Enhanced dichlorodifluoromethane decomposition with selective fluorine absorption by acidic fluoriated magnesinm oxide

Article abstract of DOI:10.1246/bcsj.77.1239

CCl₂F₂ decomposition with simultaneous halogen absorption by partially fluorinated MgO (MgF₂-MgO) was studied, focusing on the effects of the acidity of the surface. CCl₂F ₂ decomposition by MgF₂-MgO was greatly promoted by fluorination of MgO at 10 mol% or higher. CCl₂F₂ decomposition was not a catalytic process over MgF₂-MgO but basically a selective fluorine absorption reaction with MgO to form MgF₂. Chlorine was released in the form of CCl₄ regardless of reaction temperature and degree of fluorination of MgO due to low reactivity to CCl ₄. NH₃-TPD and pyridine adsorption experiments were carried out to characterize the acid sites on MgF₂-MgO samples. The amount of acid sites became maximum for 10% MgF₂-MgO and the strength of acid sites increased as fluorination proceeded. CCl₂F₂ decomposition was revealed to be initiated by Lewis acid sites on MgF ₂-MgO formed by CCl₂F₂ decomposition as well as by fluorination with hydrofluoric acid aqueous solution. Thus, formation of the Lewis acid sites was considered to be the key step for efficient CCl ₂F₂ decomposition with selective fluorine absorption.

Full text of DOI:10.1246/bcsj.77.1239

ꢀ 2004 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 77, 1239–1247 (2004) 1239
Enhanced Dichlorodifluoromethane Decomposition with Selective
Fluorine Absorption by Acidic Fluorinated Magnesium Oxide
ꢀ
Tsukasa Tamai, Koji Inazu, and Ken-ichi Aika
Department of Environmental Chemistry and Engineering, Interdisciplinary Graduate School of Science and Engineering,
Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8502
Received October 2, 2003; E-mail: kenaika@chemenv.titech.ac.jp
CCl2F2 decomposition with simultaneous halogen absorption by partially fluorinated MgO (MgF2–MgO) was stud-
ied, focusing on the effects of the acidity of the surface. CCl2F2 decomposition by MgF2–MgO was greatly promoted by
fluorination of MgO at 10 mol% or higher. CCl2F2 decomposition was not a catalytic process over MgF2–MgO but ba-
sically a selective fluorine absorption reaction with MgO to form MgF2. Chlorine was released in the form of CCl4 re-
gardless of reaction temperature and degree of fluorination of MgO due to low reactivity to CCl4. NH3-TPD and pyridine
adsorption experiments were carried out to characterize the acid sites on MgF2–MgO samples. The amount of acid sites
became maximum for 10% MgF2–MgO and the strength of acid sites increased as fluorination proceeded. CCl2F2 decom-
position was revealed to be initiated by Lewis acid sites on MgF2–MgO formed by CCl2F2 decomposition as well as by
fluorination with hydrofluoric acid aqueous solution. Thus, formation of the Lewis acid sites was considered to be the key
step for efficient CCl2F2 decomposition with selective fluorine absorption.
Development of decomposition methods for chlorofluorocar-
CCl2F2(g) þ 2MgO(s) ! CO2(g) þ MgCl2(s) þ MgF2(s)
bons (CFCs), which are one of major materials that deplete the
ꢀH0 ¼ ꢁ477 kJ molꢁ1
:
ð1Þ
stratospheric ozone layer, is an urgent issue. The catalytic hy-
drolysis of CFCs is a desirable method because it can be carried
out under mild reaction conditions,1–8 but during the decompo-
sition corrosive HCl and HF are formed and thus gas neutraliz-
ers and/or corrosion-proof systems are required. These often
cause costly set-up or complicated operation in practical use.
In most of the methods for CFC decomposition including cata-
lytic processes, the reaction effluent is washed with NaOH
aqueous solution and/or Ca(OH)2 to eliminate HF and HCl
as halides.
Recently, for decomposition of halocarbons, direct fixation
of halogens as metal halides, mainly alkaline metal and alkaline
earth metal halides, have been reported. Destructive fixation of
chlorocarbon and hydrochlorocarbon as metal chlorides has
been well investigated. Weckhuysen et al. reported CCl4 de-
composition by La2O3, CeO2, and various alkaline earth metal
oxides9,10 and Klabunde and co-workers investigated destruc-
tive adsorption of CCl4 by nano-scale structured MgO and
CaO.11–16 They found that deposition of transition metal oxides
to MgO and CaO promoted conversion of CCl4 and adsorption
capacity for chloride.11–14 However, for fluorine containing
halocarbons such as CFCs and perfluorocarbons there have
been just a few reports, including a report regarding CCl2F2 de-
composition by sodium oxalate by Crabtree et al.17–19 We have
focused on transition metal loaded MgO-based materials for
CFC decomposition,20 since MgO is inexpensive and available
as material for uses such as catalyst support. Applying these
materials to CCl2F2, which was widely used as refrigerant,
we found that addition of transition metal oxides to MgO to
be quite effective; vanadium oxides especially were excellent
additives that exhibit high conversion and selectivity to CO2
and halides. In this system, the target reaction is as follows:
Supporting vanadium oxides on MgO enabled halogen ab-
sorption efficiency to be as high as ca. 80% when CCl2F2
was introduced in stoichiometric amounts to MgO at 723 K
in a flow reaction system. Promotion of CCl2F2 decomposition
by transition metal addition was considered to be related to the
catalysis of supported transition metal oxides, i.e., the decom-
position was initiated by the reaction of transition metal oxides
with CCl2F2 to form transition metal halides. Most of these are
unstable under the reaction conditions and readily react with
MgO to generate transition metal oxides with magnesium hal-
ides as resulting products. However, this system also has a se-
rious problem for practical use. In spite of their important role
as intermediates, vanadium halides and/or oxyhalides will be
deposited in the cooler part of the reactor due to their high vol-
atility.
On the other hand, some research groups reported that acid
sites on solid catalysts work as active sites for catalytic decom-
position of CFCs6–8 and for dismutation or halogen exchange
reaction for halocarbons.21–23 The acid sites can be generated
by fluorination of the catalysts such as titanium dioxides,6,8
zeolites,7 and aluminium oxides21–23 by CFCs and HCFCs.
Their effects were considered in connection with the induction
period of the reaction. If the acid sites are introduced on MgO
by such fluorination, high conversion of CFC can be expected.
