Formation of perfluorobutene-2 epoxide in the thermal gas-phase reaction of perfluorobutene-2 with nitrogen dioxide. An experimental and DFT study

Article abstract of DOI:10.1524/zpch.2006.220.12.1595

The mechanism of the thermal reaction between perfluorobutene-2 and NO ₂ in the gas phase has been investigated in the temperature range 418.5-470.0 K. At temperatures below 432 K, equal amounts of perfluorobutene-2 epoxide (PFBE) and NO are formed. At higher temperatures, PFBE decomposes generating CF₃C(O)F and CF₃CF biradicals. The latter self-recombine reforming the parent perfluorobutene-2 molecule. The addition of O₂ does not affect the reaction course. This reaction system provides a new and promising method for PFBE preparation. Based on the perfluorobutene-2 consumption, PFBE yields larger than 90% were obtained. The energetics of the relevant reaction pathways has been calculated using various density functional theory methods. These calculations confirm the reaction mechanism based on experimental results. The computed vibrational spectra predicted for the PFBE at the B3LYP/6-311+G(3df) and B98/6-311+G(3df) levels of theory are in very good agreement with that obtained experimentally. by Oldenbourg Wissenschaftsverlag.

Full text of DOI:10.1524/zpch.2006.220.12.1595

Z. Phys. Chem. 220 (2006) 1595–1608 / DOI 10.1524/zpch.2006.220.12.1595
© by Oldenbourg Wissenschaftsverlag, München
Formation of Perfluorobutene-2 Epoxide
in the Thermal Gas-Phase Reaction of
Perfluorobutene-2 with Nitrogen Dioxide.
An Experimental and DFT Study
∗
By J. Czarnowski and C. J. Cobos
Instituto de Investigaciones Fisicoqu´ımicas Teo´ricas y Aplicadas (INIFTA),
Departamento de Qu´ımica, Facultad de Ciencias Exactas, Universidad Nacional
de La Plata, CONICET, CICPBA, Casilla de Correo 16, Sucursal 4, (1900) La Plata,
Argentina
(Received September 20, 2006; accepted September 20, 2006)
Thermal Reaction of Perfluorobutene-2 with NO2 /
Perfluorobutene-2 Epoxide / DFT Calculations
The mechanism of the thermal reaction between perfluorobutene-2 and NO2 in the gas
phase has been investigated in the temperature range 418 5–470 0 K. At temperatures
below 432 K, equal amounts of perfluorobutene-2 epoxide (PFBE) and NO are formed. At
higher temperatures, PFBE decomposes generating CF3C(O)F and CF3CF biradicals. The
latter self-recombine reforming the parent perfluorobutene-2 molecule. The addition of O2
does not affect the reaction course. This reaction system provides a new and promising
method for PFBE preparation. Based on the perfluorobutene-2 consumption, PFBE yields
larger than 90% were obtained. The energetics of the relevant reaction pathways has been
calculated using various density functional theory methods. These calculations confirm
the reaction mechanism based on experimental results. The computed vibrational spectra
predicted for the PFBE at the B3LYP 6-311+G(3df) and B98 6-311+G(3df) levels of
theory are in very good agreement with that obtained experimentally.
1. Introduction
The addition reactions of NO2 to halogenated olefins have been subject of
several studies [1–8]. These reactions are generally employed for prepara-
tive purposes, because NO2 can act either as nitrating or as oxidating agent
depending on the character of the atoms linked to the olefinic carbons. The ma-
jor products in these systems are vicinal dinitro compounds, nitrohaloacetyl
*
Corresponding author. E-mail: cobos@inifta.unlp.edu.ar

