Highly selective metal fluoride catalysts for the dehydrohalogenation of 3-chloro-1,1,1,3-tetrafluorobutane

Article abstract of DOI:10.1016/j.jcat.2011.06.013

For the first time, dehydrochlorination and dehydrofluorination reactions are studied on the same substrate, 3-chloro-1,1,1,3-tetrafluorobutane, employing nanoscopic metal fluorides AlF₃, MgF₂, CaF₂, SrF₂, and BaF₂ as catalysts that are prepared according the fluorolytic sol-gel synthesis. AlF₃ is exclusively selective toward dehydrofluorination, whereas BaF₂ is 100% selective toward dehydrochlorination. The acid-base character of the catalysts is investigated and, as a result, mechanistic proposals for the dehydrofluorination and the dehydrochlorination are given. Thus, at high conversion level, selective catalysts for both dehydrofluorination and dehydrochlorination on the same substrate have been developed.

Full text of DOI:10.1016/j.jcat.2011.06.013

Journal of Catalysis 282 (2011) 175–182
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Highly selective metal fluoride catalysts for the dehydrohalogenation
of 3-chloro-1,1,1,3-tetrafluorobutane
Katharina Teinz a, Stefan Wuttke a,b, Fabian Börno a, Johannes Eicher c, Erhard Kemnitz a,
⇑
a Institut für Chemie, Humboldt-Universität zu Berlin, Brook-Taylor-Straße 2, 12489 Berlin, Germany
b Department of Chemistry and Center for NanoScience (CeNS), University of Munich (LMU), Butenandtstraße 11, 81377 München, Germany
c Solvay Fluor GmbH, Hans-Böckler-Allee 20, 30173 Hannover, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
For the first time, dehydrochlorination and dehydrofluorination reactions are studied on the same sub-
strate, 3-chloro-1,1,1,3-tetrafluorobutane, employing nanoscopic metal fluorides AlF3, MgF2, CaF2, SrF2,
and BaF2 as catalysts that are prepared according the fluorolytic sol–gel synthesis. AlF3 is exclusively
selective toward dehydrofluorination, whereas BaF2 is 100% selective toward dehydrochlorination. The
acid–base character of the catalysts is investigated and, as a result, mechanistic proposals for the dehy-
drofluorination and the dehydrochlorination are given. Thus, at high conversion level, selective catalysts
for both dehydrofluorination and dehydrochlorination on the same substrate have been developed.
Ó 2011 Elsevier Inc. All rights reserved.
Received 29 March 2011
Revised 6 June 2011
Accepted 12 June 2011
Available online 16 July 2011
Keywords:
Dehydrochlorination
Dehydrofluorination
Heterogeneous catalysis
Nanoscopic metal fluorides
Sol–gel synthesis
Lewis acid
Lewis base
1. Introduction
rac)-a-tocopherol (vitamin E) [5,6], the one-pot synthesis of men-
thol [7,8] and vitamin K1 as well as K1-chromanol [9]. However, all
these examples are either purely acid catalyzed or bi-functional-
ized with gold that catalyzes a hydrogenation step.
The goal of research in catalysis is currently accompanied by the
development of many different methods of preparing catalysts in
an attempt to optimize their properties. The aim is to develop a
catalyst that exhibits 100% selectivity for a desired product by
100% conversion of the reaction, thus preventing by-product for-
mation and eliminating waste. By doing so, on the one side, an
understanding of the already used catalysts is necessary, but on
the other side, new synthesis routes for catalysts must be
developed.
In 2003, our group explored the fluorolytic sol–gel synthesis for
the preparation of metal fluorides for the first time [1]. This syn-
thesis starts with metal alkoxides and nonaqueous [2] (or aqueous)
HF [3] followed by the gas-phase fluorination of the obtained dry
gel with mild fluorination agents. This yields high surface area me-
tal fluorides that have different properties in comparison with con-
ventionally prepared fluorides [4]. An advantage of these new
metal fluorides are their tunable surface properties with conse-
quences for catalytic applications. These new materials were al-
ready successfully used as heterogeneous catalysts possessing
high activity as well as a high selectivity in the synthesis of (all-
In this report, we focus on the acid–base pair properties of the
metal fluorides and the consequences for dehydrohalogenation
reactions. For this purpose, aluminum fluoride and alkaline earth
metal fluorides (MgF2, CaF2, SrF2, and BaF2) were synthesized
according to the fluorolytic sol–gel synthesis. It is expected that
Lewis acidity decreases from AlF3 to BaF2 due to the increase in
the ionic radius and the decrease in the positive charge density.
At the same time, the Lewis basicity should increase. The goal is
to investigate these properties and to relate acid–base with
catalytic properties. The metal fluorides were tested for the
dehydrohalogenation of the chlorofluoroalkane, 3-chloro-1,1,1,3-
tetrafluorobutane, where either dehydrochlorination or dehydro-
fluorination can take place.
Haloalkenes are very important starting compounds for halo-
gen-containing polymers (e.g. PVC, PVdC, PVDF, and PTFE). Because
of the easy access to a huge variety of haloalkanes, they are an
interesting source for haloalkenes production. The general ap-
proach toward haloalkenes starting from haloalkanes is either a
dehydrochlorination or a dehydrofluorination reaction. In both
cases, heterogeneous catalysts are required for a green chemical
synthesis.