However, no study has been conducted on introducing acid
sites to a basic oxide, MgO. In this study we have examined
the effectiveness of fluorination of MgO to CCl2F2 decomposi-
tion by MgF2–MgO with simultaneous halogen fixation as
magnesium halides. Fluorination of MgO by soaking in hydro-
gen fluoride aqueous solution was found to be an effective
Published on the web June 9, 2004; DOI 10.1246/bcsj.77.1239

1240 Bull. Chem. Soc. Jpn., 77, No. 6 (2004)
Enhanced Decomposition of CCl2F2 by MgF2–MgO
and MCT detection.
method to generate acid sites on MgO as well as fluorination of
MgO during CCl2F2 decomposition. Introducing such acid sites
on MgO greatly enhanced the reaction of CCl2F2 with MgF2–
MgO to form MgF2, CO2, and CCl4. The acid sites were char-
acterized by NH3-TPD and pyridine adsorption experiments to
reveal their relevance to CCl2F2 decomposition and halogen
absorption by MgF2–MgO.
Results and Discussion
CCl2F2 Decomposition by Partially Fluorinated MgO.
As shown in Fig. 1(a), CCl2F2 did not substantially react with
MgO at 623 K, however, CCl2F2 was decomposed to form CO2
and CCl4 as major gas-phase product by the reaction with the
MgO samples partially fluorinated by dilute hydrofluoric acid
aqueous solution (MgF2–MgO). Almost complete conversion
of CCl2F2 was observed for all the MgF2–MgO samples that
had 5% or higher fluorination by mole. For a sample to exhibit
significant activity to CCl2F2 decomposition, 5% was a thresh-
old value for the extent of fluorination of MgO. An induction
period of about 1 h was also observed for the decomposition
by MgF2–MgO except for 5% MgF2–MgO, which required
2.5 h to attain CCl2F2 conversion of more than 80%. There
was a tendency for the induction period to be shortened with
an increase in percentage fluorination of MgO. On the other
hand, as shown in Fig. 1(b), the selectivity to CO2 was almost
constant at about 50% during the reaction time examined for
the samples other than 5% MgF2–MgO, with which an induc-
tion period of 2.5 h was observed. The rest of the carbon from
CCl2F2 was dominantly in the form of CCl4, an undesired prod-
uct, as mentioned above, although a trace amount of CCl3F was
also observed as the reaction prolonged. The observed distribu-
tion of gas-phase product indicates that decomposition of
Experimental
Materials. MgO was prepared by following procedure: Com-
mercial MgO powder (Ube Materials Industries) was suspended in
deionized water and the suspension was stirred overnight at room
temperature. After being dried, the resulting Mg(OH)2 was cal-
cined at 873 K for 3 h under helium flow to form MgO with high
surface area. Partially fluorinated MgO was prepared by adding HF
to aqueous suspensions of MgO prepared by the method mentioned
above, followed by drying procedure and calcination under the
same conditions as those for MgO. The samples are denoted, for
example, 10% MgF2–MgO for partially fluorinated MgO with
the degree of fluorination at MgF2/(MgF2 + MgO) = 0.1 mol/
mol.
CCl2F2 Decomposition. Reactions of CCl2F2 with MgO and
partially fluorinated MgO were performed using a fixed-bed flow
reactor system under atmospheric pressure. 1% CCl2F2 balanced
with helium was fed to 0.2 g of the MgO samples placed in a tub-
ular quartz reactor at a total flow rate of 30 mL/min after MgO
samples were pretreated under helium flow at 873 K for 3 h. Anal-
ysis of outlet gas was carried out by using a GC-TCD (Shimadzu
GC-14B) with a Porapak Q column to determine conversion of
CCl2F2 and selectivity to carbon dioxide. Feeding a stoichiometric
amount of CCl2F2 to the samples was attained at the reaction time
of 3.3 h. When complete conversion of CFCs was achieved, MgO
would be totally consumed to be MgCl2 and MgF2 during this
period of time.
CCl4 Decomposition. CCl4 decomposition by the MgO sam-
ple was carried out using a fixed-bed pulse reaction system directly
connected to GC-TCD. A pulse injection of CCl4 (Wako, 99%)
was made to the same reactor as that used for CCl2F2 decomposi-
tion system at 723 K with a precision syringe.
Characterization. X-ray diffraction patterns of the samples
were obtained with Rigaku Multi-flex S with Cu Kꢀ radiation as
an incident X-ray source in the range of 20 to 70 degrees as 2ꢁ.
The specific surface area of each sample was determined by nitro-
gen adsorption at 77 K with BELSORP 28SA (BEL Japan).
NH3-TPD experiments were carried out for MgO samples by us-
ing a Pyrex glass system equipped with a quartz U-tube reactor.
The samples were pretreated at 873 K for 2 h under evacuation. Af-
ter the sample was cooled down to 373 K, gaseous ammonia was
introduced to the system and then adsorption equilibrium was at-
tained within 30 min. Temperature programmed desorption of am-
monia was performed at a temperature ranging from 373 to 873 K
with detection by TCD (Shimadzu GC-8A) after removing physi-
sorbed ammonia under vacuum for 0.5 h.
The FT-IR analysis for the MgO samples were performed in a
quartz cell connected to a closed system made of Pyrex. A self-sup-
ported thin sample disc (ca. 0.1 g) was placed in the quartz IR cell
and pretreated at 873 K for 2 h under vacuum. And then about 10
Torr of pyridine was introduced to the cell at room temperature to
be adsorbed for 30 min, followed by evacuation at given tempera-
tures for 1 h. IR spectra were taken with JASCO VALOR III under
Fig. 1. Time course of CCl2F2 conversion (a) and CO2 se-
lectivity (b) in CCl2F2 decomposition at 623 K by partially
fluorinated MgO. Percentage fluorination of MgO: 0% ( ),
5% ( ), 10% ( ), 20% ( ), 40% ( ).
the following conditions: 32 repeated scans, resolution of 2 cmꢁ1
,

T. Tamai et al.
Bull. Chem. Soc. Jpn., 77, No. 6 (2004) 1241
Fig. 2. Relationship between degree of fluorination and
CCl2F2 conversion during CCl2F2 decomposition for par-
tially fluorinated MgO with different initial percentage
fluorination at 623 K. Percentage fluorination of MgO:
5% ( ), 10% ( ), 20% ( ), 40% ( ).