1596
J. Czarnowski and C. J. Cobos
halides, nitroketones and NOX (X = Cl or F). From a mechanistic point of
view these processes are interpreted by using free radical reaction schemes ini-
tiated by the N-attack on the olefinic C atom. As several reaction channels may
be possible, the elucidation of the mechanism of these reactions is important
for organic and inorganic chemistry.
Detailed mechanistic studies, kinetics parameters and product formation
have been reported for several gas-phase reactions of NO2 with perhalogenated
ethenes CF2CX2, where X = F, Br or Cl, at temperatures ranging from about
300 to 430 K [9–12]. A basic common mechanism for these reactions can be
postulated as follows:
•
NO2 +CF2CX2 → CF2 NO2 CX2
CF2 NO2 CX2 • +NO2 +M → CF2 NO2 CX2NO2 +M
CF2 NO2 CX2 • +NO2 → CF2 NO2 C O X +XNO
In the reaction of NO2 with CF2CF2, CF2(NO2 C(O)F and FNO are
formed [9]. The nitration of CF2CFBr leads to the formation of CF2(NO2 C(O)F,
CF2(NO2 CFBr(NO2 and BrNO, which is in equilibrium with Br2 and NO [10].
The addition of NO2 to CF2CFCl gives as final products CF2(NO2 C(O)F,
ClNO and CF2(NO2 CFCl(NO2 [11] and CF2(NO2 C(O)Cl, ClNO and
CF2(NO2 CCl2(NO2 are generated in the thermal reaction of NO2 with
CF2CCl2 [12]. The reaction of NO2 with CF3CFCF2 at 413–433 K, fol-
lows a different pattern, giving mostly perfluoropropene epoxide, C3F6O,and
NO [13], CF3CF(NO2 CF2(NO2 and nitroperfluoroacetone being also formed,
though in smaller quantities. The following mechanism was postulated for this
reaction:
CF3CF=CF2 +NO2 → CF3C•FCF2NO2
O
CF3C•FCF2NO2 → CF3CF−CF2 +NO
CF3C•FCF2NO2 +NO2 +M → CF3CF NO2 CF2NO2 +M
CF3C•FCF2NO2 +NO2 → CF3C O CF2NO2 +FNO
The absence of CF3C(O)F in this reaction suggests that C3F6O does not de-
compose appreciably below 433 K. It has been reported that, in the temperature
range 463–503 K, C3F6O decomposes unimolecularly to give CF3C(O)F and
the biradical difluorocarbene, CF2 [14, 15]. In addition to CF3C(O)F, the major
product of this pyrolysis is C2F4 produced by dimerization of CF2 biradicals.
Perfluorocyclopropane is also formed by addition of CF2 to CF2CF2.
Continuing our studies on the nitration of perhalogenated olefins, in the
present work the reaction between perfluorobutene-2, C4F8-2 and NO2, was
investigated in the temperature range 418 5–470 0 K. The experiments were

Formation of Perfluorobutene-2 Epoxide in the Thermal Gas-Phase Reaction
1597
complemented with quantum chemical calculations at various density func-
tional theory (DFT) levels to clarify relevant features of the potential energy
surface of this reaction.
2. Experimental technique
The experimental setup has been described elsewhere [10–13]. Therefore, only
a brief description is presented here. The experiments were performed meas-
uring the total pressure at constant volume and temperature in a conventional
static vacuum system free from moisture and grease. The employed spherical
quartz reaction vessel (270 cm3) is connected to a sensitive quartz spiral gauge
used as a null instrument and to a mercury manometer ( 0 02 Torr). The tem-
perature of the reaction cell was maintained constant within 0 1 K by means
of a Lauda thermostat filled with a Dow Corning 200 300 Fluid. The infrared
spectra were recorded on a Shimadzu IR-435 spectrophotometer, using a 10 cm
long Pyrex cell, fitted with NaCl windows.
All commercial reactants were purified. C4F8-2 (PCR, 97–98%), was re-
peatedly trap-to-trap distilled at low-pressure, the middle fraction being re-
tained each time. Small amounts of NO present in NO2 (Matheson 99.5%)
were eliminated by a series of freeze-pump-thaw cycles in presence of O2 until
the blue colour due to the formation of N2O3 disappeared. The degassed NO2
was purified by fractional condensation using the fraction that distilled between
213 and 243 K. O2 (La Ox´ıgena 99.9%) was bubbled through 98% analytical
grade H2SO4, passed slowly through a Pyrex glass coil maintained at 153 K
and stored in a Pyrex flask.
The experiments were carried out at 418.5, 432.4, 454.0, 463.0 and
470 0 K. The initial pressure of C4F8-2 was varied between 39.3 and 203 7 Torr,
that of NO2 between 11.1 and 210 4 Torr and that of O2 between 81.4 and
125 8 Torr.
3. Results and discussion
3.1 Reaction mechanism for C4F8-2 + NO2
At the lowest temperature studied, 418 5 K, no changes in the total pressure
were observed. However, in addition to the non-consumed reactants, the prod-
ucts perfluorobutene-2 epoxide, C4F8O (PFBE), and NO were observed in the
reaction mixture. The PFBE was identified by its infrared spectrum. The ab-
sorption band located at 1504 cm−1 and assigned to the ring-breathing mode
is very well reproduced by the quantum chemical calculations described in the
next section. This fundamental mode, characteristic of fluorinated olefin epox-
ides, is not present in other fluorinated compounds [16]. This band appears at
1551 cm−1 for perfluoropropene epoxide [13], at 1500 cm−1 for 1,1-dichloro-