⇑
Corresponding author. Fax: +49 3020937277.
E-mail address: erhard.kemnitz@chemie.hu-berlin.de (E. Kemnitz).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.06.013

176
K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
Cl
2,3-Dehydrofluorination
3,4-Dehydrofluorination
cis/trans-3-Chloro-1,1,1-trifluorobut-2-ene
F3C
F3C
C
C
H
CH3
CH2
Cl
C
3-Chloro-1,1,1-trifluorobut-3-ene
C
H2
F
Cl
catalyst
F3C
C
C
CH3
F
H2
3-Chloro-1,1,1,3-tetrafluorobutane
2,3-Dehydrochlorination
3,4-Dehydrochlorination
F3C
F3C
C
cis/trans-1,1,1,3-Tetrafluorobut-2-ene
C
H
CH3
CH2
F
C
1,1,1,3-Tetrafluorobut-3-ene
C
H2
Fig. 1. Reaction scheme for the dehydrohalogenation of 3-chloro-1,1,1,3-tetrafluorobutane.
For dehydrochlorination of chloroalkanes, only a few important
Hydrogen fluoride (Solvay Fluor GmbH), sodium hydrogen car-
bonate (Roth, >99.5%), aluminum isopropoxide (Aldrich, >98%),
magnesium ethoxide (Aldrich, 98%), calcium methoxide (Aldrich,
97%), strontium (Aldrich, granular 99%), and barium (Aldrich, rods
99+%) were used as received. Methanol (Aldrich, 99%), isopropanol
(Aldrich, 99%), and tetrahydrofuran (Aldrich, 99%) were dried
according the standard procedure before use. All operations were
carried out under inert conditions.
The methoxides of strontium and barium were prepared
through the reaction of the respective metal with methanol excess.
For the fluorolytic sol–gel synthesis, metal alkoxides were dissolved
or suspended in isopropanol (aluminum fluoride), tetrahydrofuran
(magnesium fluoride), or methanol (calcium, strontium, and bar-
ium fluoride), which was followed by the addition of a stoichiom-
etric amount of hydrogen fluoride dissolved in methanol. Caution:
HF is dangerous and must be handled with the appropriate precau-
tions. After fifteen hours, dry xerogels were obtained by evaporat-
ing the solvents from the wet gel under vacuum (ꢁ5 ꢂ 10ꢀ3 mbar)
and further drying at 100 °C for 2 h.
papers were published within the past decade. Mishakov et al. found
that nano-crystalline magnesium oxide is partly converted into
magnesium chloride during its reaction with 1-chlorobutane
(conversion rate: 0.004 l
mol sꢀ1 mꢀ2 at 200 °C), which gives the
alkene [10]. Furthermore, Pistarino et al. used silica–alumina sub-
strate to dehydrochlorinate 1,2-dichloropropane [11]. Kamiguchi
et al. used Nb-, Mo-, Ta-, W-, and Re-halogenide clusters for the
dehydrochlorination of 1-chloro- and 2-chloropentane and dehy-
drofluorination of 1-fluoropentane [12]. To the best of our knowl-
edge, no systematic studies have been performed so far
concerning the dehydrohalogenation of a chlorofluoroalkane where
both, dehydrochlorination and dehydrofluorination, are investi-
gated on the same substrate. Therefore, by applying different nano-
scopic metal fluorides as catalysts with tuned acid–base properties
at their surfaces, it was the intention of this investigation to under-
stand the role of the catalytic surface sites. Thus, based on gained
mechanistic insights, fine tuning of the acid–base properties and
optimized selectivity toward either of the dehydrohalogenation
reactions was anticipated.
The ‘‘BaF2 left at air’’ used for catalysis is a BaF2 xerogel that
Lewis acids are known to activate CAX bonds, which is the ini-
tial step of a dehydrohalogenation reaction. Thermodynamically,
CAF bond activation is less probable than CACl bond activation.
That is why only very strong Lewis acids should be active for dehy-
drofluorination whereas dehydrochlorination should be catalyzed
by far more Lewis acidic catalysts.
As a chlorofluorocarbon target molecule we chose 3-chloro-
1,1,1,3-tetrafluorobutane since this is an industrially interesting
and promising substrate resulting in valuable haloalkenes after HX
elimination. In Fig. 1, the reaction scheme is shown. Three different
dehydrochlorination and another three different dehydrofluorina-
tion products might be formed this way. If these two dehydrohalo-
genation reactions become catalyzed, further consecutive
reactions with the formed HX have to be considered, which would
result in two further possible side products, namely 1,1,1,3,3-penta-
fluorobutane and 3,3-dichloro-1,1,1-trifluorobutane.
after preparation was grinded at air.
2.2. Catalytic reactions
Aquartz tube(innerdiameter8 mm)with astopperofquartz wool
was used as steady flow reactor. 3-chloro-1,1,1,3-tetrafluorobutane
(Solvay Fluor GmbH, 84% 3-chloro-1,1,1,3-tetrafluorobutane, 7% 3-
chloro-1,1,1-trifluorobut-3-ene, 2% 3-chloro-1,1,1-trifluorobut-2-
ene, 2% 1,1,1,3,3-pentafluorobutane, and 5% other impurities) was
dosed via a gas saturator using nitrogen as carrier gas (nitrogen flow:
20 ml minꢀ1, total flow: 25 ml minꢀ1). If not mentioned otherwise,
the contact time was kept for 0.5 s (200–300 mg of xerogel, dumping
height: 0.4 cm) and the reaction temperature at 200 °C. The reaction
gas mixture leaving the reactor was passed through a 0.5 M sodium
hydrogen carbonate solution to trap acid gases. An analysis of the
reaction mixture was performed by gas chromatography by injecting
a sampleof 250
umn (50 m ꢂ 0.2 mm ꢂ 0.5
l
linto a Shimadzu GC17Aequipped with a PONAcol-
2. Experimental
lm), considering the impurities.