CCl2F2 by MgF2–MgO samples takes place with the following
reaction (Eq. 2):
2CCl2F2(g) þ 2MgO(s) ! CO2(g) þ CCl4(g) þ 2MgF2(s)
ꢀH0 ¼ ꢁ578 kJ molꢁ1
:
ð2Þ
Along with the reaction of Eq. 2, fluorine of CCl2F2 is fixed
in the solid phase exclusively as MgF2. It means that the de-
composition of CCl2F2 by MgF2–MgO is accompanied by se-
lective fluorine absorption to yield MgF2 as the only solid phase
product. As decomposition of CCl2F2 proceeded nearly accord-
ing to Eq. 2, the conversion of MgO can be calculated from the
yield of CO2 in this system. Figure 2 shows the dependence of
CCl2F2 conversion upon the mole percentage of MgF2 in
MgF2–MgO for each sample. Initially, CCl2F2 conversion in-
creased with an increase in mole percentage of MgF2 in
MgF2–MgO. When the mole percentage of MgF2 became
80% or higher, CCl2F2 conversion steeply decreased and the
decomposition will be terminated at around 90%. This suggests
that the induction period observed at the initial stage of the de-
composition is the period to attain a certain concentration level
of MgF2 needed to exhibit high reactivity to CCl2F2.
Fig. 3. Temperature dependence of CCl2F2 conversion (a)
and CO2 selectivity (b) in CCl2F2 decomposition by 10%
MgF2–MgO. Reaction temperature: 573 K ( ), 623 K
( ), 673 K ( ), 723 K ( ), 773 K (ꢂ).
Since reaction temperature should be an important factor af-
fecting the decomposition and product distribution, tempera-
ture dependence of CCl2F2 conversion and CO2 selectivity
was investigated for 10% MgF2–MgO. Results are shown in
Fig. 3(a) and (b), respectively. At 573 K, CCl2F2 conversion
was quite low. Therefore the result of CO2 selectivity at this
temperature was omitted due to the low and inaccurately meas-
ured yield.
Reactions at temperatures above 623 K gave almost com-
plete CCl2F2 conversion at the reaction time of 2 h and stable
CO2 formation with the selectivity of ca. 50%. At reaction tem-
peratures above 673 K, no induction period for CCl2F2 conver-
sion was observed without any changes in CO2 selectivity. It
indicates that the decomposition at 623 K or higher tempera-
tures proceeds along with the same pathway. Temperature de-
pendence of the reaction of CCl2F2 with MgO was also exam-
ined for comparison with MgF2–MgO samples and the result is
shown in Fig. 4. When the reaction temperature became 773 K,
conversion of CCl2F2 started after the induction period of 2.5 h.
Fig. 4. Temperature dependence of CCl2F2 decomposition
by MgO. Reaction temperature: 723 K ( ), 773 K ( ,
), 823 K (
(open), CO2 selectivity (close).
,
), 873 K (
,
). CCl2F2 conversion
The induction period decreased to 1.5 h at 823 K, which is com-
parable to that for MgF2–MgO at 623 K. Although the reprodu-
cibility of the induction period was not excellent probably due
to slight differences in the preparation conditions of the sam-
ples, higher reaction temperatures definitely showed shorter in-
duction periods and conversion of CCl2F2 reached 100% after

1242 Bull. Chem. Soc. Jpn., 77, No. 6 (2004)
Enhanced Decomposition of CCl2F2 by MgF2–MgO
Table 1. Particle Size of MgO and MgF2 in Partially Fluori-
nated MgO
the induction period. Once the conversion occurred, no signifi-
cant differences in stability of gas-phase product distribution,
CCl2F2 conversion, and CO2 selectivity were observed. These
observations indicate that MgO requires a temperature higher
by 150–200 K than MgF2–MgO does to be capable to decom-
pose CCl2F2 and that MgO can be activated by partial fluorina-
tion by the reaction with CCl2F2 in the same manner as fluori-
nation by hydrofluoric acid. It is also confirmed by these obser-
vations that partial fluorination of MgO greatly enhances the
decomposition of CCl2F2 with selective fluorine absorption.
XRD Analysis and NH3-TPD Measurement. As describ-
ed above, partial fluorination of MgO using hydrofluoric acid
formed highly reactive MgF2–MgO to CCl2F2 decomposition
with selective fluorine absorption. Results also suggested that
MgF2–MgO can be produced by the reaction with CCl2F2 at
temperatures above 623 K. To confirm the chemical composi-
tion of MgF2–MgO samples prepared with dilute hydrofluoric
acid aqueous solution, we performed X-ray diffraction (XRD)
analysis for five kinds of MgF2–MgO with different degrees
of fluorination, as well as MgO and MgF2 (Fig. 5). As easily ex-
pected, diffraction peaks that originated from MgF2 were de-
tected for most of the MgF2–MgO samples. Although the fluo-
rination at 5% and 10% significantly modified the base MgO, as
seen in the activity to CCl2F2 decomposition, the diffraction
peaks from MgF2 were detected only for the samples with
the degrees of fluorination of 20% or higher. This limitation
should be either due to lack of sufficient sensitivity with
XRD or due to very small particle sizes of MgF2 crystallites
in 5% and 10% MgF2–MgO samples.
Percentage
fluorination
Particle size/nm
MgOaÞ
MgF2
bÞ
0cÞ
5
10
20
8.8
8.6
7.5
n.d.dÞ
n.d.dÞ
n.d.dÞ
27
7.0
40
70
100eÞ
n.d.dÞ
n.d.dÞ
n.d.dÞ
29
25
28
a) Determined by XRD peak of MgO (200) at 2ꢁ ¼ 42:9 de-
gree. b) Determined by XRD peak of MgF2 (110) at
2ꢁ ¼ 27:3 degree. c) Conventionally prepared MgO. d) Not de-
termined because of low peak intensity or overlap of another
peak. e) Commertial MgF2.
