Rate constants and C–C bond scission ratios for hydrolysis of 2,2,3-trifluoro-3-(trifluoromethyl)oxirane determined by means of a closed-circulation reactor

Article abstract of DOI:10.1016/j.jfluchem.2018.04.013

The hydrolysis rate constant of 2,2,2-trifluoro-3-(trifluoromethyl)oxirane (hexafluoropropene oxide; HFPO), a versatile precursor of fluorinated chemicals, was determined at 279–307 K, and the rate of hydrolysis was used to estimate the tropospheric lifetime of HFPO with respect to hydrolysis in clouds or uptake by the ocean. The low solubility of HFPO in water made it difficult to determine the hydrolysis rate constant because of mass-transfer limitation between the gas and liquid. A closed-circulation reactor was used to measure the rate of decrease of the partial pressure of HFPO while an HFPO-air mixture flowed over a stirred test solution under various experimental conditions. The rate of hydrolysis increased as the OH^? concentration increased in an aqueous NaOH solution but was almost independent of the H₂SO₄ concentration in aqueous H₂SO₄ solutions. Much scissioning of C–C bonds in HFPO produced carbon monoxide and trifluoroacetate in aqueous NaOH, but similar scissioning did not in water or aqueous H₂SO₄. The first-order rate constant for the pH-independent hydrolysis (k_water in s^?1), the bimolecular rate constant for the hydroxide-catalyzed hydrolysis, and the temperature dependence of these parameters were estimated by simultaneously fitting equations based on a two-film model to the time series of HFPO partial pressures under different experimental conditions. The equations included the rate constants as common parameters. The product of k_water and the Henry's law constant, K_H (M Pa^?1), at a temperature of T (K) was determined to be k_water × K_H = 3.7 × 10^?11 exp[?3300 × (T^?1 ? 1/298.2)]. The tropospheric lifetime of HFPO estimated using this equation indicates that removal of HFPO via hydrolysis in clouds is probably not a substantial sink of HFPO and suggests that, in the absence of other atmospheric sinks of HFPO, hydrolysis of HFPO in the ocean would be the major sink of HFPO.

Full text of DOI:10.1016/j.jfluchem.2018.04.013

JournalofFluorineChemistry211(2018)109–118
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Rate constants and CeC bond scission ratios for hydrolysis of 2,2,3-trifluoro-
3-(trifluoromethyl)oxirane determined by means of a closed-circulation
reactor
Shuzo Kutsuna
National Institute of Advanced Industrial Science and Technology (AIST), 16-1 Onogawa, Tsukuba, 305-8569, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords:
The hydrolysis rate constant of 2,2,2-trifluoro-3-(trifluoromethyl)oxirane (hexafluoropropene oxide; HFPO), a
versatile precursor of fluorinated chemicals, was determined at 279–307 K, and the rate of hydrolysis was used
to estimate the tropospheric lifetime of HFPO with respect to hydrolysis in clouds or uptake by the ocean. The
low solubility of HFPO in water made it difficult to determine the hydrolysis rate constant because of mass-
transfer limitation between the gas and liquid. A closed-circulation reactor was used to measure the rate of
decrease of the partial pressure of HFPO while an HFPO-air mixture flowed over a stirred test solution under
various experimental conditions. The rate of hydrolysis increased as the OH− concentration increased in an
aqueous NaOH solution but was almost independent of the H2SO4 concentration in aqueous H2SO4 solutions.
Much scissioning of CeC bonds in HFPO produced carbon monoxide and trifluoroacetate in aqueous NaOH, but
similar scissioning did not in water or aqueous H2SO4. The first-order rate constant for the pH-independent
hydrolysis (kwater in s−1), the bimolecular rate constant for the hydroxide-catalyzed hydrolysis, and the tem-
perature dependence of these parameters were estimated by simultaneously fitting equations based on a two-film
model to the time series of HFPO partial pressures under different experimental conditions. The equations in-
cluded the rate constants as common parameters. The product of kwater and the Henry’s law constant, KH (M
Pa−1), at a temperature of T (K) was determined to be kwater × KH = 3.7 × 10−11 exp[−3300 × (T−1 − 1/
298.2)]. The tropospheric lifetime of HFPO estimated using this equation indicates that removal of HFPO via
hydrolysis in clouds is probably not a substantial sink of HFPO and suggests that, in the absence of other at-
mospheric sinks of HFPO, hydrolysis of HFPO in the ocean would be the major sink of HFPO.
Fluorinated epoxide
Physicochemical property
pH-independent hydrolysis
Hydroxide-catalyzed hydrolysis
Two-film model
Atmospheric lifetime
The rate constant for this hydrolysis, khyd, must be known to esti-
1. Introduction
mate the atmospheric lifetime of HFPO with respect to hydrolysis in
clouds or uptake by the ocean, but, to the author’s knowledge, no values
of khyd are available in peer-reviewed journals. The objective of this
study was to determine khyd at ambient temperatures.
2,2,2-Trifluoro-3-(trifluoromethyl)oxirane (hexafluoropropene oxide;
HFPO; CF3CF(eOe)CF2) is a versatile precursor of fluorinated chemicals
[1]. Its epoxide functional group has special features that allow it to play
important roles in fluorinated chemical synthesis; HFPO has therefore
been widely used in fluorochemical industrial processes. Much informa-
tion about this use has accumulated, but little is known about the fate of
HFPO after it is released into the environment. There is an absence of
information about the physicochemical properties of HFPO, such as the
rate constants for its gaseous reactions with OH radicals, which are re-
levant to processes involved in its removal from the atmosphere.
Hydrolysis of epoxides such as oxirane (CH2(eOe)CH2) is a long-
standing issue in the context of industrial applications and biological
activities, and has been experimentally and theoretically investigated
[3–12]. Furthermore, atmospheric hydrolysis of second-generation ep-
oxides derived from isoprene has recently been studied with respect to
environmental issues because these reactions can contribute to forma-
tion of secondary organic aerosols [13]. Experimental results suggest a
rate expression that includes three kinetically distinguishable paths for
the hydrolysis as follows [3,6]:
A potential mechanism for removing HFPO from the atmosphere is
hydrolysis in cloud droplets or uptake by the ocean, because HFPO is
known to hydrolyze in water at ambient temperature via Eq. (1) [2] :
khyd = kwater + ka [H3O+] + kb [OH−],
(2)
CF3CF(eOe)CF2 + 3H2O → CF3C(OH)2COOH + 3F− + 3H+
(1)
where kwater is the first-order rate constant for the pH-independent
E-mail address: s-kutsuna@aist.go.jp.
https://doi.org/10.1016/j.jfluchem.2018.04.013
Received 17 March 2018; Received in revised form 18 April 2018; Accepted 19 April 2018
Availableonline22April2018
0022-1139/©2018ElsevierB.V.Allrightsreserved.