2.1. Fluorolytic sol–gel synthesis
2.3. Catalyst characterization
In course of the present investigations, the fluorolytic sol–gel
synthesis was for the first time also applied to strontium fluoride.
The syntheses of the other fluorides are already described in the lit-
erature [13].
X-ray powder diffraction measurements were performed using
a Seifert XRD3003TT diffractometer using Cu K radiation. Mois-
ture sensitive samples were prepared in a glove box and covered
a

K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
177
with a Kapton foil. The solids used for catalysis were identified
based on the ICSD powder diffraction file [14].
The chlorine content of the catalysts was determined by the
titration of the nitric acid neutralized alkaline pulping solution
with 0.01 N mercury perchlorate solution with diphenylcarbazone
as an indicator.
Surface area measurements were performed on a Micromeretics
ASAP 2020 at 77 K by nitrogen adsorption. Before the measure-
ment, the solids were degassed at 120 °C and 5 ꢂ 10ꢀ5 mbar for
twelve hours. Isotherms were processed by the Brunauer–
Emmett–Teller method (BET).
For FT-IR photoacoustic spectroscopy with ammonia as probe
molecule (NH3-PAS), an MTEC 300 photoacoustic cell inside a Dig-
ilab Excalibur FTS 300 spectrometer was used. Before the measure-
ment, the solids were preheated at 150 °C in a nitrogen flow,
loaded with ammonia, and purged with nitrogen.
The temperature-programmed desorption profiles of ammonia
(NH3-TPD) were recorded with a Perkin Elmer FTIR-System 2000
at 930 cmꢀ1. The preheated (200 °C) solid was loaded with ammo-
nia at 120 °C. After purging with nitrogen (10 ml minꢀ1, 30 min),
the temperature was raised by 10 K minꢀ1 until reaching 500 °C
while the desorbed ammonia was recorded.
Fig. 2. XRD patterns of catalysts after reaction: AlF3 (a), MgF2 (b, ꢃ PDF 1-1196),
CaF2 (c, ° PDF 3-1088), SrF2 (d, ^ PDF 88-2294), BaF2 (f, ꢁ PDF 3-304) and xerogel
BaF2 (e, + PDF 1-533).
IR spectra were recorded on a Nicolet Magna 550 spectrometer
equipped with a liquid nitrogen cooled MCT detector in transmis-
sion mode and purged by a dry nitrogen flow. The resolution was
4 cmꢀ1 and 128 scans were co-added for each spectrum, in a trans-
mission range of 1000–4000 cmꢀ1. The cell for IR analysis con-
sisted of vitreous silica and was equipped with CaF2 windows. A
moveable quartz sample holder permits the adjustment of the pel-
let in the infrared beam for acquisition of spectra and its displace-
ment into a furnace at the top of the cell for thermal treatments.
The cell was connected to a vacuum line and a glass injection loop
with a known volume (2.15 cm3). The sample was pressed into a
self-supporting pellet (area, 2 cm2; pressure, 1.5 tons; mass, 10–
20 mg) in air or in a dry glove box. After connecting the IR cell to
the vacuum line, the sample was heated at 473 K for 12 h. After
the activation of the sample, different probe molecules were added
in small doses.
barium fluoride that was left in air was significantly higher in com-
parison with the barium fluoride which was handled under inert
conditions. After four hours of reaction time, the maximum of chlo-
rine content was reached and remained constant over at least six-
teen additional hours of reaction. Based on these data, for the
catalytic active phase of in-air handled barium fluoride was calcu-
lated a total formula of BaCl0.7F1.3
.
3.1.3. Surface area
As shown in Table 2, the used fluorides provide different surface
areas. The main reason for these differences is due to the different
crystallization temperature of the xerogels. It was found that the
crystallization of nanoscopic MgF2 depends on the gas-phase com-
position at around 250 °C whereas AlF3 crystallizes above 300 °C
[15,16]. The quite diverse surface areas of the barium fluorides
can be rationalized by their different chemical composition. The in-
ert handled BaF2 represents BaCl0.3F1.7, whereas BaF2 air corre-
sponds to BaCl0.5F1.5 (Section 3.1.2) during catalysis.
3. Results
3.1. Catalyst characterization
3.1.4. Acidity: NH3-PAS and NH3-TPD
With NH3-PAS, strong Lewis acidic sites (1340 cmꢀ1) and med-
ium Brønsted acidic (1430 cmꢀ1) sites were found for aluminum
fluoride. Calcium fluoride and strontium fluoride show only
Brønsted acidic sites (1450 cmꢀ1). The occurrence of Brønsted
acidic sites is caused by the adsorption of hydrogen fluoride on Le-
wis acidic sites during the catalytic reaction [17]. In the case of alu-
minum fluoride, the adsorption of hydrogen fluoride was also
proven by MAS NMR (not shown here). Unfortunately, NH3-PAS
is not sensitive enough to find any acidic sites on magnesium fluo-
ride and barium fluoride – although we have already shown med-
ium Lewis acidity for magnesium fluoride probed with CO [18].