1598
J. Czarnowski and C. J. Cobos
Table 1. Experimental data for the thermal reaction between C4F8-2 and NO2. ∆p indi-
cates the increase of the total pressure during the reaction time ∆t.
T
(K)
∆t
(min)
∆p
(Torr)
C4F8-2
(Torr)
NO2
(Torr)
O2 initial O2 final
PFBE CF3C(O)OF
(Torr)
(Torr)
(Torr)
(Torr)
418.5
432.4
432.4
432.4
432.4
455.0
455.0
301.3
306.0
181.5
294.0
210.0
258.7
249.8
–
2.0
–
2.2
1.0
2.4
2.5
8.2
3.5
100.0
39.9
39.3
199.3
101.6
203.6
41.9
11.1
198.5
12.1
42.4
40.0
–
–
81.4
–
–
81.2
6.7
36.6
2.5
37.1
18.0
35.6
37.6
33.9
43.0
–
4.0
–
4.0
2.0
4.8
5.0
16.4
7.0
125.8
125.6
39.1
–
–
–
–
–
–
–
–
208.8
210.4
47.6
463.0 2340.0
470.0 243.0
42.3
203.7
2,2-difluoroethene epoxide [17], at 1545 cm−1 for chlorotrifluoroethene epox-
ide [17] and at 1540 cm−1 for bromotrifluoroethene epoxide [18].
The above findings are consistent with the overall reaction
C4F8-2+NO2 → C4F8O+NO
∆n = 0
(I)
which occurs without change in the total number of moles and, therefore,
without change in total pressure. In the temperature range 432 4–470 0 K,
a significant increase of pressure was observed. Under these conditions, in add-
ition to PFBE and NO, the formation of CF3C(O)F was detected by its infrared
absorption band located at 1897 cm−1 [19, 20]. The total reaction in this tem-
perature range can be described by the stoichiometric equation:
2C4F8-2+3NO2 → C4F8O+3NO+2CF3C O F
∆n = 1
(II)
Representative experimental data are listed in Table 1. Here, ∆p denotes
the increase in total pressure corresponding to a reaction time of ∆t. According
to Eq. (II), the pressure of the formed CF3C(O)F is equal to 2∆p.
In order to quantify the amount of PBFE produced, an IR calibration curve
was constructed. For this, pure PFBE was prepared at 463 0 K by allowing to
react the mixture of 42 3 Torr of C4F8-2 and 210 4 Torr of NO2 during 39 h to
achieve the total consumption of C4F8-2. Afterwards, the gaseous mixture was
cooled to a room temperature of 298 K and then 111 1 Torr of O2 were added to
convert the formed NO into NO2. Subsequently, the reaction vessel was rapidly
cooled to liquid nitrogen temperature and O2 evacuated. The remaining reac-
tion mixture was separated by fractional condensation. The fraction volatile
at 153 K consisted of CF3C(O)F. Pure PFBE was separated between 153 and
173 K and NO2 remained as condensed residue at 173 K.
The obtained pure PFBE was employed to draw the plot depicted in Fig. 1.
The selected infrared band located at 1504 cm−1 is not interfered by other

Formation of Perfluorobutene-2 Epoxide in the Thermal Gas-Phase Reaction
1599
Fig. 1. Plot of the absorbance at 1504 cm−1 as a function of PFBE pressure. The straight
line is the result of a least-squares fit (see text).
bands. The linearity of the plot indicates that the Beer–Lambert law is clearly
obeyed under the present conditions. Therefore, a reliable absorption cross-
section for PFBE at 298 K can be determined from the slope. The obtained
value is 3 1 0 2 ×10−20 cm2 molecule−1. This small value is consistent with
the low intensity observed experimentally and predicted computationally (see
the next section) for this band. This calibration allowed us to determine the
PFBE pressures corresponding to the experiments given in Table 1.
The absence of pressure change below 432 K and the similar vapour pres-
sure values for C4F8-2 and PFBE which exclude their separation by fractional
condensation, make impracticable a kinetic study similar to those performed
for other olefins [9–13]. Therefore, the reaction mechanism was inferred from
the observed reaction products and previously related studies. The proposed
reaction scheme is supported by the detailed quantum chemical calculations
presented below.
By contrast to the nitration reactions of CF2CYX (X = Cl or Br and Y =
Cl or F) in presence of O2 [21, 22], the addition of O2 to the system C4F8-2 +
NO2 does not affect the course of this reaction, as shown in Table 1. The oxida-
tion of CF2CYX ethylenes has been explained by the following initial reaction
sequence:
CF2CYX+NO2 → CF2 NO2 CYX•
CF2 NO2 CYX• +O2 +M → CF2 NO2 CYXO2• +M
2CF2 NO2 CYXO2• → 2CF2 NO2 CYXO• +O2
CF2 NO2 CYXO• → CF2 NO2 C O Y +X