Table 2. BET Surface Area and Results of NH3-TPD for
Partially Fluorinated MgO
Percentage
fiuorination
BET surface
area/m2 gꢁ1
NH3-adsorbedaÞ
mmol gꢁ1 mmol mꢁ2
0
5
192
179
233
187
127
58
18
86
153
146
127
76
0.09
0.48
0.66
0.78
1.00
1.31
—
10
20
40
70
100
<1
trace
As shown in Table 1, the particle size of MgO, which was
calculated from full width at half height (FWHM) of a diffrac-
tion peak at 2ꢁ of 42.9ꢃ that originated from the (200) facet, be-
came slightly smaller as MgO became highly fluorinated. As
reported by Shimbo et al., covering MgO particles with hydro-
fluoric acid can prevent the particles from sintering during heat
a) Adsorbed at 373 K, desorbed from 373 to 873 K.
treatment and hence core MgO in MgF2–MgO particles can
keep their size as small as they were originally even after treat-
ment at 873 K.24 On the other hand, for MgF2 in MgF2–MgO,
the difference in degree of fluorination caused no clear tenden-
cy for changes in the particle size, which was calculated from
FWHM of a peak at 2ꢁ of 27.3ꢃ that originated from the
(110) facet. The particle sizes of MgF2 of MgF2–MgO samples,
ranging from 25 to 30 nm, were coincidentally similar to that of
commercial MgF2 (28 nm).
NH3-TPD was carried out to estimate the amount and the
strength of acid sites on MgF2–MgO samples. The results of
BET surface area and NH3-TPD measurement are listed in
Table 2 and the NH3-TPD profiles are shown in Fig. 6. The
largest BET surface area of MgF2–MgO samples was obtained
with 10% MgF2–MgO (233 m2 gꢁ1) and the degree of fluorina-
tion higher than 10% gave smaller surface area. Low specific
surface area of MgF2 can account for this tendency, even
though the particle size of MgO was kept small during the fluo-
rination in a similar manner to the case described above. The
amount of adsorbed ammonia increased drastically with an in-
crease in degree of fluorination up to 10%. 10% MgF2–MgO
had the largest amount of adsorbed ammonia of 153 mmol gꢁ1
.
Higher degree of fluorination than 10% gave smaller amount of
adsorbed ammonia, however, surface density of acid sites,
i.e., specific amount of adsorbed ammonia by surface area, in-
creased with an increase in the degree of fluorination. Commer-
cial MgF2 had quite small surface area and no meaningful
amount of NH3 adsorption.
Fig. 5. XRD patterns of partially fluorinated MgO. Percent-
age fluorination of MgO: 0% (MgO) (a), 5% (b), 10% (c),
20% (d), 40% (e), 70% (f), 100% (MgF2) (g). MgO ( ),
MgF2 ( ).

T. Tamai et al.
Bull. Chem. Soc. Jpn., 77, No. 6 (2004) 1243
Fig. 6. NH3-TPD profiles of partially fluorinated MgO. Per-
centage fluorination of MgO: 0% (MgO) (a), 5% (b), 10%
(c), 20% (d), 40% (e), 70% (f), 100% (MgF2) (g).
The NH3-TPD profiles shown in Fig. 6 indicate that desorp-
tion temperature of NH3 shifted to higher as fluorination pro-
ceeded, and from 20% of fluorination a broad peak that spread
from 450 to 750 K was observed. It clearly indicates that there
were acid sites with different strength on MgF2–MgO, and that
the strength of acid sites increased by partial fluorination of
MgO as well as the number of acid sites. Commercial MgF2
has desorption peaks at 450 K, which shows weak acidity.
IR Observation of Pyridine Adsorption. Pyridine has
been used as a typical probe adsorbate molecule to characterize
surface acidity. It has also been found that pyridine interacts
with ionic metal oxides in three different modes. The first mode
corresponds to transfer of a proton from surface hydroxy group
to form a pyridinium ion (Brꢁnsted, BPy). The second one cor-
responds to coordination of the lone pair of nitrogen atom with
a Lewis acid site such as some metal ion (LPy bonding mode).
The last one, the weakest interaction among the three, is made
through hydrogen bonding of pyridine to surface OH groups.
Since MgO is a basic oxide, there have been quite a few reports
dealing with the acidity of MgO.25,26 According to the above
NH3-TPD results, acid sites on MgF2–MgO samples were con-
sidered to be generated by fluorination of MgO. To characterize
the acid sites, we carried out pyridine adsorption experiments
for partially fluorinated MgO by means of FT-IR. Before pyri-
dine was introduced onto the samples, each sample was pre-
treated under vacuum at 873 K for 2 h and then cooled down
to room temperature. The IR spectrum after pretreatment was
used as background spectrum for pyridine adsorption experi-
ments for each sample. Pyridine was introduced onto the sam-
ples for 30 min at ambient temperature and then evacuated at
elevated temperatures from 298 to 573 K. The spectra obtained
are shown in Fig. 7 for MgO (a), 10% MgF2–MgO (b), and 40%
MgF2–MgO (c). In the range from 1400 to 1630 cmꢁ1, four dis-
tinct peaks were observed at 1440–1451, 1480–1496, 1570–
1580, and 1595–1615 cmꢁ1. These peaks originate from the
ring vibration mode of pyridine assigned to 19b, 18a, 8b, and
8a modes by Kline and Turkevich.27 19b and 8a modes are of-
ten used to distinguish between LPy mode, physisorbed, and
BPy. Physisorbed or hydrogen-bonded pyridine gave the bands
at 1440–1445 cmꢁ1 (19b) and 1580–1600 cmꢁ1 (8a), while pyri-
dine in LPy mode gave the bands at 1445–1460 cmꢁ1 (19b)
Fig. 7. IR spectra of pyridine adsorbed on MgO (a), 10%
MgF2–MgO (b), and 40% MgF2–MgO (c) at desorption
temperatures ranging from 298 to 573 K.
and 1602–1634 cmꢁ1 (8a). BPy mode gives the bands at
.
1540 cmꢁ1 and 1640 cmꢁ1 28 After evacuation at room temper-
ature, pyridine on MgO shows distinct adsorption bands of 19b
and 8a at 1445 cmꢁ1 and 1598 cmꢁ1, respectively. However,
the peaks from excess and weakly adsorbed pyridine became
smaller by evacuation at higher temperatures and they disap-
peared at 473 K. This observation indicates that adsorption of
pyridine on MgO is a weak physical adsorption and thus the
peaks at 1445 cmꢁ1 (19b) and 1598 cmꢁ1 (8a) can be assigned
as originating from physically adsorbed pyridine or as adsorbed

1244 Bull. Chem. Soc. Jpn., 77, No. 6 (2004)
Enhanced Decomposition of CCl2F2 by MgF2–MgO
on very weak Lewis acid sites, which can remain on the surface
under vacuum at temperatures below 473 K.
changed by partial fluorination to form MgF2–MgO. MgO
was also fluorinated during CCl2F2 decomposition, and 5%
MgF2–MgO was employed as a typical MgF2–MgO sample
to observe the fluorination of MgO by the reaction with CCl2F2.
