Reaction between carbon dioxide and elementary fluorine

Article abstract of DOI:10.1016/j.jfluchem.2006.09.002

Reactions between carbon dioxide and fluorine were examined at temperatures of 303-523 K under various pressure and mixture ratios of both gases. Reactions were carried out similarly under the existence of NaF, CsF and EuF₃. After the reaction, fluorine was removed and the reaction products were analyzed using FT-IR, GC/FT-IR and GC/MS. The major products were CF₃OF, COF₂, CF₄ and CF₂(OF)₂. The best yield of COF₂ was 11.1% under the reaction condition of CO₂/F₂ = 76 kPa/76 kPa with temperature of 498 K for 72 h in a direct reaction. The formation rate of COF₂ in the direct reaction was estimated as 0.232 dm³ mol^-1 h^-1 under the reaction conditions of CO₂/F₂ = 76 kPa/76 kPa, at 498 K. In the presence of CsF, it was estimated as 1.88 dm³ mol^-1 h^-1 at CO₂/F₂ = 76 kPa/76 kPa at 498 K. The activation energy of the COF₂ formation in the direct reaction was estimated as 45.7 kJ mol^-1 at CO₂/F₂ = 76 kPa/76 kPa at 498 K. In addition, 24.2 and 38.9 kJ mol^-1 were evaluated at CO₂/F₂ = 76 kPa/76 kPa at 498 K, respectively, in the presence of CsF and EuF₃.

Full text of DOI:10.1016/j.jfluchem.2006.09.002

Journal of Fluorine Chemistry 128 (2007) 17–28
www.elsevier.com/locate/fluor
Reaction between carbon dioxide and elementary fluorine
b
Yasuo Hasegawa , Reiko Otani , Susumu Yonezawa , Masayuki Takashima
b
b,*
a
a
Department of Materials Science and Engineering, Faculty of Engineering, Fukui University, 3-9-1 Bunkyo, Fukui 910-8507, Japan
b
Forensic Science Laboratory, Fukui Prefectural Police Headquarters, 3-17-1 Ote, Fukui 910-8515, Japan
Received 11 July 2006; received in revised form 29 August 2006; accepted 5 September 2006
Available online 9 September 2006
Abstract
Reactions between carbon dioxide and fluorine were examined at temperatures of 303–523 K under various pressure and mixture ratios of both
gases. Reactions were carried out similarly under the existence of NaF, CsF and EuF3.
After the reaction, fluorine was removed and the reaction products were analyzed using FT-IR, GC/FT-IR and GC/MS. The major products were
CF
3
OF, COF
The best yield of COF was 11.1% under the reaction condition of CO /F = 76 kPa/76 kPa with temperature of 498 K for 72 h in a direct
2 4 2 2
, CF and CF (OF) .
2
2
2
3
ꢀ1 ꢀ1
h
reaction. The formation rate of COF in the direct reaction was estimated as 0.232 dm mol
under the reaction conditions of CO2/
/F = 76 kPa/76 kPa at 498 K.
formation in the direct reaction was estimated as 45.7 kJ mol at CO /F = 76 kPa/76 kPa at 498 K. In
addition, 24.2 and 38.9 kJ mol were evaluated at CO /F = 76 kPa/76 kPa at 498 K, respectively, in the presence of CsF and EuF .
2
3
ꢀ1 ꢀ1
F
2
= 76 kPa/76 kPa, at 498 K. In the presence of CsF, it was estimated as 1.88 dm mol
The activation energy of the COF
h at CO
2
1
2
ꢀ
2
2
2
ꢀ1
2
2
3
#
2006 Elsevier B.V. All rights reserved.
Keywords: Carbon dioxide; Fluorine gas; Carbonyl fluoride; Reaction mechanism; Catalyst
1
. Introduction
2. Results and discussion
The reaction between CO and F to form COF takes place
2
2.1. Reaction product
2
2
spontaneously even at room temperature because the Gibbs free
energychangeof the reactioniscalculated asꢀ220 kJ molꢀ [1].
1
Fig. 1 shows FT-IR spectra of the reaction products of the
direct reaction of CO and F have many peaks. The absorption
peaks corresponded to O–F, C–F and CF O bonds that appeared,
2
2
CO2 þ F2 ! COF2 þ ð1=2ÞO2 þ ðꢀDGÞ 220 kJ molꢀ
1
ꢀ
1
respectively, at 800–1000, 1200–1300, and around 1930 cm .
ꢀ1
The absorption peaks at 2360 and 3700 cm indicate the
Some reports have addressed preparation of COF [2,3] and
2
CF OF [2] from CO and F ; few reports have examined the
3
2
presence of residual CO . Some fluorocarbon compounds,
2
reaction between CO and F . Only three reports have described
2
2
including CF OF, were inferred to have been produced.
3
formation not of COF , but of CF (OF) , using CsF as a catalyst
2
2
2
Gram-Schmide spectra of the reaction product measured
using GC/FT-IR are shown in Fig. 2. They confirmed that three
products existed, with retention times of 18.95–18.96, 19.04–
[
4–6]. To confirm the conversion of CO2 into COF , the
2
reaction between CO2 and F2 has been tried at various
temperatures from 303 to 523 K under various pressure and
mixing ratios of both gases. This reaction was carried out
19.06 and 19.23–19.27 min.
The IR spectra of the products separated by GC/FT-IR are
similarly in the presence of NaF, CsF and EuF . Catalytic
3
shown in Fig. 3. IR spectra of the product sampled at the
retention times of 18.95–18.96, 19.04–19.06 and 19.23–
abilities of these metal fluorides to form COF were examined.
2
In addition, the reaction mechanism and reaction rate are
discussed in this paper.
19.27 min are also shown in Fig. 3(a)–(c). Fig. 3(a) shows
that the product with retention time of 18.95–18.96 min had a
ꢀ
1
characteristic absorption peak at 1282 cm corresponding to a
C–F bond. Similarly, the product with retention time of 19.04–
*
Corresponding author at: Faculty of Engineering, University of Fukui, 3-9-1
ꢀ
1
19.06 min had characteristic absorption peaks at 1929 cm
(CF O), 1256 cm (C–F) and 977 cm (C–F), as shown in
Bunkyo, Fukui 910-8507, Japan.
ꢀ ꢀ1
1
E-mail address: yonezawa@matse.fukui-u.ac.jp (S. Yonezawa).
0022-1139/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2006.09.002