1600
J. Czarnowski and C. J. Cobos
The X atoms, adding in presence of O2 to the double bond, initiate the
chain reaction leading to the formation of the oxidation products. The ma-
jor final products for CF2CCl2 + NO2 + O2 are CF2ClC(O)Cl, F2CO and
Cl2CO [21] and those for CF2CFBr + NO2 + O2 are CF2BrC(O)F, F2CO and
FBrCO [22]. In both reactions very small amounts of CF2(NO2 C(O)Cl and
CF2(NO2 C(O)F are also formed, respectively. Therefore, the nitration of the
halogenated ethenes in the presence of O2 generates reaction products different
from those observed in the C4F8-2 NO2 O2 system.
The above considerations indicate that the CF3CF(NO2 CFCF3 radical is
not formed in the initial step or, more probably, presents a lifetime too short to
be stabilized which precludes its reaction with O2. This will be analyzed in the
next section.
It appears interesting to compare the present system with the oxidation
of C4F8-2 at 323–352 K initiated and propagated by CF3O radicals generated
by F atom abstraction from CF3OF by C4F8-2 [23]. Here, CF3OC(O)F and
CF3C(O)F are the main products and no formation of perfluorobutene-2 epox-
ide has been observed. All the experimental results were very well rationalized
by a chain reaction mechanism in which a C–C bond of the intermediate radical
CF3OCF(CF3 C(O)FCF3 is broken. In particular, the CF3C(O)F is generated in
the following way:
CF3O• +C4F8-2 → CF3OCF CF3 C•FCF3
CF3OCF CF3 C•FCF3 +O2 +M → CF3OCF CF3 C O2 •FCF3 +M
•
2CF3OCF CF3 C O2 •FCF3 → 2CF3OCF CF3 C O FCF3 +O2
•
CF3OCF CF3 C O FCF3 → CF3C O F +CF3OC•F CF3
•
The CF3OC•F(CF3 radicals then recombine with O2 to form CF3OC(O2
F(CF3 radicals which generate CF3OC(O)•F(CF3 by self-reaction. This
species undergoes unimolecular C–C bond scission to form the other reaction
product CF3OC(O)F and CF3 radicals. The latter add to O2 forming CF3O2 rad-
icals which subsequently regenerate the chain propagating species CF3O by
self-reaction.
The lack of O2 effect on the reaction C4F8-2 + NO2 shows that in this sys-
tem the CF3CF(NO2 CFCF3 radical is not stabilized. Then, due to the fact that
CF3C(O)F is also formed in absence of O2, it can be assumed that this com-
pound is formed through a reaction mechanism different from that postulated
to interpret the C4F8-2 + CF3O system.
As below 432 4 K the only product formed is perfluorobutene-2 epoxide,
the following primary step is proposed:
O
O−−NO
|
|
C4F8-2+NO2 → CF3CF−CFCF3 → CF3CF−CFCF3 +NO
(1)

Formation of Perfluorobutene-2 Epoxide in the Thermal Gas-Phase Reaction
1601
An analogous four-member ring intermediate was employed to explain the
formation of C3F6O in the thermal reaction of C3F6 with NO2 [13]. As shown
in Table 1, at temperatures above 432 K a pressure increase was measured
and the concomitant formation of CF3C(O)F is evidenced. To explain this, we
assume an unimolecular pathway in which a concerted C–C and C–O bond
cleavage generates CF3C(O)F and fluoro(perfluoromethyl)carbene, CF3CF:
O
CF3CF−CFCF3 → CF3C O F +CF3CF:
(2)
A related process has been proposed to interpret the thermal cracking of
C3F6O into CF3C(O)F and singlet CF2 [14]. Taking into account the molecu-
lar complexity of PFBE, reaction (2) is probably, under the present conditions,
very close to the high pressure limit. The CF3CF generated in reaction (2),
afterwards recombine reforming the parent C4F8-2,
CF3CF: +CF3CF: +M → C4F8-2+M
(3)
This process has been proposed in the mechanism of the pyrolysis of
CF3CF2SiF3 to give the products SiF4 and C4F8-2 above 433 K [24]. In add-
ition, this work shows that CF3CF does not add to the double bond of C4F8-2,
probably due to steric hindrance effects. This effect is corroborated by the fact
that the reactivity of FNO with C2F4 is much higher than with C4F8-2 [25]. The
latter reaction required temperatures above 523 K.
The obtained results indicate that the reaction between C4F8-2 with NO2
provides a new and simple method for PFBE preparation. An analysis of the
data of Table 1 based on consumed C4F8-2, leads to a PFBE yield larger than
90%. For comparison, a C3F6O yield of 63–89% is obtained in the reaction
of C3F6 with NO2 at 413–433 K [13]. The PFBE has been also synthesized
by passing at room temperature a stream of C4F8-2 gas through a solution
of potassium hydroxide in water and a mixture of hydrogen peroxide and
methanol [16]. More recently, a number of epoxides have been prepared with
yields ranging from 69 to 94% treating perfluorinated alkenes with M(OX)n
(M=Na, K, Ba, Ca and X=Cl, Br, being n = 1, 2) [26].
3.2 Density functional theory calculations
We present here the results obtained from a theoretical study of the mechanism
of the thermal reaction between C4F8-2 and NO2. The popular hybrid B3LYP
density functional of the DFT was used to compute fully optimized geometries
of the reactants, products, various intermediates and transition states using ana-
lytical gradient methods. Harmonic vibrational frequencies employed to make
zero-point energy corrections (ZPE) and to characterize the nature of the sta-
tionary points have been calculated via analytical second derivates methods.