The NH3-TPD profiles for all of the studied fluorides are shown
in Fig. 3. For aluminum fluoride, an intense profile is found with
two maxima at ꢁ200 °C and ꢁ400 °C evidencing the presence of
strong acid sites [1]. For magnesium fluoride, calcium fluoride,
and strontium fluoride, ammonia desorption is already completed
at 300 °C due to medium acidic sites. Barium fluoride did not ad-
sorb any detectable amounts of ammonia.
If not mentioned, otherwise the characterization of the solids
was done after catalytic runs over two hours.
3.1.1. XRD patterns
In Fig. 2, XRD patterns of the studied fluorides are shown. Alu-
minum fluoride is X-ray amorphous. The alkaline earth metal fluo-
rides show broad reflexes and the maxima of diffractograms
correspond to the crystalline fluorides (see PDF card numbers in
figure caption). With the aid of Debeye–Scherrer equation, a crys-
tallite size of about 10 nm was estimated. XRD patterns of barium
fluoride after catalysis do not show reflexes for BaF2 as the xerogel
(inert as well as exposed to air), but they can be assigned to BaClF.
Obviously, a chlorination of the BaF2 xerogel occurs during the
reaction with 3-chloro-1,1,1,3-tetrafluorobutane. This chlorofluo-
ride phase is observed even after 20 h of reaction time, which indi-
cates its stability under the reaction conditions. Moreover, no
indications for BaCl2 formation were found.
3.1.2. Chlorine content
3.1.5. Acidity: Adsorption of CO on different BaF2-samples
In Table 1, chlorine contents of the used catalysts are shown.
Only for strontium fluoride as well as barium fluoride, chlorine
could be detected after the reaction. The chlorine content of the
CO is a suitable probe molecule for the characterization of Lewis
acidity because the
m
(CO) frequency is very sensitive to the local
(CO)
cationic environment [19]. The wave number of the linear
m

178
K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
Table 1
2162
Chlorine content of catalysts after reaction.
Catalysts
Reaction time
(h)
Chlorine content
(wt.%)
Chemical
composition
0.05
AlF3
MgF2
CaF2
2
2
2
–
–
–
–
–
–
SrF2
BaF2
BaF2 air
BaF2 air
BaF2 air
2
2
2
4
4
6
10
13
13
SrCl0.1F1.9
BaCl0.3F1.7
BaCl0.5F1.5
BaCl0.7F1.3
BaCl0.7F1.3
2154
2166
20
Table 2
Surface area of catalysts after reaction determined by the BET model.
2180
2170
2160
(cm-1)
2150
2140
Catalysts
Surface area (m2 gꢀ1
)
AlF3
MgF2
CaF2
220
40
250
40
Fig. 4. CO stretching region of the IR spectra of adsorbed CO on BaCl0.7F1.3
outgassed at 473 K and after adding successive doses of CO at 100 K and finally
1 Torr of CO at equilibrium pressure.
SrF2
BaF2
20
BaF2 air
60
of adsorbed CO on barium cations in zeolites [21] and on BaO/ZrO2
[22]. Hence, all the BaF2 samples possess very weak Lewis acid
sites.
3.1.6. Basicity: Adsorption of CHCl3 and pyrrole on the different BaF2-
samples
By using CHCl3 and pyrrole as probe molecules [18], we have
detected the basicity of a solid fluoride material for the first time.
For the prepared sol–gel MgF2, some Lewis basicity is observed, but
at the same time, the Lewis acid centers are the dominating sites in
the acid–base pair when a probe molecule is adsorbed. However,
by increasing the ionic radius with constant charge (from MgF2
to BaF2), a decrease in the Lewis acidity as well as an increase in
the Lewis basicity is expected.
CHCl3 fulfills the criteria for interacting separately with either a
Lewis acid site or a base site or simultaneously with both sites
[23,24]. Therefore, it can be predicted which of the acid–base pair
reaction is playing the key role during a catalytic process. CHCl3
adsorption reveals that all BaF2 samples (BaF2, BaF2 air, and
BaCl0.7F1.3) display the same bands and consequently the same
interactions occur with this probe molecule. Fig. 5 presents the
spectra for BaCl0.7F1.3, which is recorded after the introduction of
1 Torr of CHCl3 at equilibrium pressure (Fig. 5a). This is followed
by a progressive evacuation at room temperature (Fig. 5b–f). After
the adsorption of 1 Torr of CHCl3 at equilibrium pressure, the spec-
tra display a broad band at 3009 cmꢀ1 as well as a small band at
Fig. 3. NH3-TPD profiles of catalysts after reaction: AlF3 (a), MgF2 (b), CaF2 (c), SrF2
(d), BaF2 (e).
stretching vibration of molecules, adsorbed at Lewis acidic centers
without electronic back-donation, is shifted to higher wave num-
bers than in gaseous CO (m
gas = 2143 cmꢀ1). Thus, the higher the
3033 cmꢀ1. The blue shift of the
m(CH) band (m )
gas = 3019 cmꢀ1
blue shift, the higher the Lewis acidity. As it was already shown
with CO, AlF3 possesses very strong Lewis acidity [20], whereas
MgF2 shows acid centers with medium strength [18]. Therefore,
we investigated here the different BaF2 samples in more detail in
order to compare all these samples and to better understand the
catalytic results below.