18
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
Fig. 1. FT-IR spectrum of the product in the direct reaction. CO
Pa:76 kPa, reaction temperature 498 K, and reaction time 5 h.
2 2
:F = 76 k-
Fig. 3(b), and the product with the retention time of 19.23–
ꢀ
1
1
9.27 min had characteristic absorption peaks at 2331 cm
ꢀ ꢀ1
CO ), 2361 cm (CO ) and 3700 cm (CO ), as shown in
2 2 2
1
(
Fig. 3(c). Matching with HR Nicolet Vapor Phase Library
Thermo Nicolet Japan Inc.) revealed that the respective
products were CF [7], COF [8] and CO .
(
4
2
2
Fig. 4 shows the total ion chromatogram of the gaseous
reaction product and GC/MS spectra in the direct reaction.
Three peaks existed at the retention times of 9.02, 9.15 and
9
.45 min. The numbers in the figure indicate the m/z values.
Fig. 4(b) shows the mass chromatogram of m/z = 69, 47, 66 and
4. Ions with m/z = 69, 47, 66 and 44 corresponded,
4
Fig. 3. GC/FT-IR spectra of the products at retention times of 18.95–18.96 min
(a), 19.04–19.06 min (b) and 19.23–19.27 min (c). CO :F = 50.7 k-
+
+
+
+
2
2
respectively, to CF , COF , COF and CO2 [9]. Fig. 4(c)
3
2
Pa:101.3 kPa, reaction temperature 498 K, and reaction time 50 h.
and (d) were mass spectra of the products sampled at 9.02 and
.15 min, where the peaks attributable to N and O were
9
2
2
subtracted as a background. However, as shown in Fig. 4(c), one
ion indicated m/z = 32 because the oxygen in air mixed with the
sample injection could not be subtracted sufficiently. The ion
the result of positive CI–MS also suggested that CF [7] and
4
COF [8] were produced.
2
It is noteworthy that the ions with m/z = 85 and 54
+
+
corresponding, respectively, to CF O and OF2 were not
+
with m/z = 69, corresponding to CF3 was detected at the
retention time of 9.02 min; ions with m/z = 47 and 66
3
observed by GC/MS analysis. However, CF OF was confirmed
3
+
+
corresponding to COF and COF2 were detected at the
retention time of 9.15 min. Comparing the spectra to the data in
NIST Library, the MS spectra at 9.02 and 9.15 min correctly
corresponded, respectively, to CF [7], COF [8]. In addition,
by FT-IR spectra, as shown in Fig. 1. The characteristic
absorption peak corresponding to OF was also observed at
2
ꢀ1
860 cm . Moreover, CF OF and OF were not observed by
3
2
GC/FT-IR and GC/MS [10] because these products had
adsorbed on the column strongly and reacted with the column.
Some fluorocompounds, such as CH SiF , (CH ) SiF and
4
2
3
3
3 2
2
fluorobenzene, were detected during measurements.
Results show that CF OF, COF and CF were the major
products of the direct reaction between CO and F around
3
2
4
2
2
473 K. It is noteworthy that the reaction between CO and F
2 2
proceeded very slowly at room temperature. Confirmation of the
CF OOCF formation was notpossible inthe direct reaction. The
generated OF at the reaction, along with F , was eliminated by
3
3
2
2
vaporization at 77 K. Little OF remained in the product.
2
2.2. In the presence of metal fluorides
Reaction between CO and F in the presence of metal
2
2
fluoride was carried out under various conditions. In the
presence of NaF and EuF , the results of this reaction were
similar to that of the direct reaction. However, the FT-IR
Fig. 2. Gram-Schmide spectra of the product measured using GC/FT-IR.
CO :F = 50.7 kPa:101.3 kPa, reaction temperature 498 K, and reaction
time 50 h.
3
2
2

Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
19
Fig. 4. GC/EI–MS spectra of the products. (a) Total ion chromatogram, (b) mass chromatogram at m/z = 69, 66, 47 and 44, (c) mass spectra at the retention time of
.02 min, and (d) at the retention time of 9.15 min. CO :F = 50.7 kPa:101.3 kPa, reaction temperature 498 K, reaction time 40 h, and GC oven temperature 223 K.
9
2
2
ꢀ
3
spectrum of the reaction in the presence of CsF differed from
that in the absence of metal fluoride and the presence of EuF3
and NaF.
In the presence of 260 mg (1.65 ꢁ 10 mol) CsF at various
temperatures, CF OF and OF formed and CF (OF) did not
3
2
2
2
form, even at 303 K. The result was similar to that of the direct
reaction when the amount of CsF was small.
The FT-IR spectra of the reaction product in the presence of
ꢀ2
3
absorption profile changed between 1100 and 1300 cm
g (2 ꢁ 10 mol) CsF are shown in Fig. 5. The shape of the
Fig. 6 shows the mass chromatogram and MS spectra of the
gaseous reaction product between CO and F in the presence of
CsF. Fig. 6(a) shows that there were three peaks detected at the
ꢀ1
.
Absorption peaks at 1167 and 1202 cm newly appeared and
2
2
ꢀ
1
ꢀ
1
ꢀ1
the absorption peak at 1285 cm
shifted to 1289 cm .
retention time of 9.04, 9.17, and 9.39 min. Peaks with m/z = 69
+
corresponding to CF3 appeared in the MS spectra (Fig. 6(b)
and (d)) at the retention time of 9.04 and 9.39 min. Ions with m/
Fig. 5(b) shows that the reaction product was clearly different at
each reaction temperature. In the case reacted at 303 K, the
product displayed strong absorption peaks at 1267, 1248, 1202,
+
z = 47 and 66, corresponding, respectively, to COF and
ꢀ
1
ꢀ1
+
COF , were observed at the retention time of 9.17 min.
Especially, as shown in Fig. 6(d), the product had peaks with m/
1
186 cm , etc. The strong peak at 1202 cm became weak at
2
temperatures over 373 K. The amount of the product having
this strong peak decreased with increasing temperature. It
disappeared completely at 473 K. On the other hand, the strong
+
z = 151 and 170 corresponding to C F O and C F O .
+
2
5
2
2 6 2
Comparing the spectra to the data in the NIST Library, the
products at 9.04, 9.17 and 9.39 min corresponded, respectively,
to CF , COF and CF OOCF .
ꢀ
1
peak of 1167 cm
appeared clearly at 498 K. Strong
ꢀ1
absorption peaks at 1267, 1248, 1202, 1186 cm , etc. were
attributed to CF (OF) [11]. At high temperatures over 373 K,
the amount of CF OF in the products increased at an increasing
4
2
3
3
Fig. 7 shows the reaction temperature dependence of the
mass chromatogram of the reaction products in the presence of
CsF. Two peaks at 8.96 and 8.99 min were detected in the
profile, corresponding, respectively, to m/z = 47 and 66. The
2
2
3
ꢀ1
rate [12]. Moreover, the peak at 1167 cm appeared at 498 K
and was identified as that of CF OOCF [13,14].
3
3