The time dependences of CCl2F2 conversion and products yield
for 5% MgF2–MgO at 723 K are shown in Fig. 8. Since CCl2F2
decomposition by MgF2–MgO is exclusively selective fluorine
absorption with formation of CO2 and CCl4 described above as
Eq. 2, the fluorination degree of MgO can be calculated based
on the yields of CO2 at corresponding reaction times. As de-
scribed above, there was an induction period for CCl2F2 con-
version, probably due to active acid site generation by the reac-
tion. The degree of MgO fluorination simply increased with an
increase in reaction time although the rate of the increase got
smaller when the fluorination degree approached 90%.
10% MgF2–MgO and 40% MgF2–MgO were used as repre-
sentative MgF2-MgO samples, since they showed distinct dif-
ferences in the length of induction period in CCl2F2 decompo-
sition at 623 K (Fig. 1) and NH3-TPD profile (Fig. 6). Based on
the wavenumber of the peaks from 19b and 8a modes, pyridines
on both samples were assigned to be adsorbed at Lewis acid
sites. However, a small amount of physically adsorbed pyridine
was also included at 298 K, and 373 K for both samples. The
intensity of all four peaks decreased and the wavenumbers of
the peaks from 19b and 8a modes slightly shifted to higher val-
ues as evacuation temperature rose. For example pyridine ad-
sorbed on 40% MgF2–MgO showed the peaks at 1448 cmꢁ1
(19b) and 1608 cmꢁ1 (8a) at 298 K, while at 573 K this value
shifted to 1451 cmꢁ1 (19b) and 1614 cmꢁ1 (8a). This fact indi-
cates that MgF2–MgO samples have Lewis acid sites that are
different in strength. Stronger Lewis acid sites, from which
pyridine did not desorb at higher evacuation temperature, have
been reported to show the peaks at higher wavenumber as
19b29,30 and 8a modes.21 Since the geometry of adsorbed pyri-
dine is almost the same for all the samples used, Lewis acidity
of acid sites on MgF2–MgO samples can be compared by the
wavenumber of the peaks from 19b and 8a modes. Pyridine
on 40% MgF2–MgO showed wavenumbers of the peaks from
19b and 8a modes higher than those on 10% MgF2–MgO at
any temperature employed; for example, at the evacuation tem-
perature of 573 K, the wavenumber of the peaks were 1451
cmꢁ1 (19b) and 1614 cmꢁ1 (8a) for 40% MgF2–MgO, while
they were 1449 cmꢁ1 (19b) and 1611 cmꢁ1 (8a) for 10%
MgF2–MgO. The peaks remaining at 573 K, which originated
from stronger Lewis acid sites, were more intense for 40%
MgF2–MgO than for 10% MgF2–MgO. This fact indicates that
40% MgF2–MgO has more and stronger acid sites than 10%
MgF2–MgO. 20% MgF2–MgO gave almost the same results
as for 40% MgF2–MgO, although the results are not shown
here. This indicates that the degree of fluorination over 20%
can give sufficient Lewis acid sites in both number and
strength. No spectrum obtained in this study showed any bands
at 1540 cmꢁ1 and 1640 cmꢁ1 from Brꢁnsted acid sites. The fea-
sibility of observation of Brꢁnsted acid sites on MgO by pyri-
dine adsorption is still obscure, because previously reported re-
sults have some discrepancies: Itoh et al. reported that pyridine
The result of XRD analysis at different reaction times is
shown in Fig. 9. The intensity of the diffraction peaks from
MgO and the corresponding increase in the peaks from MgF2
were observed with an increase of reaction time. No substantial
diffraction peak from MgCl2 was observed for all the samples.
Fig. 8. Time course of CCl2F2 decomposition by 5%
MgF2–MgO at 723 K. CCl2F2 conversion ( ), CO2 yield
( ), CCl4 yield ( ), CCl3F yield ( ), ratio of MgF2 in
the sample ( ).
did not interact with Brꢁnsted acid sites on MgO,25 while Lopez
´
and Gomez reported the observation of PyB modes on MgO
´
prepared by a sol–gel method hydrolyzed with hydrochloric
acid.26 In our partially fluorinated MgO samples, it might be
possible that trace amounts of residual hydroxy groups which
remain after thermal treatment at 873 K have Brꢁnsted acid
sites, which are not detectable by infrared spectroscopy using
pyridine as probe molecule. Nevertheless, such Brꢁnsted acid
sites cannot be so important since there is no water and hydro-
gen source to generate Brꢁnsted acid sites in the system em-
ployed in this study, while Brꢁnsted acidity is an important fac-
tor in hydrolysis of CCl2F2 over solid catalysts where abundant
water vapor exists in the system.
Changes of 5% MgF2–MgO during CCl2F2 Decomposi-
tion. The results shown above reveal that the reactivity to
CCl2F2, acid characteristics, and the morphology of MgO are
Fig. 9. XRD patterns of 5% MgF2–MgO during CCl2F2 de-
composition at 723 K. Reaction time: 0 h (a), 0.5 h (b), 1 h
(c), 2 h (d), 3 h (e), 5 h (f), 10 h (g). MgO ( ), MgF2 ( ).

T. Tamai et al.
Bull. Chem. Soc. Jpn., 77, No. 6 (2004) 1245
Table 3. Changes in Particle Size of MgO and MgF2 in 5%
MgF2–MgO during CCl2F2 Decomposition
Reaction time
/h
Particle size/nm
MgOaÞ
MgF2
bÞ
0
0.5
1
2
3
8.6
9.5
9.5
n.d.cÞ
n.d.cÞ
n.d.cÞ
n.d.cÞ
n.d.cÞ
n.d.cÞ
22
23
24
5
10
26
26
Reaction condition: 1% CCl2F2, 723 K, 0.2 g of 5% MgF2–
MgO. a) Determined by XRD peak of MgO (200) at
2ꢁ ¼ 42:9 degree. b) Determined by XRD peak of MgF2
(110) at 2ꢁ ¼ 27:3 degree. c) Not determined because of low
peak intensity or overlap of another peak.