[23] at 3033 cmꢀ1 indicates an interaction of the lone pair of a
chlorine atom of the chloroform molecule with a Lewis acid site
(Mx+ ClACHCl2). However, this interaction lowers the symmetry
of CHCl3 from C2 for the isolated molecule to the Cs point group.
m
This leads to a splitting of the m4 band (m
gas = 2116 cmꢀ1) [23],
All BaF2 samples (BaF2, BaF2 air, and BaCl0.7F1.3) show the same
bands after CO adsorption. Fig. 4 presents the spectra of BaCl0.7F1.3
recorded after adsorption of small doses of CO at 100 K. The first
dose displays a band at 2166 cmꢀ1 that slightly shifted at higher
CO coverage. After the adsorption of 1 Torr of CO at equilibrium
pressure, the band is observed at 2162 cmꢀ1. Since no perturbation
which could not be observed (Fig. 5). This result suggests that this
interaction is very weak and plays an insignificant role, which is
supported by the CO results that reveal very weak Lewis acidity
for coordinatively unsaturated Ba2+ sites. On the other hand, the
broad and intensive band at 3009 cmꢀ1 indicates the predomi-
nance of the Fꢄ ꢄ ꢄHACCl3 interaction. This is the first time that in
a fluoride material, the base is the dominating site in the acid–base
pair couple. This result is supported by the m4 band, which gives
rise to one band at 1219 cmꢀ1, meaning that the symmetry of
CHCl3 is the same as in the gas phase. Therefore, CHCl3 reveals
the dominating presence of Lewis basic sites in the acid–base pair
for all BaF2 samples.
occurs in the m(OH) range (not shown), this band can be assigned
to CO adsorbed on coordinatively unsaturated Ba2+ sites. Moreover,
the shoulder at lowest frequency (around 2154 cmꢀ1) only devel-
ops at high partial pressures and this is typical of weak and unspe-
cific adsorption of CO. The unusually low frequency of the
m(CO)
band on the unsaturated Ba2+ sites is in agreement with the result

K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
179
3383
1219
0.02
3220
0.01
a
3009
b
c
3300
d
0.01
3033
e
1240
1220
1200
4000 3800 3600 3400 3200 3000 2800 2600
(cm-1)
a
b
c
d
e
f
Fig. 6. IR spectra of BaCl0.7F1.3 outgassed at 473 K after adding 1 Torr of pyrrole at
equilibrium pressure (a) and then outgassed under primary vacuum for 5 min (b)
and 10 min (c) and under secondary vacuum for 5 min (d) and 10 min (e).
and barium fluoride handled in air, long-term measurements were
performed: even after twenty hours, the reaction rate of 3-chloro-
1,1,1,3-tetrafluorobutane was 0.06
in the case of aluminum fluoride and 0.15
version) for barium fluoride handled in air. The barium fluoride
l
mol sꢀ1 mꢀ2 (98% conversion)
3150
3100
3050
3000
(cm-1)
2950
2900
l
mol sꢀ1 mꢀ2 (92% con-
that had been handled under inert conditions displays a decrease
Fig. 5. IR spectra of BaCl0.7F1.3 outgassed at 473 K after adding 1 Torr (a) and
0.5 Torr (b) of CHCl3 at equilibrium pressure and then outgassing (primary vacuum)
for 1 min (c), 2 min (d), 5 min (e), and 10 min (f) (a–e are subtracted spectra from
the background).
in reactivity from 0.43
minutes to 0.34
mol sꢀ1 mꢀ2 (72% conversion) after two hours.
For magnesium fluoride, the reaction rate raises strongly during
the first ninety minutes and approaches 0.30
mol sꢀ1 mꢀ2 (83%
l
mol sꢀ1 mꢀ2 (90% conversion) after thirty
l
l
conversion) after two hours. The reaction rates found here are
one to two magnitudes higher than these observed by Mishakov
et al. for conversion of 1-chlorobutane over MgO at 200 °C
However, CHCl3 is not a suitable probe molecule to elucidate the
strength of Lewis basicity. For this purpose, pyrrole is a better probe
molecule because it is able to form H-bonded species via the NH
group with moderately basic surface centers [25]. Spectra recorded
on BaCl0.7F1.3 after introducing 1 Torr of pyrrole at equilibrium pres-
(0.004 l
mol sꢀ1 mꢀ2) [10].
Despite the different catalytic activity and surface area of the
both barium fluoride phases, they also differ in their degree of
chlorination (Table 1). The barium fluoride handled in air exhibited
significantly higher chlorine content than the inert handled one.
Obviously, chlorination is faster when moisture is present. Proba-
bly, traces of water accelerate the chlorination of barium fluoride
due to the formation of a thin liquid film of hydrogen chloride dis-
solved in water thus transforming the solid/gas into a solid/liquid
reaction. BaCl0.7F1.3 formed this way is thermodynamically stable
(Section 4) and the most active catalytic phase in the barium chlo-
ride fluoride system.
