Separate tests with model materials shoved that
both decomposition methods gave comparable results.
Absolute differences in fluorine or chlorine values obtained
by both methods were usually lower than 1%. Samples of
chromia aerogels used in some experiments under static
conditions were extracted with chloroform or ether. Before
extraction, used catalysts were evacuated at 423 K for 2 h.
For each extraction two aliquots of fresh solvent were used.
In each extraction the solid sample was shaken with the
solvent for 30 min then solvent was removed by filtration.
Extracted solids were dried at 322 K and cooled in a
desiccator.
The behaviour of representative chromia aerogels and
g-alumina towards [36Cl]-labelled anhydrous hydrogen
chloride at room temperature was examined using a
Geiger–Mu¨ller direct monitoring method, originally devel-
oped for [14C] adsorption measurements [27]. Two, inter-
calibrated Geiger–Mu¨ller counters were positioned within
the evacuable Pyrex counting vessel to enable [36Cl] activ-
ity from the vapour phase and from the vapour plus surface
to be monitored concurrently. The fraction retained on the
surface was obtained from the difference of the two values.
It should be noted that due to the strong self-absorption
of the bÀ emitter [36Cl], only the activity from the surface
can be monitored in this way. Incorporation of H36Cl
into the bulk solid is therefore not detectable by this
technique. Further details can be found in previous reports
[4,24].
3. Conclusions
Doping of chromia aerogels with low levels of either
ZnII or MgII with subsequent fluorination leads to stable
materials that have catalytic activity for transformations of
CCl2FCClF2. However, the promotional effect of MII,
observed in a previous study [9], was not observed. This
is accounted for by a combination of two factors. The level
of doping achievable from solution impregnation of aerogel
materials is limited. The catalyst–substrate interactions,
required for catalytic fluorination using anhydrous HF (the
reaction studied in [9]) and those appropriate for dynamic
behaviour of CCl2FCClF2 under HF-free conditions, studied
here, are apparently different.
4. Experimental
4.1. Preparation of catalytic materials
The preparation procedure for chromia aerogels consisted
of three separate steps, preparation of oxidic precursors
(stage S), conditioning (stage VH) and fluorination (stage
F). Precursor aerogels were prepared from an aqueous
solution of chromium(VI) oxide (Merck, minimum 99%)
and a large excess of ethanol (Carlo Erba, minimum 99.8%).
The liquid mixtures thus obtained were dried in the auto-
clave at supercritical conditions (573 K). Two series of
chromia aerogels doped with magnesium(II) or zinc(II)
cations were prepared. For both series hydrated metal
acetates (Merck, minimum 99.5%) were the dopant sources.
Due to the limited solubility of metal(II) acetates in ethanol,
two methods of dopant addition were tested for the zinc(II)
doped series. Method A: CrO3(aq.) (81 g, 43.2 wt.% CrO3)
was mixed with the majority of ethanol (280 cm3), then a
corresponding quantity of zinc(II) acetate, dissolved in
100 cm3 of ethanol, was added. Method B: corresponding
quantity of zinc(II) acetate was dissolved in 380 cm3 of
ethanol, then CrO3(aq.) (same quantity as in A) was added to
the solution. For higher dopant loadings only method B
could be used. For the magnesium(II) doped series only
method B was used.
Before fluorination, precursor chromia aerogels were con-
ditioned at 723 K under dynamic vacuum to reduce the

92
H. Bozorgzadeh et al. / Journal of Fluorine Chemistry 121 (2003) 83–92
[7] J. Kijowski, G. Webb, J.M. Winfield, Appl. Catal. 27 (1986) 181–
4.3. Catalytic reactions
193.
[8] L.E. Manzer, M.J. Nappa, Appl. Catal. A: Gen. 221 (2001) 267–274.
[9] D.W. Bonniface, J.R. Fryer, P. Landon, J.D. Scott, W.D.S. Scott, M.J.