Fig. 10. NH3-TPD profiles of 5% MgF2–MgO during
CCl2F2 decomposition. Reaction time: 0 h (a), 0.5 h (b),
1 h (c), 2 h (d), 3 h (e), 5 h (f), 10 h (g).
Table 4. BET Surface Area and Results of NH3-TPD for 5%
MgF2–MgO during CCl2F2 Decomposition
MgF2–MgO are almost independent of the method of fluorina-
tion of MgO but are dependent on the degree of fluorination.
However, it would be also true that some characteristics of
the acid sites are different for the MgF2–MgO samples prepared
by these two different methods, because a little amount of
chlorine could be adsorbed on MgF2–MgO formed by CCl2F2
decomposition and lateral distribution of acid sites should be
different, which may affect the characteristics of acid sites.
Generation of Acid Sites by Fluorination. The above re-
sults revealed that Lewis acid sites were formed by fluorination
of both HF and CCl2F2. Commercial MgF2 had quite small sur-
face area and poor acidity (Table 2), and crystalline MgF2 in
partially fluorinated MgO was similar to commercial MgF2 as
judged from particle size of MgF2 (Tables 1 and 3). Thus, gen-
erated acidity is not derived from MgF2, but is derived from the
interface of MgF2 and MgO. The acidity of MgF2 is reported to
be weak, which is inactive to disproportionation reactions of
chlorofluorocarbons.31–33 As described in the introduction,
there are many reports about fluorination of metal oxide and ef-
fects of fluorination, although there is no report about fluorina-
tion of MgO. The reason for increase in acidity of metal oxides
by fluorination is briefly explained as follows. Partial replace-
ment of surface oxygen and/or hydroxy groups by more elec-
tronegative fluorine increased the polarity of the metal oxide
lattice, making the nearby Brꢁnsted and Lewis acid sites more
acidic.6
Role of Acid Sites on MgF2–MgO in Total Decomposition
of CCl2F2. Partially fluorinated MgO effectively decomposed
CCl2F2 to CO2 and CCl4 with fluorine absorption as MgF2,
while the desirable reaction is the total decomposition to CO2
and solid halides. If CCl4 is completely decomposed with
chlorine fixation as MgCl2, total decomposition of CCl2F2 with
simultaneous absorption of both chlorine and fluorine is at-
tained. The reason why selective fluorine absorption exclusive-
ly takes place should be related to the reactivity of MgF2–MgO
to CCl4. In order to confirm this, we performed CCl4 decompo-
sition experiments by MgO and partially fluorinated MgO using
a pulse-reaction system. Results are shown in Fig. 11. Regard-
less of the degree of fluorination, in other words independent of
the acidity of the sample, prompt deactivation within a few
Reaction time
/h
Surface area
/m2 gꢁ1
NH3-adsorbedaÞ
mmol gꢁ1 mmol mꢁ2
0
0.5
1
2
3
179
137
113
78
58
31
86
141
140
95
78
56
0.48
1.03
1.24
1.22
1.35
1.79
2.52
5
10
15
39
Reaction condition: 1% CCl2F2, 723 K, 0.2 g of 5% MgF2–
MgO. a) Adsorbed at 373 K, desorbed from 373 to 873 K.
These observations agree with the result of gas-phase product
analysis and Eq. 2. The crystallite size of MgO was almost con-
stant at 8 to 9 nm, as shown in Table 3, while the size of MgF2
slightly increased from 22 to 26 nm after the reaction for 5 h.
This result indicates that a part of MgO is once fluorinated by
the reaction with CCl2F2 to have a higher degree of fluorina-
tion, then fluorine absorption in CCl2F2 decomposition pro-
ceeds dominantly on the MgF2–MgO with higher degree of
fluorination until MgO in the MgF2–MgO is almost completely
consumed to be MgF2, then another MgF2–MgO with low or
initial degree of fluorination starts to react with CCl2F2. In
other words, CCl2F2 decomposition with fluorine absorption
proceeds sequentially from the top layer at upstream to bottom
part in downstream.
BET surface area and the amount of adsorbed ammonia on
MgF2–MgO after different periods of reaction are listed in
Table 4. The surface area of MgF2–MgO monotonously de-
creased as the reaction proceeded. The amount of adsorbed am-
monia, i.e., the number of acid sites, increased until 0.5 h, at
which initially employed 5% MgF2–MgO changed to 13%
MgF2–MgO (Fig. 8) and a broad desorption peak of ammonia
from 450 to 750 K started to develop in NH3-TPD (Fig. 10).
This result is consistent with that for partially fluorinated
MgO prepared with HF, for example, 10% MgF2–MgO had
the largest number of acid sites and a broad desorption peak
in NH3-TPD. In this way, characteristics of acid sites on

1246 Bull. Chem. Soc. Jpn., 77, No. 6 (2004)
Enhanced Decomposition of CCl2F2 by MgF2–MgO
MgCl2 is clearly more difficult than absorption of fluorine
which can proceed until conversion of MgO reaches about
90%. Thermodynamically favored formation of MgF2 would
be a reason for it. It can also be the reason that the solid-state
ion exchange reaction between oxygen and chlorine is slower
than that between oxygen and fluorine because the ion size of
chloride is larger than that of fluoride (ion radii of O2ꢁ, Fꢁ,
and Clꢁ are 1.28, 1.18, and 1.67 A, respectively).