Aluminum fluoride is highly selective for the dehydrofluorina-
tion reaction of 3-chloro-1,1,1,3-tetrafluorobutane. The selectivity
toward dehydrochlorination increases in the alkaline earth metal
fluoride series with increasing atom number. Independent of its
handling, barium fluoride is very selective for dehydrochlorination
(Table 3).
In the case of calcium fluoride and strontium fluoride, which
catalyze both, the dehydrochlorination as well as the dehydrofluo-
rination of 3-chloro-1,1,1,3-tetrafluorobutane, the formation of
1,1,1,3,3-pentafluorobutane is observed. Very probably, it is
formed by the reaction of cis/trans-1,1,1,3-tetrafluorobut-2-ene,
formed via dehydrochlorination, with hydrogen fluoride also
formed during the dehydrofluorination reaction. Hence, it is obvi-
ously a consecutive reaction product of the dehydrochlorination
product with HF.
sure (Fig. 6a) show in the
3220 cmꢀ1. These two bands are red-shifted with respect to that of
the gas phase (
gas = 3497 cmꢀ1) indicating the presence of two
m
(NH) range two bands at 3383 cmꢀ1 and
m
different Lewis basic sites. The band at 3220 cmꢀ1 is not sensitive
to the outgassing procedure, whereas that at 3383 cmꢀ1 is red
shifted to 3300 cmꢀ1. This can be due to the fact that pyrrole ad-
sorbed on the fluorine atom sites is affected by inductive interac-
tions through the surface caused by pyrrole adsorption on
neighboring sites. Therefore, removing adsorbed pyrrole could
strength the other sites resulting in the observed red shift. This
behavior suggests the presence of two different Lewis basic sites:
the first being weak and the second with medium strength.
However, pyrrole adsorption that was found to be dissociative
on the pure BaF2 samples suggests the presence of medium/strong
Lewis basic centers. For these cases, pyrrole is not any more mean-
ingful due to the fact that it changes the acid–base surface proper-
ties of the sample. It can be concluded from the pyrrole adsorption
on the different BaF2 samples that BaCl0.7F1.3 possesses moderately
strong Lewis basicity whereas the pure BaF2 samples possess
stronger basic sites.
3.2. Catalytic results
After an initial period of ca. 90 min, stable reaction rates of 3-
chloro-1,1,1,3-tetrafluorobutane were observed for aluminum
fluoride, calcium fluoride, strontium fluoride, and barium fluoride
the latter handled in air (Fig. 7, Table 3). For aluminum fluoride
The desired product of the dehydrofluorination as well as the
dehydrochlorination reaction of 3-chloro-1,1,1,3-tetrafluorobutane
is that one with an end standing double bond (3-chloro-1,1,1-

180
K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
100
90
80
70
60
50
40
0.45
0.40
0.35
AlF3
MgF2
CaF2
SrF2
BaF2 inert
BaF2 air
0.30
0.25
0.20
0.15
0.10
0.05
0.00
30
60
90
120
30
60
90
120
reaction time in minutes
reaction time in minutes
Fig. 7. Conversion of 3-chloro-1,1,1,3-tetrafluorobutane during 120 min with different catalysts and corresponding reaction rates.
Table 3
Catalytic data for the dehydrohalogenation of 3-chloro-1,1,1,3-tetrafluorobutane (C: conversion, S: selectivity).
Catalysts
C3-chloro-1,1,1,3-tetrafluorobutane
Reaction rate (
l
mol sꢀ1 mꢀ2
)
Sdehydrofluorination
Sdehydrochlorination
S1,1,1,3,3-pentafluorobutane
AlF3
MgF2
CaF2
SrF2
BaF2
>99%
83%
>99%
96%
72%
98%
(0.06)
0.30
(0.06)
0.28
0.34
0.16
1.0
0.9
0.4
0.2
0.0
0.0
0.0
0.1
0.5
0.7
1.0
1.0
0.0
0.0
0.1
0.1
0.0
0.0
BaF2 air
Table 4
Dependency of the desired product selectivity on contact time.
Catalyst
AlF3
Contact time
C (%)
S3-chloro-1,1,1-trifluorobut-3-ene
Catalyst
BaF2 air
Contact time (s)
C (%)
S1,1,1,3-tetrafluorobut-3-ene
0.25
0.50
>99
>99
0.78
0.69
0.25
0.50
82
98
0.79
0.73
trifluorobut-3-ene, 1,1,1,3-tetrafluorobut-3-ene), which is kineti-
cally preferred. With decreasing contact time, the selectivity to-
ward these products increases, giving proof of the kinetic control
(Table 4).
Table 5
Reaction enthalpies: Reaction of metal fluoride with HCl to
metal chloride and HF.
DRH (kJ molꢀ1
)
AlF3
MgF2
CaF2
SrF2
263
121
71
24
ꢀ14
4. Discussion
Lewis acids are known to activate CAX bonds, which is the ini-
tial step of a dehydrohalogenation reaction. Thermodynamically
CAF bond activation is less probable than CACl bond activation
(490 kJ moleꢀ1 vs. 340 kJ molꢀ1) [26]. That is why, only very strong
Lewis acids are expected to be active for dehydrofluorination in
contrast to dehydrochlorination, which should be privileged for
most of Lewis acidic catalysts.
As predicted, aluminum fluoride as a strong Lewis acid cata-
lyzes the dehydrofluorination reaction. In contrast to our expecta-
tion, the dehydrochlorination reaction is completely suppressed.