1602
J. Czarnowski and C. J. Cobos
Fig. 2. Optimized geometry (in Å and degrees) for PFBE computed at the B3LYP 6-
311+G(3df) level of theory.
The intermediates were characterized by all the real frequencies, and the transi-
tion states, as confirmed by normal-mode analysis, present only one imaginary
frequency. The B3LYP approach employs the Becke’s three-parameter nonlo-
cal exchange functional [27] coupled to the nonlocal correlational functional of
Lee, Yang and Parr [28]. The standard 6-31G(d) split-valence Pople’s basis set
was employed for this calculations [29]. The derived vibrational frequencies
were scaled by a standard factor of 0.96 [30]. To improve total energies, single-
point energy calculations with the extended 6-311+G(3df) triple split va-
lence basis set [29] were then performed (B3LYP 6-311+G(3df) B3LYP 6-
31G(d)). The particular contributions of exchange and correlation functionals
that will give the best results in a given instance are difficult to predict. For
this reason, as in recent works from this laboratory [31–35], we have em-
ployed several combination of them. In this way, the recent DFT formulations
B98 [36], B97-2 [37], B1B95 [38] and O3LYP [39] were also used in the
single-point computations. As a measure of the spin contamination in these
calculations, the expectation value of the ꢀS2ꢁ operator was not greater than
0.76, close to the exact value of 0.75 for doublet states. The uncertainties in
the methods employed allow to get an error estimate of about 3 kcal mol−1 in
the energetics [40]. We cannot afford for the present fifteen heavy atoms sys-
tem more accurate and demanding computations such as provided by high level
composite ab initio model chemistries [41]. All calculations were carried out
with the Gaussian 03 program package with default integration grids [42].
Figure 2 shows the B3LYP 6-311+G(3df) optimized structure for PFBE.
Mean bond distances and angles differing respectively in only 0.007 Å and
0.15 degrees were computed at the B98 6-311+G(3df) level. For the sake of
simplicity they have not been reported here. In these calculations the 104 va-
lence electrons of PFBE were accommodated in molecular orbitals which

Formation of Perfluorobutene-2 Epoxide in the Thermal Gas-Phase Reaction
1603
Table 2. Harmonic vibrational frequencies (in cm−1 , infrared intensities (in km mol−1
and mode description for PFBE.
experimental
B3LYP 6-311+G(3df)
B98 6-311+G(3df)
assignment
frequency Irel frequency intensitya frequency intensitya
1504
1367
1321
1250
–
1217
–
1163
4
7
68
100
–
80
–
30
19
9
37
–
–
10
–
–
–
–
–
–
–
–
–
–
1507
1353
1293
1229
1211
1185
1179
1148
1124
1092
951
812
765
731
635
608
587
579
529
524
438
428
397
329
293
279
267
232
195
119
83
25 (4)
32 (5)
307 (48)
643 (100)
138 (21)
325 (51)
26 (4)
107 (17)
96 (15)
19 (3)
212 (33)
2 (0.3)
0.1 (0.02)
74 (11)
2
14
0.6
4
4
6
1
0.7
0.4
1508
1368
1303
1252
1234
1207
1201
1169
1130
1093
955
812
773
737
639
613
592
585
534
530
440
431
398
331
293
280
267
231
192
118
83
26 (4)
31 (5)
333 (53)
C2C3 stretch
C1C2 s-stretch
C1C2 a-stretch
628 (100) CF3 a-stretch
138 (22)
339 (54)
27 (4)
121 (19)
66 (11)
38 (6)
206 (33)
0.6 (0.1)
0.07 (0.01)
79 (13)
2
13
0.8
4
4
7
1
0.6
0.4
–
CF3 s-stretch
–
CF a-stretch
COC deformation
CCCC deformation
COC a-stretch
1117
1092
950
–
–
717
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
CF3 bend
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.6
0.9
4
0.7
0.5
0.007
0.7
0.5
0.2
0.7
0.9
4
0.6
0.5
0.01
0.7
0.5
0.2
–
–
–
–
–
–
–
–
–
–
–
–
–
51
51
–
–
35
0.07
36
0.07
a
Relative intensity values (Irel in parentheses.
comprise 507 basis functions based on 754 primitive Gaussians. In addition,
calculations carried out at the more modest B3LYP 6-31G(d) level (195 basis
functions and 364 primitive Gaussians) only differ in 0.003 Å and in 0.26 de-
grees from those computed at the B3LYP 6-311+G(3df) level. These small
differences between the two levels of calculations justify the single-point en-
ergy calculations given below. Unscaled harmonic vibrational frequencies,
infrared intensities and approximate mode assignations are listed in Table 2.