ꢂ
In the decomposition experiments of CCl2F2 at 623 K for
MgF2–MgO, the reactivity was greatly dependent on the degree
of fluorination. The highest conversion of CCl2F2 during the re-
action was obtained with 10% and 20% MgF2–MgO, both of
which have large specific acids. Thus, conversion should be re-
lated to the amount of acid sites. The induction period for
CCl2F2 decomposition became shorter as the degree of fluori-
nation became higher. 20% MgF2–MgO and 40% MgF2–
MgO can be considered to have similar acid strengths accord-
ing to NH3-TPD profiles (Fig. 6(d), (e)) and FT-IR analysis
of pyridine adsorption. 20% MgF2–MgO has a larger amount
of acid sites than 40% MgF2–MgO, while the surface density
of acid sites of 40% MgF2–MgO was larger than that of 20%
MgF2–MgO (Table 2); hence, the induction period should be
related to the surface density of acid sites when the strengths
of acid sites of the samples are comparable. For initiation of
CCl2F2 decomposition by MgF2–MgO, strong Lewis acid sites
with high density should be effective, since the CCl2F2 mole-
cule has an electron distribution localized at the fluorine atoms
due to its high electronegativity to have fluorine atoms some-
what negatively charged. Partially fluorinated magnesium ox-
ides have the coordinatively unsaturated Mg sites with neigh-
boring oxygen, which are generated by surrounding fluorides
depending on the degree of fluorination. Because the coordina-
tively unsaturated Mg sites behave as Lewis acids, MgF2–MgO
with high degree of fluorination reacts with CCl2F2 more easily
than those with lower degrees of fluorination. From a mechanis-
tic point of view, the first step of CCl2F2 decomposition will be
dissociative chemisorption of fluorine to form CCl2 from
CCl2F2 on the Lewis acid sites. Then extraction of O2ꢁ ion
from the MgO lattice occurs in succession to form phosgene,
followed by conversion of COCl2 to CCl4 and CO2 via two
ways: the one is a disproportionation of COCl2, the other is
the dissociative chemisorption of COCl2 onto MgO to form
CO2 and magnesium chloride, and then halogen exchange be-
tween MgCl2 and CCl2F2 takes place to form CCl4 and MgF2.
Formed CCl4 are not decomposed downstream due to the low
reactivity of MgF2–MgO to CCl4 as mentioned above. As
shown in Table 4, by procession of the reaction, the surface
density of fluoride becomes higher to create stronger Lewis
acid sites with higher surface density. Conversion of CCl2F2 di-
minished as the amount of acidic Mg sites with neighboring
oxygen decreased. In this way, selective fluorine absorption
takes place accompanied by formation of CCl4 and CO2 in
equimolar amounts by MgF2–MgO with moderately dense
and strong Lewis acid sites.
Fig. 11. CCl4 decomposition by MgO ( ), 5% MgF2–MgO
(
), and 10% MgF2–MgO ( ) at 723 K.
pulses of CCl4 injection was observed, though at the first pulse
of CCl4 the conversion was as high as 89, 67, and 70% for 10%
MgF2–MgO, 5% MgF2–MgO, and MgO, respectively. 10%
MgF2–MgO with a high specific surface area of 233 m2 gꢁ1
showed slightly higher reactivity to CCl4 than the other two.
The ratio of surface-exposed MgO to MgO in the bulk was cal-
culated to be 14.5%, based on the specific surface area and lat-
tice constant of 4.21 A for MgO.34 The final conversion of MgO
ꢂ
calculated by the cumulative amount of CO2 formed by 10
pulses of CCl4 was about 3% showing that just one-fifth of sur-
face-exposed MgO reacted with CCl4. It is also found by BET
measurement and XRD analysis that, even with 1 pulse of
CCl4, serious sintering of MgO crystallites occurred at the sur-
face. From these results, we conclude that the acidity of MgO
samples is not correlated to the reactivity to CCl4. Thus, most
of the CCl4 formed by CCl2F2 decomposition cannot be de-
composed by partially fluorinated MgO, regardless of the de-
gree of fluorination.
Weckhuysen et al. compared the reactivity of various alka-
line earth metal oxides (MgO, CaO, SrO, and BaO) to CCl4
and showed that the order in the reactivity to CCl4 was as fol-
lows: BaO > SrO > CaO > MgO.10 This order is almost the
same as that for the basicity and probably related to the size
of the alkaline earth cation, while the acidity of the samples
is quite important for the CCl2F2 decomposition in this study,
as mentioned above. Decomposition of CCl4 and CCl2F2 are
clearly different in the reaction mechanism or the pathway to
final products. Generally speaking, catalytic decomposition of
CCl2F2 is more difficult than that of CCl4 since C–F bonding
is stronger than C–Cl bonding.2,5 Although the inclination
seems to be in the opposite direction in decomposition with hal-
ogen fixation by MgO-based materials, dissociation of halo-
gen–carbon bonding must be easier for CCl4 than for CCl2F2
in the decomposition studied here. This can be seen in the ex-
perimental fact that initial CCl4 conversion with MgO was
high, as shown in Fig. 11 while for CCl2F2 decomposition at
the same temperature of 723 K, MgO didn’t show any reactivity
at all (Fig. 4).
In this reaction, dissociation of halogen–carbon bonding is
important for initiation of the reaction, however, the halogen
fixation step (formation of halides) and the solid-state halogen
exchange reaction to regenerate the reactive sites is rather im-
portant for continuous conversion. Absorption of chlorine as
As the other byproduct, a small amount of CCl3F was ob-
served, as shown in Fig. 8. In this case, formation of CCl3F be-
came maximum around 80% of MgF2 content. Moreover, obvi-
ous CCl3F formation was observed under circumstances that
showed the difficulty in C–F bond dissociation: CCl2F2 decom-

T. Tamai et al.
Bull. Chem. Soc. Jpn., 77, No. 6 (2004) 1247
2
Y. Takita, H. Wakamatsu, M. Tokumaru, H. Nishiguchi, M.
position by 10% MgF2–MgO at 573 K and CCl2F2 decomposi-
tion by 5% MgF2–MgO at 623 K. Those conditions of low re-
action temperature and weak acidity of the sample would cause
incomplete dissociation of C–F bonds to form CCl2F inter-
mediates, which would combine with Cl to form CCl3F. In gen-
eral, decomposition of CCl3F is easier than that of CCl2F2 due
to lower binding energy of both C–Cl and C–F bonding; then
further halogen exchange to form CCl4 would easily occur.
This would be the reason for small amount of CCl3F formation
and exclusive formation of CCl4 as final product, if there are
sufficient active sites. The equation for selective fluorine ad-
sorption and formation of CCl3F can be described as Eq. 3.
The lack of CClF3 as a product definitely shows that dispropor-
tionation of CCl2F2 to form CCl3F and CClF3 does not occur.
Ito, and T. Ishihara, Appl. Catal., A, 194, 55 (2000).
Y. Takita, M. Ninomiya, R. Matsuzaki, H. Wakamatsu, H.
Nishiguchi, and T. Ishihara, Phys. Chem. Chem. Phys., 1, 2367
3
(1999).
4
(2001).
5
X. Deng, Z. Ma, Y. Yue, and Z. GaO, J. Catal., 204, 200
M. Tajima, M. Niwa, Y. Fujii, Y. Koinuma, R. Aizawa, S.
Kushiyama, S. Kobayashi, K. Mizuno, and H. Ohuchi, Appl.