Basically, dehydrohalogenation should proceed via an activated
surface complex of the respective haloalkane, which is initiated
in the first step by the interaction of the alkane halogen atom with
a coordinatively unsaturated surface site of the catalyst. The Lewis
acidity of the non-saturated surface sites is expected to play a cru-
cial rule regarding activity (conversion) and selectivity (dehydro-
fluorination vs. dehydrochlorination). However, additional factors
must play a role because thermodynamically dehydrochlorination
should be privileged at all the Lewis acidic catalysts used, meaning
CACl bond activation should be easier than CAF bond activation.
Moreover, barium fluoride possesses only very weak Lewis acid
BaF2
sites but catalyzes significantly and selectively the dehydrochlori-
nation reaction. Thus, we believe that for the dehydrochlorination
of 3-chloro-1,1,1,3-tetrafluorobutane, the Lewis acidity is not the
determining factor. In contrast, by decreasing the Lewis acidity
and increasing the Lewis basicity, the selectively toward dehydro-
chlorination increases.
Thus, we believe that the affinity of the metal toward the
respective halogen (strength of the metal-halogen bond) is a major
decisive factor. In Table 5, the reaction enthalpies of chloride for-
mation by reacting the respective metal fluoride with HCl are pre-
sented [27]. The formation of AlF3 is significantly favored over the
formation of AlCl3. As a consequence, a selective interaction of the
alkane F-atom with under coordinated aluminum surface sites of
AlF3 is expected in accord with the experimentally determined
behavior. The situation changes within the alkaline earth metal
series. With raising atom number, the thermodynamic tendency
of fluoride formation drops down. In case of barium halogenide

K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
181
CH2R
X
X
C
F
CH2CF3
Cl
Cl
C
CH2CF3
F
H
C
H
C
H
C
C
H
H
H
H
H
H
Al
Ba
Fig. 8. Proposed catalytic mechanism for the dehydrofluorination and dehydrochlorination of 3-chloro-1,1,1,3-tetrafluorobutane.
formation, the reaction enthalpy is close to zero, that is, why no
reaction pathway is favored and a barium chloride fluoride will
be formed. In correspondence with this thermodynamic prediction,
the reaction. Moreover, we successfully developed catalysts that
are highly selective in the catalyzed dehydrofluorination or dehy-
drochlorination reaction for a substrate that can perform both
reactions. We are now working to apply these catalysts on other sub-
strates to extend this application. The final goal is to demonstrate
that the sol–gel prepared nano-metal fluorides can be used as selec-
tive catalysts for the production of different haloalkenes from
haloalkanes.
chlorination of BaF2 resulting in the formation of BaCl0.7F1.3
which is the catalytically active phase – has been observed.
–
Thus, the different dehydrohalogenation selectivities of alumi-
num fluoride and barium fluoride are an interrelation between
the surface acid–base properties and the thermodynamic tendency
of the under coordinated metal surface sites with the respective
halogen atom. Going from aluminum fluoride to barium fluoride,
the Lewis acidity decreases dramatically and, at the same time,
the Lewis basicity increases. Moreover, also with this behavior,
the predominance of the acid–base pair changes from acid site to
basic site dominance. Last but not least, the affinity toward fluorine
decreases from aluminum fluoride to barium fluoride, whereas the
chlorine affinity increases. Therefore, a molecule like 3-chloro-
1,1,1,3-tetrafluorobutane that can interact (like CHCl3) with either
a Lewis acid site or a base site changes in the predominant type of
adsorption from AlF3 to BaF2. Calcium fluoride and strontium fluo-
ride catalyze both, the dehydrochlorination and the dehydrofluori-
nation of 3-chloro-1,1,1,3-tetrafluorobutane. These two fluorides
lay between aluminum fluoride and barium fluoride, meaning that
in these samples neither the acid–base pairs are single dominant
sites nor the affinity toward fluorine or chlorine is favored. Hence
both reactions, dehydrochlorination and dehydrofluorination,
might be rationalized as a consequence of the dominant hard–hard
(AlAF) and soft–soft (BaACl) interaction of the different types of
metal fluoride catalysts used.
Acknowledgments
We thank S. Bäßler and A. Zehl (Humboldt-University, Berlin)
for NH3-TPD and NH3-PAS analysis and chlorine content determi-
nation. K.T. is a stipendiate of the graduate school ‘‘Fluorine as
key element’’ (GRK 1582) of Deutsche Forschungsgemeinschaft,
DFG.
References
[1] E. Kemnitz, U. Groß, S. Rüdiger, C.S. Shekar, Angew. Chem. 115 (2003) 4383;
E. Kemnitz, U. Groß, S. Rüdiger, C.S. Shekar, Angew. Chem. Int. Ed. 42 (2003)
4251.
[2] S. Rüdiger, U. Groß, E.J. Kemnitz, Fluorine Chem. 128 (2007) 353.
[3] S. Wuttke, S.M. Coman, G. Scholz, H. Kirmse, A. Vimont, M. Daturi, S.L.M.
Schroeder, E. Kemnitz, Chem. Eur. J. 14 (2008) 11488.
[4] S. Wuttke, S.M. Coman, J. Kröhnert, F.C. Jentoft, E. Kemnitz, Catal. Today 152
(2010) 2.