1604
J. Czarnowski and C. J. Cobos
Mode assignations were derived from the animation of the normal modes cor-
responding to the fundamental vibrational frequencies. Several of the modes
are significantly mixed and therefore the consigned assignations are only ap-
proximated. A comparison between the experimental frequencies and those
estimated with the B3LYP and B98 approaches led to mean absolute devia-
tions of 14 and 8 cm−1 respectively. At the same levels of theory, the deviations
in the relative infrared intensities are of only 8 and 7%. The very good agree-
ment found between the experimental and estimated parameters supports the
presence of PFBE in the system.
The total energy of the reactants C4F8-2 + NO2 and the relative ener-
gies (both with ZPE corrections) for the relevant species computed at the
different levels of theory are compiled in Table 3. Due to fact that the real
exchange-correlation functional remains unknown (Hohenberg–Kohn theo-
rem [43]), a significant difference in the relative energies is appreciable.
However, all calculations predict similar potential energy features. The cal-
culated reaction pathways on the basis of the structural evolutions are shown
in Fig. 3. Other channels are discarded because of their high endothermici-
ties. In Fig. 3 the total energy for C4F8-2 + NO2 is set as zero. The indicated
relative energies are the average values resulting from all quantum chemical
calculations listed in Table 3. The C4F8-2 + NO2 reaction starts by the for-
mation of a C4F8–NO2 transition state (TS1) which generates the nitro radical
CF3CF(NO2 CFCF3 (denoted as C4F8NO2 . This N-addition can find support
from spin distribution analysis. In fact, at the B3LYP 6-311+G(3df) level,
the spin density of NO2 radical is 0.528e on N atom and 0.236e on O atoms.
The N atom with the larger spin density should be the most reactive site. This
finding is consistent with the above discussed nitration reactions where the ni-
tro radicals are stabilized. The calculations give a transition state with a C–N
bond length of 1.979 Å and one imaginary frequency of 407i cm−1. The mean
electronic barrier for TS1 is 17 4 1 3 kcal mol−1. From this value an acti-
vation energy of Ea ∞ = ∆H0# +2RT = 19 2 kcal mol−1 can be estimated for
the NO2 addition to C4F8-2 at 450 K. For comparison, values of 7.5, 10.9,
10.8 and 10 5 kcal mol−1 have been respectively reported for the ethylenes
CF2CF2 [9], CF2CFBr [10], CF2CFCl [11] and CF2CCl2 [12]. A larger value
of 15 4 1 2 kcal mol−1 has been recently measured for the nitration of the
larger olefin C3F6 [13]. These results together with the present estimation
for C4F8-2 suggest that the activation energy of nitration reactions increases
with the number of C atoms in the olefin. After forming, the vibrationally
excited C4F8NO2 adduct can either be collisionally stabilized or more prob-
ably, due to the small exit barrier mostly attributed to the internal rotation
around the formed N–C bond (TS2), undergo an intramolecular rearrange-
ment generating the vibrationally excited four-center radical c-C4F8NO2 in-
dicated in reaction (1). This fast step is responsible for the short life of the
C4F8NO2 intermediate and justifies the lack of nitroderivated compounds in
the present system. The reaction proceeds via decomposition of the energized