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
hydrolysis; ka is the bimolecular rate constant for the acid-catalyzed
hydrolysis; and kb is the bimolecular rate constant for the hydroxide-
catalyzed hydrolysis. For oxirane, values at 298 K have been reported to
be kwater = 5.7 × 10−7 s−1
;
ka = 9 × 10−3 M−1 s−1
;
and
kb = 1 × 10−4 M−1 s−1 [8]. The hydrolysis begins with the cleavage of
a CeO bond of the epoxide ring. In the case of hydroxide-catalyzed and
pH-independent paths, hydrolysis begins with a bimolecular nucleo-
philic substitution that involves the breaking of the epoxide CeO bond
and formation of a covalent bond between the nucleophile and epoxide.
The dominant hydrolytic products are diols such as glycols without
CeC bond scissions for most epoxides.
For halogenated epoxides, hydrolysis of chlorinated ethene oxides
such as trichloroethylene oxide [10,11] and tetrachloroethylene oxide
[12] has been studied with particular emphasis on biological issues. The
values of kwater and ka for trichloroethylene oxide are more than 105
and 3 times, respectively, those for oxirane. The fact that carbon
monoxide is the main hydrolytic product from trichloroethylene oxide
and tetrachloroethylene oxide indicates that a CeC bond scission occurs
after CeO bond cleavage of the epoxide ring during hydrolysis of these
halogenated epoxides.
The relatively low solubility in water and absence of near-ultra-
violet and visible absorption of HFPO may make it difficult to de-
termine the value of khyd for HFPO. In this study, a reactor with a closed
circulation system was used to observe decreases of the partial pressure
of HFPO at various stirring speeds in test solutions with different pH
values (e.g. deionized water and aqueous NaOH). The dependence of
the rate of decrease of the partial pressure of HFPO on the stirring speed
of the test solutions suggested that the rates of gaseous HFPO hydrolysis
were limited by mass-transfer between the gas and liquid. A two-film
model [14] was therefore used to determine values of khyd by fitting
simultaneous equations, with kwater and kb as common parameters, to
the time series of HFPO partial pressures observed under different ex-
perimental conditions. Degradation products were found to differ be-
tween reactions in water and aqueous NaOH, and much CeC bond
scissioning occurred in aqueous NaOH.
Fig. 1. Time course of the residence ratio of gas-phase HFPO (Pt/P0) when the
HFPO-air mixture was allowed to flow over deionized water stirred at rates
ranging from 0 to 1200 rpm at a temperature of 295.9 K. The values of P0 were
20.6, 20.8, 19.9, 20.6, 20.0, and 20.3 Pa for each experimental run in order of
increasing stirring rates from 0 to 1200 rpm. Red lines indicate the values
calculated from the fitting procedure (Section 2.2) (For interpretation of the
references to colour in this figure legend, the reader is referred to the web
version of this article).
in Fig. 1 alone could not be used to determine the value of khyd. A
relationship such as Eq. (2) imposed further constraints on the value of
khyd, as discussed later (Section 3.2). Therefore, similar experiments
were performed for 10–50 mM aqueous NaOH solutions stirred at 800,
1000, and 1200 rpm at 295.9 K (Fig. 2, panels a, b, and d) and for
10–30 mM aqueous H2SO4 solutions stirred at 1000 rpm at 295.9 K
(Fig. 2, panel c).
The partial pressure of HFPO decreased with time according to first-
order kinetics (Eq. (3)) in each experimental run (Fig. 2). Fig. 3 plots
the k1 values obtained from the data in Fig. 2 against the nominal molar