Watson, G. Webb, J.M. Winfield, Green Chem. 1 (1999) 9–11.
[10] H. Kim, H.S. Kim, B.G. Lee, H. Lee, S. Kim, J. Chem. Soc., Chem.
Commun. (1995) 2383–2384.
Catalytic reactions of CCl2FCClF2 were studied under
steady flow and static conditions, usually at 623 K. Experi-
mental methods for both sets of experiments and analytical
methods for the quantification of reaction products were the
same as those used in the previous studies [2,4]. The contact
time for experiments performed under steady flow condi-
tions was 1 s, for static experiments the contact times ranged
from 24 to 72 h.
[11] E. Kemnitz, A. Hess, G. Rother, S. Troyanov, J. Catal. 159 (1996)
332–339.
[12] B. Adamczyk, A. Hess, E. Kemnitz, J. Mater. Chem. 6 (1996) 1731–
1735.
[13] K. Tanabe, T. Sumijoshi, K. Shibata, T. Kiyoura, J. Kitagawa, Bull.
Chem. Soc. Jpn. 47 (1974) 1064–1066.
[14] E. Kemnitz, Y. Zhu, B. Adamczyk, J. Fluorine Chem. 114 (2002)
163–170.
Acknowledgements
[15] J.N. Armor, E.J. Carlson, W.C. Conner, React. Solids 3 (1987) 155–
159.
We thank the European Union for support of this work
under contract ENV4-CT97-0601. Additional funding for T.
Skapin was provided by the Ministry of Education, Science
and Sport of the Republic of Slovenia.
[16] T. Skapin, E. Kemnitz, J. Non-Cryst. Solids 225 (1998) 163–167.
[17] T. Skapin, E. Kemnitz, Catal. Lett. 40 (1996) 241–247.
[18] T. Skapin, J. Non-Cryst. Solids 285 (2001) 128–134.
[19] A. Hess, E. Kemnitz, J. Catal. 149 (1994) 449–457.
[20] A. Zecchina, S. Coluccia, E. Guglielminotti, G. Ghiotti, J. Phys.
Chem. 75 (1971) 2774–2783.
[21] M. Schraml-Marth, A. Wokaun, H.E. Curry-Hyde, A. Baiker, J.
Catal. 133 (1992) 415–430.
References
[22] H. Lee, H.D. Jeong, Y.S. Chung, H.G. Lee, M.J. Chung, S. Kim, H.S.
Kim, J. Catal. 169 (1997) 307–316.
[1] M. Blanchard, L. Wendlinger, P. Canesson, Appl. Catal. 59 (1990)
123–128.
[23] M. Schraml-Marth, A. Wokaun, H.E. Curry-Hyde, A. Baiker, J.
Catal. 133 (1992) 431–444.
[2] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M.
Winfield, J. Fluorine Chem. 110 (2001) 181–189.
[3] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M.
Winfield, J. Fluorine Chem. 107 (2001) 45–52.
[24] C.H. Barclay, H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry,
D.E.M. Ross, T. Skapin, J. Thomson, G. Webb, J.M. Winfield, J.
Chem. Soc., Dalton Trans. (2002) 40–47.
[25] J. Thomson, G. Webb, J.M. Winfield, J. Mol. Catal. 67 (1991) 117–
126.
[4] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, G.D.
Tate, J.M. Winfield, J. Fluorine Chem. 112 (2001) 225–232.
[5] L. Rowley, J. Thomson, G. Webb, J.M. Winfield, A. McCulloch,
Appl. Catal. A: Gen. 79 (1991) 89–103.
[26] E. Kemnitz, A. Kohne, I. Grohmann, A. Lippitz, W.E.S. Unger, J.
Catal. 159 (1996) 270–279.
[27] A.S. Al-Ammar, G. Webb, J. Chem. Soc., Faraday Trans. I 74 (1978)
195–205.
[6] A. Farrokhnia, B. Sakakini, K.C. Waugh, J. Catal. 174 (1998)
219–230.