Table 3. Computed total and relative energies of reactants, intermediates, transition states (TS) and products for the reaction of
C4F8 with NO2 at different DFT levels with B3LYP 6-31G(d) optimized geometries. The total energies are in units of a.u., and
the relative energies (including scaled ZPE corrections) are in units of kcal mol−1
.
species
ZPE
B3LYP
B1B95
B98
B97-2
O3LYP
6-311+G(3df)
6-311+G(3df)
6-311+G(3df)
6-311+G(3df)
6-311+G(3df)
C4F8 + NO2
C4F8NO2
c-C4F8NO2
PBFE+ NO
CF3C(O)F + CF3CF
TS1
TS2
TS3
TS4
33.5
35.4
35.8
33.6
27.7
34.1
35.3
33.4
28.4
−1156.579459
14.0
−1156.291568
12.0
−1156.178771
11.1
−1156.215818
13.1
−1156.224778
17.5
1.6
−3.0
−3.3
0.0
5.3
−19.7
−25.3
34.5
−21.6
31.6
−21.6
−17.6
31.5
34.8
35.0
17.5
17.3
19.2
31.7
17.4
15.3
21.4
33.9
15.7
14.4
17.4
31.0
17.1
16.4
21.5
35.6
19.3
21.3
23.7
38.3

1606
J. Czarnowski and C. J. Cobos
Fig. 3. Schematic representation of the potential energy surface for the C4F8-2 + NO2 re-
action. The indicated energies are average DFT values (see text).
agrees with the sequence proposed in reaction (1) to explain the mechanism
below 432 K.
The observed formation of CF3C(O)F at higher temperatures can be also
rationalized by the quantum chemical calculations. In fact, when the thermal
energy of reactants increases, higher vibrational levels of C4F8O are populated,
so that the unimolecular exit channel leading CF3C(O)F and the singlet state of
the biradical CF3CF is certainly more favoured. The efficiency of this channel
is determined by the interplay established between the collisional stabilization
of the highly vibrationally excited PFBE and intramolecular processes leading
to CF3C(O)F molecule and CF3CF: biradical. In addition, Table 3 shows that
the B3LYP, B97-2 and O3LYP functionals predict small electronic barriers of
0 2–3 3 kcal mol−1 for the reverse CF3C(O)F + CF3CF association, while the
B1B95 and B98 methods suggest a simple bond fission (barrierless) process.
Our results for the decomposition of thermalized PFBE are in agreement with
previous ones calculated at a lower level of theory. In fact, employing the pure
functional BPW91 cc-pVDZ, a barrier high of 52 1 kcal mol−1 has been re-
ported [44]. As shown in Table 3, our computed values are comprised between
51.2 and 59 8 kcal mol−1.
In summary, the present work shows that the principal products of the re-
action between C4F8-2 and NO2 below 432 K are PFBE and NO, and indicates
that CF3CF biradicals are generated by thermal decomposition of PFBE above

Formation of Perfluorobutene-2 Epoxide in the Thermal Gas-Phase Reaction
1607
this temperature. This system provides a new synthetic method that can pro-
duce PFBE with high efficiency. The thermal decomposition of substituted
epoxides to generate perfluoroalkylcarbenes is expected to be useful to pro-
duce synthetic and biological macromolecules and polymers analogously to the
difluorocarbene [14, 15]. The presented DFT calculations support the experi-
mental findings and allow predicting the energetics of the reaction pathways of
the reaction mechanism.
Acknowledgement
This research project was supported by the Universidad Nacional de La Plata,
the Consejo Nacional de Investigaciones Cient´ıficas y Te´cnicas (CONICET)
(PIP 5777), the Comisio´n de Investigaciones Cient´ıficas de la Provincia de
Buenos Aires (CICPBA) and the Max Planck Institute for Biophysical Chem-
istry Göttingen through the “Partner Group for Chlorofluorocarbons in the
Atmosphere”.
References
1. D. D. Coffman, M. S. Raasch, G. W. Rigby, P. L. Barrick, and W. E. Hanford, J. Org.
Chem. 14 (1949) 747.
2. R. N. Haszeldine, J. Chem. Soc. (1953) 2075.
3. D. A. Barr and R. N. Haszeldine, J. Chem. Soc. (1960) 1151.
4. E. R. Bissel, J. Org. Chem. 26 (1961) 5100.
5. I. L. Knunyants, A. V. Fokin, and V. A. Komarov, Zh. Vses. Khim. Obshchestva im.
D. I. Mendeleeva 7 (1962) 709.
6. I. L. Knunyants, A. V. Fokin, and V. A. Komarov, Izv. Akad. Nauk, Ser. Khim. 3
(1966) 466.
7. A. V. Fokin and A. T. Uzun, Z. Obshchei Khimii. 36 (1966) 117.
8. B. L. Dyatkin, E. P. Mochalina, and I. L. Knunyants, Russian Chem. Rev. 35 (1966)
417.
9. Ch. W. Spicer and J. Heicklen, Int. J. Chem. Kinet. 4 (1972) 575.
10. R. M. Romano and J. Czarnowski, Z. Phys. Chem. 219 (2005) 849.
11. R. M. Romano, C. O. Della Vedova, and J. Czarnowski, Z. Phys. Chem. 216 (2002)
1203.
12. J. Czarnowski and H. J. Schumacher, Int. J. Chem. Kinet. 18 (1986) 907.
13. R. M. Romano and J. Czarnowski, Z. Phys. Chem. 218 (2004) 575.
14. P. Krusic, D. Christopher Roe, and B. E. Smart, Israel J. Chem. 39 (1999) 117.
15. W. Mahler and P. R. Resnick, J. Fluor. Chem. 3 (1973 74) 451.
16. E. I. Du Pont de Nemours and Co, British Patent 904 877 (1962).
17. D. Chow, M. H. Jones, M. P. Thorne, and E. C. Wong, Can. J. Chem. 47 (1969)
2491.
18. J. Czarnowski, J. Fluor. Chem. 127 (2006) 736.
19. J. Pacansky, R. L. Waltman, and Y. Ellinger, J. Phys. Chem. 98 (1994) 4787.
20. G. D. Bent, E. Zerrad, G. W. Trucks, K. Wiberg, and L. Taing, J. Phys. Chem. 104
(2000) 370.
21. J. Czarnowski, Bull. Polish Akad. Sciences, Chem. 39 (1991) 49.