20
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
after 18 days storage. It is inferred that CF (OF) decomposed
2
2
in the glass container to form SiF . It seems that one of two
4
pathways of COF formation must occur through the process of
2
the decomposition of CF (OF) [11] in the GC column; the
2
2
other pathway requires COF to be produced in the reaction
2
between CO and F . The product that had m/z = 69 at the
2
2
retention time of ca. 9.37 min was also CF OOCF . The
3
3
amount of CF OOCF that was produced at 303 K was less than
3
3
that at 498 K. Apparently, the formation of CF OOCF of more
3
3
than 498 K occurred under the low fluorine content condition. It
was unrelated with the decomposition of CF (OF) at 303 K.
Formation of CF did not proceed compared to the direct
2
2
4
reaction.
The electronic states of metal fluorides before and after the
reaction were analyzed using XPS. The color of EuF and NaF
3
changed from white to light yellow, whereas XPS profiles of the
F 1s electron of EuF and NaF did not change through the
3
reaction. The color of CsF changed from white to pink in parts
that contacted the nickel container. The color of CsF did not
change in the parts that did not contact with the container. After
a few days, the pink sample turned to yellow-green.
Fig. 9 shows XPS spectra [15] of the F 1s electron in CsF
before (a) and after (b) used as the catalyst. The peak at
6
6
83.9 eV appeared in the case of CsF before use. The peak at
85.6 eV is visible in Fig. 9(a). This peak must correspond to
the F 1s electron of the organic compound adsorbed in CsF.
This peak disappeared at 685.6 eV after 20 min of Ar ion
etching, as shown in Fig. 9(a)-2. Therefore, it seems that the
peak at 685.6 eV corresponded to some impurities on the
surface of CsF. In Fig. 9(b), the peak at 687.1 eV appeared, no
peak was detected at 684 eV before etching. The peak at
684.3 eV appeared after Ar ion etching, indicating that there
must be two types of fluoride. The CsF surface was covered by
some organic compound.
2.3. Yields
Quantitative analysis using the calibration line in Fig. 15
was carried out and the yield was calculated as a rate to CO2.
The yield of the whole gaseous reaction product after
eliminating F is shown in Table 1. In the direct reaction,
Fig. 5. FT-IR spectra of products in the presence of CsF at various tempera-
tures. CO :F = 76 kPa:76 kPa and reaction time 50 h. Sampling amount 2 kPa
a) and 0.4 kPa (b).
2
2
2
(
yields of COF , CF OF, CF and CF OOCF increased with
3
2
3
4
3
increasing reaction temperature and increasing reaction time,
as shown in Fig. 10. The reaction rate under the condition of
CO /F = 76 kPa/76 kPa was smaller than that of CO /
peak at 8.99 min disappeared at temperatures greater than
23 K. The MS spectrum for the product at 8.96 min was
determined as COF . The two peaks in Fig. 7(a) proved that
4
2
2
2
F = 50.7 kPa/101.3 kPa. However, the yield of COF of the
2 2
2
COF was produced in two different pathways.
2
reaction at CO /F = 76 kPa/76 kPa was larger than that at
2 2
The GC/MS oven temperature was reduced to 224 K from
42 K to separate the two peaks. Fig. 8(a) shows a mass
CO /F = 50.7 kPa/101.3 kPa.
2 2
2
In the presence of CsF (3 g), the yields of COF , CF OF,
2 3
chromatogram of m/z = 47, 66 and 69. The two peaks in
Fig. 7(a) were separated and two peaks with the fragment ion of
m/z = 47 and 66 were apparent at the respective retention times
of 9.16 and 9.23 min; the MS spectra of the two peaks also
CF , CF OOCF and CF (OF) were measured under the
4 3 3 2 2
conditions of CO /F = 76 kPa/76 kPa, with 50 h reaction time.
2
2
The yields, which are shown in Fig. 11, did not increase
monotonously with increasing reaction temperature. The yield
of CF (OF) decreased with increasing temperature. The yields
agreed with COF , as shown in Fig. 8(b)-2 and (b)-3. The
2
2
2
intensity ratio of the two peaks was about 1:3. This peak
intensity at the retention time of 9.23 min tended to decrease
during storage of the sample. The peak disappeared completely
of COF2 and CF OF had the highest values. In addition,
3
CF OOCF was detectable at temperatures higher than 498 K.
3
3
The yields of COF and CF OF become larger in the presence
2
3

Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
21
Fig. 6. Mass chromatogram and MS spectra of the products in the presence of CsF. (a) Mass chromatogram at m/z = 69, 66 and 47, (b) mass spectra at the retention
time of 9.04 min, (c) at the retention time of 9.17 min and (d) at the retention time of 9.17 min. CO :F = 76 kPa:76 kPa, reaction temperature 498 K, reaction time
0 h, and GC oven temperature 223 K.
2
2
5
of CsF than in the direct reaction at temperatures of 373–473 K.
Furthermore, the yields of CF OOCF appeared at the high-
temperature range over 473 K. However, CF was only slightly
seemed that the yields of COF , CF OF and CF at the direct
2 3 4
reaction increased with increasing reaction time, as shown in
Table 1.
In summary, the best yield of COF was 11.1% under the
3
3
4
observed in the presence of CsF, as shown in Fig. 11. Total
yields of the reaction products were equal, about 23%, for both
reactions at 498 K.
2
reaction condition of CO /F = 76 kPa/76 kPa, with reaction
2
2
temperature of 498 K and reaction time of 72 h. In the presence
of CsF, the best yield of COF was 6.5% under the reaction
On the other hand, in the presence of 0.25 g (1.65 ꢁ
2
ꢀ
3
1
reaction. The different yields of COF and CF OF can be
0
mol) CsF, the reaction product was similar to the direct
conditions of CO /F = 76 kPa/76 kPa, with reaction tempera-
2
2
ture of 473 K and reaction time of 50 h.
2
3
explained using the reaction mechanism.
ꢀ
3
In the presence of 0.35 g (1.65 ꢁ 10 mol) EuF under the
2.4. Reaction rate
3
condition of CO /F = 76 kPa/76 kPa at the reaction time of
2
2
5
0 h, the yields of COF , CF OF, CF and CF OOCF are
3
The reaction of COF2 generation can be written as
CO + F = COF . For the reaction rate, v is k[CO ][F ], where
2
3
4
3
shown in Fig. 12. Those yields were identical to those of the
direct reaction. It appears that the amount of EuF3 was
insufficient to hasten the reaction.
2
2
2
2
2
k is a rate constant.
The whole amount of COF produced during the reaction
2
Fig. 13 shows the yield of the direct reaction under the
condition of CO /F = 76 kPa/76 kPa at 498 K. In this case, it
process was calculated as the amount of the products that were
produced through COF pathways such as COF , CF OF, CF
4
2
2
2
2
3

22
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
2.6. Reaction mechanism
The radical reaction probably takes place, especially in the
gas phase between CO and F . The reaction mechanism will be
considered as follows:
2
2
D
ꢂ
F2ꢀ! F
(1)
(2)
ꢂ
ꢂ
CO2 þ F ! CFðOÞO
ꢂ
ꢂ
CFðOÞO þ F ! COF þ OF
(3)
2
2
ꢂ
ꢂ
OF þ F ! OF þ F
(4)
2
2
ꢂ
ꢂ
COF2 þ F ! CF3O
(5)
ꢂ
ꢂ
CF3O þ F2 ! CF3OF þ F
(6)
ꢂ
ꢂ
CF3OF þ F ! CF3 þ OF2
CF3 þ F2 ! CF4 þ Fꢂ
(7)
ꢂ
(8)
ꢂ
ꢂ
CF O þ CF O ! CF OOCF
(9)
3
3
3
3
ꢂ
ꢂ
CF O þ CF O ! CF OF þ COF
(10)
(11)
3
3
3
2
ꢂ
ꢂ
OF þ OF ! F þ O
2
2
ꢂ
The fluorine radical (F ) is produced by the dissociation of
ꢂ
F in reaction (1); subsequently, F might go on to react with
2
ꢂ
CO . In this case, the F formation is hastened by heating in the
2
initiation step (reaction (1)). The reaction (1) proceeds very
slowly at room temperature; it proceeds remarkably faster at
4
73 K.
Some subsequent steps are shown in reactions (2)–(8), and
COF , CF OF and CF are produced. These products were
Fig. 7. Mass chromatograms of the products obtained in the presence of CsF at
a) 303 K, (b) 423 K, (c) 473 K and (d) 498 K. CO :F = 76 kPa:76 kPa,
reaction time 50 h, and GC oven temperature 243 K.
(
2
2
2
3
4
detected as described above. Many other subsequent steps are
possible; they might proceed in parallel.
Similarly, the termination steps are shown in reactions (9)–
(11), and CF OF and CF OOCF are produced. Thus, this direct
3 3 3
and CF OOCF in the direct reaction. However, under the
3
reaction process, including the reactions (1)–(11), involves a
free radical chain reaction [16]. In our present experiment, it
seemed that reaction (9) proceeded only slightly.
3
existence of CsF (3 g) it was calculated as the total of COF2,
CF and CF OOCF .
The COF formation rates and rate constants are shown in
4
3
3
On the other hand, another reaction between CO and F can
2
2
2
Table 2. The COF formation rates in the presence of CsF were
2
occur. Hohorst and Shreeve report that when the reaction
temperature is near room temperature and the F amount is
five times larger than those without any catalyst.
2
sufficient under the existence of CsF, the reaction product
between CO and F is CF (OF) [12]. However, CF OF and
OF are formed when the reaction temperature rises and the F
2
.5. Activation energy of COF formation
2
2
2
2
2
3
2
2
The activation energy of COF is estimated according to the
2
amount is insufficient. Identical results were confirmed through
our experiments.
In the first step, the adsorption of F on the CsF surface must
following reaction. The relation between the rate constant k and
the temperature T is usually expressed as an Arrhenius reaction:
2
take place. The first ionization energy of Cs is 3.89 eV. In
addition, fluorine has the largest electronegativity. Therefore,
CsF is the material that has the highest ionic bond. This highly
ionic bond apparently causes the larger polarization of F2
k ¼ A eꢀ
E=RT
The value of the activation energy E was determined from
the gradient of the plots of ln k versus ꢀ1/T.
molecule adsorbed on CsF. This polarization of the F molecule
2
The values of the activation energy are shown in Table 2. The
activation energy in case of using CsF was the smallest among
all that were obtained.
might promote the reaction between CO and F .
2
2
It has also been reported that atomic fluorine [17] must be
generated easily at the CsF interface. The atomic fluorine might

Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
23
Fig. 8. Mass chromatogram of the products in the presence of CsF at the reaction temperature of 303 K. (a) Mass chromatograms at m/z = 69, 66 and 47 after (a)-1 0
days and (a)-2 4 days. (b) Mass spectra of the sample in (a)-1 having the retention time of (b)-1 9.01 min, (b)-2 9.16 min, (b)-3 9.23 min and (b)-4 9.37 min.
2 2
CO :F = 76 kPa:76 kPa, reaction time 50 h, and GC oven temperature 223 K.
Fig. 9. XPS F 1s spectrum of CsF before and after the reaction. (a) Before reaction: (a)-1 non-etching and (a)-2 after etching; (b) after reaction: (b)-1 non-etching and
b)-2 after etching.
(

24
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
Table 1
Yields of the products
Temperature
Time
(h)
Yield (%)
COF CF
(K)
2
3
OF
CF
4
CF
3
OOCF
3
2 2
CF (OF)
(
a) Direct reaction
CO = 76 kPa/76 kPa
2
/F
03
73
73
23
23
73
73
98
98
98
98
98
98
98
98
98
25
2
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
0
0.7
0.7
0.8
1.4
8.5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
5
0
0
0.1
0.1
0.1
0.5
0.2
0.3
0.1
0.6
0.3
0.7
1.1
1.9
4.6
2.6
50
5
0.8
0.5
2.6
0.2
0.8
0.4
1.3
1.3
2.5
2.2
4.5
11.1
3.4
10
17
3.5
49
2
5
8
10
20
30
40
50
60
72
50
6.6
Fig. 10. Yields of products in the direct reaction at various temperatures.
CO :F = 76 kPa:76 kPa and reaction time 50 h. (*) CF O, (~) CF OF, (&)
CF , (^) CF
9.5
2
2
2
3
13.8
16.4
20
4
3
OOCF
3
2 2
and (ꢁ) CF (OF) .
19.1
32.1
26.7
dꢀ
d+
of[CF(O)O Cs F]and[CF (OF)O Cs F][18–20]shown
dꢀ
d+
2
in the following reactions must be formed:
CO
4
2
/F
98
98
98
98
98
98
98
98
98
2
= 50.7 kPa/101.3 kPa
dꢀ
dþ
2
5
CO þ F þ CsF ! ½CFðOÞO Cs Fꢃ; CsFðCFðOÞOFÞ
2
2
4
(
12)
4
10
4
20
30
40
50
60
72
2.4
30.8
0.8
0
0
dþ
dþ
½
CFðOÞOdꢀ Cs Fꢃ þ F2 ! ½CF2ðOFÞOdꢀ Cs Fꢃ
(13)
(14)
4
4
dꢀ
dþ
4
6
51.5
45.2
43.4
1.5
1.9
1.8
0
0
0
0
0
0
½CF ðOFÞO Cs Fꢃ þ F ! CF ðOFÞ þ CsF
At temperatures higher than 303 K or under low fluorine gas
pressure, CF OF seems to form, leaving OF by reaction (15)
instead of proceeding to reaction (14):
2
2
2
2
4
4.3
5.3
4
3
2
(
b) In the presence of CsF (3 g)
CO = 76 kPa/76 kPa
2
/F
03
03
73
73
23
23
73
73
98
98
98
98
98
98
98
98
98
2
½CF2ðOFÞOdꢀ Cs Fꢃ þ CsF ! CF3OF þ OF2 þ Cs2O
dþ
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
5
50
5
0.3
0
0
0
0
0
19.4
36.6
0.6
2.7
4
0. 1
(
15)
50
5
31.5
30.1
23.3
11.1
27.4
0
0
9
5
0
0
0
Cs2O þ F2 ! CsF þ OF2 (16)
The concentration of F adsorbed on the surface must
50
5
0
0
2
decrease with increasing temperature. Therefore, it seems to
be difficult for the intermediate compound to react with
50
2
6.5
3.7
2.5
0.1
0
0.1
0.4
0.2
5
10
20
30
40
50
60
72
0
2.9
17.2
0
1.5
0
proceed to react with CO as well as fluorine radical:
2
CsF
F2ꢀ! 2F
In many reactions, the reaction intermediates should be
postulated. The XPS analyses of the CsF surface give some
information about the organic compounds as intermediates.
Fig. 11. Yields of products in the presence of CsF at various temperatures.
In the second step, it seems to form CF (OF) by passing
2
2
CO :F = 76 kPa:76 kPa and reaction time 50 h. (*) CF O, (~) CF OF, (&)
2
2
2
3
through reactions (12)–(14). The possible reaction intermediates
CF
4
, (^) CF
3
OOCF
3
2 2
and (ꢁ) CF (OF) .

Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
25
Fig. 11 shows that the yield of CF OF increased with rising
3
reaction temperature in the temperature range of 373–473 K.
The formation and stability of the possible intermediates such
dꢀ
d+
[CF(O)O Cs F],
dꢀ
d+
[CF (OF)O Cs F]
as
and
CF O Cs F] must be examined to explain the change in
2
dꢀ
d+
[
CF OF yield. This situation will be estimated from XPS
3
3
analysis of the CsF used for the reaction. The binding energy of
the F 1s electron on the CsF surface was analyzed using XPS
(Fig. 9). After the reaction, the XPS profile of F 1s electron
shows a peak at 687.1 eV that corresponds to the C–F bond. The
XPS profile after etching has a peak at 684.3 eV that
corresponds to CsF. Results showed that the organic
compounds including C–F bond were adsorbed onto the CsF
surface. These organic compounds must be intermediates.
On the other hand, it is considered that the radical [24]
Fig. 12. Yields of products in the presence of EuF
CO :F = 76 kPa:76 kPa and reaction time 50 h. (*) CF
CF , (^) CF OOCF (OF)
and (ꢁ) CF
3
at various temperatures.
2
2
2
O, (~) CF OF, (&)
3
reaction must contribute to the increased CF OF yield at
3
4
3
3
2
2
.
temperatures of 373–473 K. However, comparing Fig. 10 with
Fig. 11 at 498 K, the radical reaction must only slightly
contribute to the increase of the CF OF yield. Fig. 11 shows that
the F molecule. The possible reaction intermediate [CF (O-
2
2
3
dꢀ
d+
F)O Cs F] reacts with CsF, and CF OF must be formed by
the yields of COF and CF OF decreased over 473 K. That is,
3
3
2
reactions (15) and (21). In fact, Cs O reacts immediately with
2
they have a maximum value around 473 K. With the change of
the reaction temperature, it seems that the kinds of the reaction
intermediates change on the CsF surface. At low temperatures,
F2, as in reaction (16), and it will return to CsF. The low F2
concentration on the CsF surface induced by the low F2
pressure yields the similar result to that shown above.
Fig. 11 shows the temperature dependence of the yield of
dꢀ
d+
the possible reaction intermediates of [CF(O)O Cs F] and
dꢀ
d+
[CF (OF)O Cs F] must be formed by reactions (12)–(14);
2
COF . It seems that pathways (17)–(21) are considered to form
2
CF (OF) is formed dominantly at room temperature. The
2
2
COF and CF OF [21–23]. With rising temperature, OF would
2
COF and CF OF must be formed dominantly from the reaction
2 3
2
3
be similarly produced as follows; OF would remain in the
2
intermediates by reactions (15)–(21) at temperatures of 373–
473 K. The formation of CF OF becomes dominant. At
products at higher temperature like in case of CF OF:
3
3
temperatures greater than 498 K, the reaction through the
dꢀ
dþ
½
½
½
CFðOÞOdꢀ Cs Fꢃ þ F ! COF þ OF þ CsF
(17)
(18)
(19)
(20)
(21)
d+
possible intermediate [CF O Cs F] will become dominant.
2
2
2
3
Furthermore, some reactions exist in which the possible
dꢀ
CFðOÞOdꢀ Cs Fꢃ þ CsF ! COF2 þ Cs2O
dþ
d+
reaction intermediates of [CF(O)O Cs F], [CF (O-
2
dꢀ
d+
F)O Cs F] and [CF O Cs F] might undergo competing
dꢀ
d+
CF2ðOFÞOdꢀ Csdþ Fꢃ þ F2 ! COF2 þ OF2 þ CsF
CF2O þ CsF ! ½CF3Odꢀ Cs Fꢃ
3
reactions to give different products such as COF , CF OF, CF
2
3
4
dþ
and CF OOCF .
3 3
CF OOCF is formed by the following reaction:
3
3
dꢀ
dþ
½
CF3O Cs Fꢃ þ F2 ! CF3OF þ CsF
dþ
2
½CF3Odꢀ Cs Fꢃ ! CF3OOCF3 þ CsF
(22)
As inferred from the results shown in Fig. 11, reaction (22)
must take place significantly at temperatures greater than
73 K. Yields of CF OF might decrease because of the
4
3
dꢀ
d+
consumption of [CF O Cs F] by reaction (23). Table 1
3
shows that CF OOCF was also produced at 303 K in 0.1%.
3
3
This fact can not be explained using the reaction scheme
described above. There is probably a reaction path way through
which CF OOCF is produced from CF (OF) , which is the
3
3
2
2
major product at 303 K.
The formation of CF described by reaction (23) scarcely
occurred, as shown in Fig. 11:
4
dꢀ
dþ
½
CF3O Cs Fꢃ þ F2 ! CF4 þ O2 þ CsF
(23)
As described above, it became the possible major
dꢀ
d+
intermediate instead of [CF (OF)O Cs F] at temperatures
2
Fig. 13. Yields of products in the direct reaction at 498 K with various reaction
times. CO :F = 76 kPa:76 kPa and reaction time 50 h. (*) CF O, (~) CF OF,
&) CF , (^) CF OOCF (OF)
and (ꢁ) CF
greater than 473 K. It is easily converted to CF OF and
3
2
2
2
3
dꢀ
CF OOCF because [CF O Cs F] must react with F to
3 3 3 2
d+
(
4
3
3
2
2
.