Catal., B, 9, 167 (1996).
6
7
8
S. Karmakar and H. L. Greene, J. Catal., 151, 394 (1995).
S. Karmakar and H. L. Greene, J. Catal., 148, 524 (1994).
X. Fu, W. A. Zeltner, Q. Yang, and M. A. Anderson, J.
Catal., 168, 482 (1997).
9
B. M. Weckhuysen, M. P. Rosynek, and J. H. Lunsford,
Phys. Chem. Chem. Phys., 1, 3157 (1999).
10 B. M. Weckhuysen, G. Mestl, M. P. Rosynek, T. R.
Klawietz, J. F. Haw, and J. H. Lunsford, J. Phys. Chem. B, 102,
3773 (1998).
11 K. J. Klabunde, J. Stark, O. Koper, C. Mohs, D. G. Park, S.
Decker, Y. Jiang, I. Lagadic, and D. Zhang, J. Phys. Chem., 100,
12142 (1996).
12 Y. Jiang, S. Decker, C. Mohs, and K. J. Klabunde, J. Catal.,
180, 24 (1998).
13 K. J. Klabunde, A. Khaleel, and D. Park, High Temp. Mater.
Sci., 33, 99 (1995).
3CCl2F2(g) þ 2MgO(s) ! CO2(g) þ 2CCl3F(g) þ 2MgF2(s)
ꢀH0 ¼ ꢁ532 kJ molꢁ1
:
ð3Þ
The mechanism of CCl2F2 decomposition and the pathway
for product formation should become independent of reaction
temperature between 623 and 773 K and the degree of fluorina-
tion higher than 5%, since in all such cases the reactivity to
CCl4 was quite low. Decomposition of CCl4 and fixation of
chlorine can be attained by using other alkaline earth metal ox-
ides such as CaO, SrO, or BaO. Thus, formation of CCl4 as
product is not a crucial problem for practical application of this
method. Combination of the systems using partially fluorinated
MgO and using other alkaline earth metal oxides can allow both
chlorine and fluorine of CCl2F2 to be separately fixed.
14 S. P. Decker, J. S. Klabunde, A. Khaleel, and K. J.
Klabunde, Environ. Sci. Technol., 36, 762 (2002).
15 O. Koper, I. Lagadic, and K. J. Klabunde, Chem. Mater., 9,
838 (1997).
16 O. Koper, E. A. Wovchko, J. A. Glass, J. T. Yates, and K. J.
Klabunde, Langmuir, 11, 2054 (1995).
Conclusions
Partially fluorinated magnesium oxides (MgF2–MgO) were
found to be highly reactive in CCl2F2 decomposition with se-
lective fluorine absorption as MgF2. The reactivity was related
to the degree of fluorination of MgO. The amount and strength
of acid sites on MgF2–MgO analyzed by NH3-TPD drastically
changed depending on the degree of fluorination. Pyridine ad-
sorption experiments using FT-IR to characterize acid sites
for MgO and partially fluorinated MgO indicated that partially
fluorinated MgO had Lewis acid sites and no Brꢁnsted acid
sites. By increasing the degree of fluorination, Lewis acidity be-
came slightly stronger, as suggested by slight shift in the ab-
sorption bands by ring vibration modes 8a and 19b of pyridine.
Thus, the initial stage of CCl2F2 decomposition was considered
to occur on Lewis acid sites to dissociate C–F bonds. Mg–F
bonding was subsequently formed after the dissociation of C–
F in CCl2F2 and highly electronegative fluorine atoms generat-
ed coordinatively unsaturated Mg sites at their surroundings.
As a result, Lewis acid sites are also generated by partially
fluorination of MgO by CCl2F2, so CCl2F2 decomposition with
fluorine fixation proceeds at high efficiency. CCl4 decomposi-
tion by MgO and partially fluorinated MgO proceeded very
slowly. This low reactivity to CCl4 was considered to be a dom-
inant factor for selective MgF2 formation in CCl2F2 decompo-
sition by MgO and MgF2–MgO.
17 J. Burdenic and R. H. Crabtree, Science, 27, 340 (1996).
18 M. C. Lee and W. Choi, Environ. Sci. Technol., 36, 1367
(2002).
19 D. P. Dufaux and M. R. Zachariah, Environ. Sci. Technol.,
3, 2223 (1997).
20 T. Tamai, K. Inazu, and K. Aika, Chem. Lett., 32, 436
(2003).
21 A. Hess and E. Kemnitz, J. Catal., 149, 449 (1994).
22 A. Alonso, A. Morato, F. Medina, Y. Cesteros, P. Salagre,
and J. E. Sueiras, Appl. Catal., B, 40, 259 (2003).
23 A. Hess, E. Kemnitz, A. Lippitz, W. E. S. Unger, and D. H.
Menz, J. Catal., 148, 270 (1994).
24 M. Shimbo, Z. Nakagawa, Y. Ohya, and K. Hamano, J.
Ceram. Soc. Jpn., 97, 640 (1989).
25 H. Itoh, S. Utanapanya, J. V. Stark, K. J. Klabunde, and J.
R. Shulup, Chem. Mater., 5, 71 (1993).
´
´
26 T. Lopez and R. Gomez, J. Sol-Gel Sci. Technol., 13, 1043
(1998).
27 A. H. Kline and J. Turkevich, J. Chem. Phys., 12, 300
(1944).
28 G. Connell and J. A. Dumesic, J. Catal., 101, 103 (1986).
29 H. Noller, J. A. Lercher, and H. Vinex, Mater. Chem. Phys.,
18, 577 (1998).
30 J. W. Wald, J. Catal., 10, 34 (1968).
31 S. Okazaki, Shokubai, 10, 242 (1968).
32 E. Kemnitz, Y. Zhu, and B. Adamczyk, J. Fluorine Chem.,
114, 163 (2002).
33 M. Wojciechowska, M. Zielinski, and M. Pietrowski, J.
Fluorine Chem., 120, 1 (2003).
34 Y.-X. Li and K. J. Klabunde, Langmuir, 7, 1388 (1991).
References
1 Y. Takita, G.-R. Li, R. Matsuzaki, H. Wakamatsu, H.
Nishiguchi, Y. Moro-oka, and T. Ishihara, Chem. Lett., 1997, 13.