1608
J. Czarnowski and C. J. Cobos
22. V. Arce, M. dos Santos Afonso, R. M. Romano, and J. Czarnowski, J. Argent. Chem.
Soc. 93 (2005) 123.
23. R. M. Romano, C. O. Della Ve´dova, and J. Czarnowski, Int. J. Chem. Kinet. 35
(2003) 532.
24. K. G. Sharp and T. D. Coyle, Inorg. Chem. 11 (1972) 1259.
25. S. Andreades, J. Org. Chem. 27 (1962) 4163.
26. I. P. Kolenko, T. I. Filyakova, A. Y. Zapevalov, and E. P. Lure, Izvest. Akad. Nauk
SSSR, Ser. Khim. 11 (1979) 2509.
27. A. D. Becke, Phys. Rev. A 38 (1988) 3098.
28. C. Lee, W. Yang, and R. G. Parr, Phys. Rev. B 37 (1988) 785.
29. M. J. Frisch, J. A. Pople, and J. S. Binkley, J. Chem. Phys. 80 (1984) 3265 and ref-
erences therein.
30. A. P. Scott and L. Radom, J. Chem. Phys. 100 (1996) 16502.
31. C. J. Cobos, J. Mol. Struct. (Theochem) 581 (2002) 17.
32. M. E. Tucceri, M. P. Badenes, and C. J. Cobos, J. Fluor. Chem. 116 (2002) 135.
33. J. E. Sicre and C. J. Cobos, J. Mol. Struct. (Theochem) 620 (2003) 215.
34. M. P. Badenes, A. E. Croce, and C. J. Cobos, Phys. Chem Chem. Phys. 6 (2004)
747.
35. C. J. Cobos, J. Mol. Struct. (Theochem) 714 (2005) 147.
36. H. L. Schmider and A. D. Becke, J. Chem. Phys. 108 (1998) 9624.
37. P. J. Wilson, T. J. Bradley, and D. J. Tozer, J. Chem. Phys. 115 (2001) 9233.
38. A. D. Becke, J. Chem. Phys. 104 (1996) 1040.
39. W.-M. Hoe, A. J. Cohen, and N. C. Handy, Chem. Phys. Lett. 341 (2001) 319.
40. A. G. Baboul, L. A. Curtiss, P. C. Redfern, and K. Raghavachari, J. Chem. Phys. 110
(1999) 7650.
41. L. A. Curtiss, P. C. Redfern, V. Rassolov, G. Kedziora, and J. A. Pople, J. Chem.
Phys. 114 (2001) 9287.
42. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheese-
man, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam,
S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li,
J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski,
P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Za-
krzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Ko-
maromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayak-
kara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gon-
zalez, and J. A. Pople, Gaussian 03, Revision C.02, Gaussian, Inc., Pittsburgh PA
(2004).
43. P. Hohenberg and W. Kohn, Phys. Rev. 136 (1964) B864.
44. C. J. Cramer and M. Hillmyer, J. Org. Chem. 64 (1999) 4850.