26
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
Table 2
Formation rate of COF
2
and rate constant and activation energy of COF
2
formation
Formation rate,
Catalyst
None
Pressure (kPa)
CO
50.7
Temperature,
Rate constant,
ꢀ3
Activation
ꢀ3
ꢀ1
ꢀ1 ꢀ1
ꢀ1
T (K)
v (mol dm
h
)
k (dm mol
h
)
energy (kJ mol )
2
F
2
2 ꢁ 10ꢀ
4
1.63
101.3
76
498
373
423
473
498
ꢀ
ꢀ
5
4
7
7
7
7
6
6
6
6
0.4 ꢁ 10
0.00594
0.0859
0.208
0.232
45.7
76
0.5 ꢁ 10
ꢀ4
76
1 ꢁ 10
ꢀ4
76
1 ꢁ 10
ꢀ
ꢀ
ꢀ
6
5
4
CsF (3 g)
76
76
76
76
303
373
498
6 ꢁ 10
0.00966
0.0402
1.88
24.2
38.9
7
7
6
6
2 ꢁ 10
5 ꢁ 10
ꢀ
ꢀ
ꢀ
ꢀ
5
5
4
4
EuF
3
(0.35 g)
76
76
76
76
76
373
423
473
498
0.5 ꢁ 10
0.8 ꢁ 10
0.6 ꢁ 10
0.8 ꢁ 10
0.00629
0.00932
0.0789
0.121
7
7
7
6
6
6
COF is converted into CF OF, the best yield of COF was
2
3
2
1
7
1.1% under the reaction conditions of CO /F = 76 kPa/
2 2
6 kPa, with temperature of 498 K and 72 h reaction time in the
direct reaction. The reaction mechanism in the presence of CsF
can be explained using the formation of possible reaction
dꢀ
d+
dꢀ
d+
intermediates [CF (OF)O Cs F] and [CF O Cs F] at
2
3
each temperature. The best yield of COF was 6.5% under the
2
reaction conditions of CO /F = 76 kPa/76 kPa with 473 K
2
2
reaction temperature and 50 h reaction time.
4. Experimental
Scheme 1. Scheme of competitive reaction processes. Blanketed compounds
are the possible intermediates.
Fig. 14 shows a gas line made of SUS 316L for the reaction
between CO and F , which was originally designed to control
2
2
and monitor reaction pressures and temperatures precisely. This
apparatus has three separate parts: a gas–gas reaction line, a
trap-to-trap separation line, and a vacuum line.
generate CF . The reactions (21) and (22) proceed rapidly more
4
than reaction (23), thereby giving low CF yield.
4
The competitive reaction processes mentioned here are
summarized in Scheme 1.
The reactor was filled with F gas at 20 kPa; it was kept at
2
523 K for 1 day to passivate the inner wall of the line before use.
The F gas in the exhaust was eliminated by passing it through
2
3
. Conclusion
the activated alumina column. The reactor vessel capacities
were 59.6 ml. The amount of F gas that was fixed to the
2
3
activated alumina was about 300 dm /450 g Al O .
By the reaction between CO and F , COF was obtained at
2
2
2
2 3
high yield in conditions of 473 K during 50 h in a direct
reaction. The by-products were CF OF and CF . All of NaF,
CaF and EuF were tested as catalysts. Results showed that CsF
The CO was introduced into the reactor and the reactor was
2
cooled to 77 K (liquid N ). Then, F gas was introduced into the
2
3
4
2
reactor. The reaction temperature was controlled at 303–523 K.
Residual F in the product was eliminated by vaporization at
77 K.
3
can play a role as the catalyst. The COF formation rates in the
2
2
presence of CsF were five times larger than that without
catalyst. Using CsF, the yields of COF and CF OF become
For the reaction with the metal fluorides, the metal fluoride
(1 equiv. mol or 3 g) was first put into a nickel container. Before
use, the metal fluorides were kept all through the night in F gas
2
3
higher than that in the direct reaction for temperatures of 373–
73 K. In addition, CF (OF) was produced at 303 K and its
yield decreased with increasing temperature; CF OOCF was
4
2
2
2
at 503 K to eliminate trace amounts of water and oxides.
After mixing CO and F , the temperature at the surface of
3
3
detected more at 473 K. Yields of COF and CF OF become
2
3
2
2
higher than that in the direct reaction at temperatures of 373–
73 K. The slightly lower amount of CF less than 0.1% was
produced in this case. The CsF affected the reaction pathways.
the reactor was measured. The relation between the lapse of
time and pressure change was measured to monitor the reaction
progress. The total pressure was fixed at 152 kPa.
4
4
ꢂ
For the reaction without CsF, the F formation plays an
important role in the products’ composition. Considering that
The reaction product after eliminating F was introduced
2
into the 10 cm path length IR cell having BaF or CaF window
2
2

Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
27
Fig. 14. Apparatus for the reaction between CO
2
and F
2
. (a) Gas–gas reaction line, (b) trap-to-trap separation line and (c) vacuum line. (1) CO
gas pressure gauge, (7) F
), and (13) vacuum pump (oil rotary).
2
gas storage, (2) F
2
gas
storage, (3) reactor (SUS 316L), (4) constant volume container, (5) reactor pressure gauge, (6) CO
2
2
gas pressure gauge, (8) Pirani gauge, (9)
Ar or He gas cylinder, (10) gas sampler, (11) activated alumina, (12) trap (liquid N
2
at 2 or 0.4 kPa (FT-IR 8960 PC; Shimadzu Corp.). Moreover,
the product was analyzed by GC/FT-IR (Nexus; Thermo
Nicolet Japan Inc.) with a sample amount 20 ml, split-less and
GC/EI–MS, CI–MS: for EI (6890/5970 GC/MDS; Hewlett
Packard Co.), m/z 15–300, with a sample amount 20 ml, split-
less; for CI with methane (QP 5000A; Shimadzu Corp.), m/z
for the GC was 100 m (for GC/FT-IR, 200 m) ꢁ 0.25 mm
ꢁ 0.5 mm (DB-PETRO; J&W Scientific Inc.). The oven
temperature was controlled at 223, 243 and 303 K. Before
GC/MS analysis, helium gas was introduced into the reactor so
that the total pressure was set at 202.6 kPa.
The calibration line of the reaction product was prepared
with IR data and MS intensity.
4
5–300, with a sample amount 1 ml, split. The capillary column
ꢀ1
Fig. 15. Calibration lines of CO
2
(a), COF
2
(b), CF OF (c) and CF
3
4
(d). (b)-1 and (b)-2 correspond, respectively, to data at 1930 cm in Ref. [25] and Ref. [26]. (c)-1,
ꢀ1
ꢀ1
(c)-2 and (c)-3 correspond to data at 1282 cm in Ref. [27], at 1282 and at 945 cm measured in this study.