[5] S.M. Coman, S. Wuttke, A. Vimont, M. Daturi, E. Kemnitz, Adv. Synth. Catal. 350
(2008) 2517.
[6] N. Candu, S. Wuttke, E. Kemnitz, S.M. Coman, V.I. Parvulescu, Appl. Catal. A:
Gen. 391 (2011) 169.
[7] A. Negoi, S. Wuttke, E. Kemnitz, D. Macovei, V.I. Parvulescu, C.M. Teodorescu,
S.M. Coman, Angew. Chem. Int. Ed. 49 (2010) 8134.
[8] S.M. Coman, V.I. Parvulescu, S. Wuttke, E. Kemnitz, Chem. Commun. (2009)
460.
[9] S.M. Coman, V.I. Parvulescu, S. Wuttke, E. Kemnitz, ChemCatChem 2 (2010) 92.
[10] (a) I.V. Mishakov, A.F. Bedilo, R.M. Richards, V.V. Chesnokov, A.M. Volodin, V.I.
Zaikovskii, R.A. Buyanov, K.J. Klabunde, J. Catal. 206 (2002) 40;
(b) V.B. Fenelenov, M.S. Mel’gunov, I.V. Mishakov, R.M. Richards, V.V.
Chesnekov, A.M. Volodin, K.J. Klabunde, J. Phys. Chem. B 125 (2001) 3937.
[11] C. Pistarino, E. Finocchio, M.A. Larrubia, B. Serra, S. Braggio, M. Baldi, Ind. Eng.
Chem. Res. 40 (2001) 3262.
As a consequence, we would describe the catalytic mechanism
as a simultaneous interaction of the halogen with the metal site
and of a catalysts surface fluorine with a proton (Fig. 8). Thereby,
the character of the metal site (affinity toward either to fluorine
or chlorine, acid–base properties) defines the reaction pathway
(dehydrochlorination or dehydrofluorination).
5. Conclusions
[12] S. Kamiguchi, M. Watanabe, K. Kondo, M. Kodomari, T. Chihara, J. Mol. Cat. A:
Chem. 203 (2003) 153.
[13] U. Gross, S. Rüdiger, E. Kemnitz, Solid State Sci. 9 (2007) 838.
[14] PDF-2 Database [Sets 1-51 plus 70-89], International Centre for Diffraction
Data, 2001.
[15] S. Ruediger, U. Groß, M. Feist, H.A. Prescott, S.C. Shekar, S.I. Troyanov, E.
Kemnitz, J. Mater. Chem. 15 (2005) 588.
[16] S. Wuttke, G. Scholz, S. Rüdiger, E. Kemnitz, J. Mater. Chem. 17 (2007) 4980.
[17] S. Rüdiger, G. Eltanany, U. Groß, E. Kemnitz, J. Sol–Gel Sci. Technol. 41 (2007)
299.
[18] S. Wuttke, A. Vimont, J.C. Lavalley, M. Daturi, E. Kemnitz, J. Phys. Chem. C 114
(2010) 5113.
[19] K.I. Hadjiivanov, G.N. Vayssilov, Adv. Catal. 47 (2002) 307.
[20] T. Krahl, A. Vimont, G. Eltanany, M. Daturi, E. Kemnitz, J. Phys. Chem. C 111
(2007) 18317.
[21] P. Li, Y. Xiang, V.H. Grassian, S.C. Karsen, J. Phys. Chem. B 103 (1999) 5058.
[22] V. Sadykov, R. Bunina, G. Alikina, A. Ivanova, D. Kochubei, B. Novgorodov, E.
Paukshtis, V. Fenelonov, V. Zaikovsky, T. Kuznetsova, S. Beloshapkin, V.
Aluminum fluoride and the alkaline earth metal fluorides
prepared via the fluorolytic sol–gel synthesis are very selective
dehydrohalogenation catalysts. It was found that dehydrofluorina-
tion is selectively catalyzed by strong Lewis acids (AlF3, MgF2), while
dehydrochlorination is completely suppressed. Interestingly
enough, BaF2 carries very weak acidic surface sites but catalyzes
the dehydrochlorination selectively. The thermodynamic data show
that the metal affinity toward chlorine or fluorine plays a decisive
role. Reasonably, the reaction pathway is determined by the charac-
ter of the metal site (affinity toward fluorine or chlorine, acid–base
properties). Thus, we could successfully relate acid–base and
thermodynamic properties of the catalysts with the catalytic behav-
ior of the solid catalysts and with this giving mechanistic insight of

182
K. Teinz et al. / Journal of Catalysis 282 (2011) 175–182
Kolomichuk, E. Moroz, V. Matyshak, G. Konin, A. Rozovsky, J.R.H. Ross, J. Breen,
J. Catal. 200 (2001) 117.
[23] S. Huber, H. Knözinger, J. Mol. Catal. A 141 (1999) 117.
[24] J.C. Lavalley, Catal. Today 27 (1996) 377.
[25] C. Binet, A. Jadi, J. Lamotte, J.C. Lavalley, J. Chem. Soc., Faraday Trans. 92 (1996)
123.
[26] NIST Chemistry WebBook – NIST Standard Reference Database Number 69, US
Secretary of Commerce on behalf of the United States of America, 2008.
[27] D.R. Lide, H.P.R. Frederikse, CRC Handbook of Chemistry and Physics, CRC
Press, Inc., 1996.