28
Y. Hasegawa et al. / Journal of Fluorine Chemistry 128 (2007) 17–28
Fig. 15(a) portrays a CO calibration line in which the
2
Acknowledgement
absorbance was plotted against the pressure. Into the IR cell,
CO gas was introduced at a proper pressure; the IR spectra
We gratefully acknowledge that the GC/FT-IR spectra were
acquired for us by Mr. Mamoru Komatsu of Thermo Nicolet
Japan Inc.
2
ꢀ
1
were measured at resolution of 4 cm . The absorption peak
ꢀ
1
at 2360 cm was chosen for the calibration line of CO . For
2
COF , CF OF, CF (OF) and CF OOCF [2], the relations
2
3
2
2
3
3
between the pressure and the absorbance were referred from
the literature [4,11,13,14,25–27]. The calibration line was
obtained by plotting the absorbance against the pressure.
Fig. 15(b) shows the COF2 calibration line. Line 1 in
References
[
1] J.C. Amphlett, J.R. Dacey, G.O. Pritchard, J. Phys. Chem. 75 (1971) 3024–
3026.
ꢀ
1
Fig. 15(b) was drawn using the absorbance at 1930 cm in
the literature [25]. The corrected calibration line, line 2 in
Fig. 15(b) was obtained by plotting the absorbance against
the pressure measured by Central Glass Co. Ltd. [26]. The
slopes of line 1 and line 2 were approximated, respectively,
[2] G.H. Cady, Inorg. Synth. 8 (1966) 165–170.
[
3] Kwasnik, in: G. Brauer (Ed.), 2nd ed., Handbook of Preparative Inorganic
Chemistry, vol. 1, Academic Press Inc., New York, 1963, pp. 206–208.
4] F.A. Hohorst, J.M. Shreeve, J. Am. Chem. Soc. 89 (1966) 1809–1810.
5] R.L. Cauble, G.H. Cady, J. Am. Chem. Soc. 89 (1967) 1962.
[
[
[
6] R.L. Cauble, G.H. Cady, J. Am. Chem. Soc. 89 (1967) 5161–5162.
ꢀ
1
as 1803 and 1254. Fig. 15(c) shows the CF OF calibration
3
[7] Selected data for CF
+
4
: FT-IR (CaF
+
2
2
, cm ): 1280vs; EI–MS: m/z(%): 69
+
, 10); CI–MS: m/z(%): 69 ([M-19] , 100),
(
CF
3
, 100), 50 (CF
line. Line 1 in Fig. 15(c) was drawn using the absorbance at
ꢀ
+
89(M + H , 6).
1
1
282 cm in the literature [27]. However, the sum of CO2
residual pressure and pressure of CF OF was greater than
ꢀ1
[
8] Selected data for COF
2
: FT-IR (CaF
2
, cm ): 1957vs, 1943vs, 1929vs,
3
+
256vs, 1231vs, 977s and 959s; EI–MS: m/z(%): 47 (COF , 100), 66
1
+
+
+
2
when the pressure of CF OF was estimated using line 1 in
kPa, which was the total pressure at the measurement,
(
2
COF , 60); CI–MS: 67 (M + H , 100), 47 ([M-19] , 25).
[9] L.R. Anderson, W.B. Fox, J. Am. Chem. Soc. 89 (1967) 4313–4315.
10] J.M. Huston, M.H. Studier, J. Fluorine Chem. 13 (1979) 235–249.
11] P.G. Thompson, J. Am. Chem. Soc. 89 (1966) 1811–1813.
12] F.A. Hohorst, J.M. Shreeve, Inorg. Syn. 11 (1968) 143–147.
13] A.J. Arvia, P.J. Aymonino, Spectrochim. Acta 18 (1962) 1299–1309.
14] R.S. Porter, G.H. Cady, J. Am. Chem. Soc. 79 (1957) 5628–5631.
3
[
[
[
[
[
Fig. 15(c). The corrected calibration line 2 was determined
by considering the effect of IR spectrum resolution against
line 1. The slopes of line 1 and line 2 were, respectively,
approximately 366 and 153. The pressures evaluated using
these corrected calibration lines were about 50% lower than
those by calibration lines with plotting the values in the
literature [4,11,13,14,25,27]. In this case, the resolution
greatly affected the absorption intensity. Line 3 in Fig. 15(c)
[15] NIST XPS Database, 2003.
[
16] G. Chiltz, P. Goldfinger, G. Huybrechets, G. Martens, G. Verbeke, Chem.
Rev. 63 (1963) 355–372.
[
[
[
17] G.H. Cady, Ann. Assoc. Quim. Argentina 59 (1959) 618.
18] B.S. Ault, Inorg. Chem. 21 (1982) 756–759.
ꢀ1
was obtained by plotting the absorbance at 945 cm against
the pressure. Line 3 was used instead of line 2 when the
19] L.R. Anderson, W.B. Fox, Inorg. Chem. 9 (1970) 2182–2183.
[20] L. Lawlor, J. Passmore, Inorg. Chem. 18 (1979) 2923–2924.
ꢀ
1
[
[
[
21] M. Wechsberg, G.H. Cady, J. Am. Chem. Soc. 91 (1969) 4432–4436.
22] M.E. Redwood, C.J. Willis, Can. J. Chem. 43 (1965) 1893–1898.
23] M. Lusting, A.R. Pitochelli, J.K. Ruff, J. Am. Chem. Soc. 89 (1967) 2841–
absorbance at 1282 cm
calculation.
Fig. 15(d) is the CF calibration line. This line plotted the
could not be used for the
4
2
843.
MS intensity against concentration of CF . The concentration
4
[24] R.F. Merritt, J. Org. Chem. 32 (1967) 4124–4126.
of CF and its MS intensity were measured using GC/MS with
4
[25] A.H. Nielsen, T.G. Burke, P.J.H. Woltz, E.A. Jones, J. Chem. Phys. 20
(
1952) 596–604.
the standard sample. The CF concentration was measured as a
4
[
26] Private communication with Central Glass Co. Ltd. for COF
2
: FT-IR
, cm ), (T%): 1957vs (32%), 1943vs (25%), 1929vs (23%), 1256vs
24%), 1231vs (23%), 977s (70%) and 959s (70%).
volume percentage of the GC injection gas. It was assumed
that the MS intensities of COF and CF OOCF were equal to
ꢀ1
(
CaF
2
2
3
3
(
that of CF4 to draw the calibration line of COF2 and
CF OOCF .
[27] R.T. Legemann, E.A. Jones, P.J. Woltz, J. Chem. Phys. 20 (1952) 1768–
1771.
3
3