concentrations of NaOH and H2SO4. The k1 values increased with in-
creasing concentration of NaOH but were almost independent of the
concentration of H2SO4 in the concentration range examined. Fig. 3
thus suggests that Eq. (4) applies to the hydrolysis of HFPO at NaOH
concentrations of 0–50 mM.
2. Results
2.1. Decreases of HFPO during the experimental runs at various stirring
speeds of the test solutions with different pH values
khyd = kwater + kb [OH−
]
(4)
Fig. 1 shows the residence ratio, Pt/P0, of HFPO on a logarithmic
scale as a function of time for each experimental run in which an HFPO-
air mixture flowed over deionized water at 295.9 K in the closed-cir-
culation reactor. The parameter Pt is the partial pressure of HFPO at
time t, and P0 is the initial partial pressure of HFPO. In each experi-
mental run, the deionized water was stirred at a prescribed rate that
ranged from 0 to 1200 rpm. At 60 min, the circulation route was
changed so that the gas mixture flowed over the deionized water. The
resulting increase in the total volume corresponded to a 29% decrease
in the partial pressure of HFPO and a corresponding abrupt decrease of
the residence ratio at 60 min (Fig. 1).
When the HFPO-air mixture flowed over deionized water, the par-
tial pressure of gaseous HFPO decreased with time. This observation,
combined with the detection of degradation products such as F− (de-
scribed later), clearly indicated that hydrolysis of HFPO proceeded in
the deionized water. The partial pressure of HFPO decreased with time
according to first-order kinetics:
2.2. Evaluation of hydrolysis rate constants of HFPO by fitting to the time
series of HFPO partial pressures observed under different reaction conditions
Under the experimental conditions examined, the decrease of gas-
eous HFPO was apparently limited by mass transfer between the gas
and liquid phases (Fig. 1). Using a two-film model to describe the mass
transfer process [14] together with Eq. (4) enabled me to simulate the
time series of gaseous HFPO as follows:
dPt /dt = − a1 km (KH Pt – Ct)
(5)
(6)
dCt /dt = a2 km (KH Pt – Ct) – (kwater + kb [OH−]) Ct,
where km, in dm3 s−1, is the volumetric mass transfer coefficient; KH, in
M Pa−1, is the Henry's law constant of HFPO; and Ct, in M, is the bulk
concentration of HFPO in the test solutions. Here the first-order rate
constant for the pH-independent hydrolysis of HFPO, kwater, is re-
presented in s−1; the bimolecular rate constant for the hydroxide-cat-
ln(Pt/P0) = − k1 t,
(3)
alyzed hydrolysis, kb, is represented in M−1
a2 are defined as follows:
s
−1. The constants a1 and
where k1 is the first-order rate constant for the rate of change of the
partial pressure of gaseous HFPO.
That values of k1 increased with increasing stirring speeds of the
deionized water suggested that the rate of change of gaseous HFPO was
limited by mass transfer of HFPO between the gas and liquid. The data
a1 =RTa / (10−3 VG),
(7)
(8)
a2 = 1/VL,
110

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
Fig. 2. Time series of the residence ratio of gas-phase HFPO (Pt/P0) after the HFPO-air mixture was allowed to flow over aqueous NaOH (10–50 mM, stirred at 800,
1000, or 1200 rpm) or aqueous H2SO4 (10–30 mM, stirred at 1000 rpm) at a temperature of 295.9 K. Red lines indicate the values calculated from the fitting
procedure (section 2.2) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).
0.71, as explained in Section 2.1. Eqs. (5) and (6) were used to describe
all the time series of HFPO partial pressures for deionized water and
aqueous NaOH test solutions at 295.9 K.
The fitting was performed using the parameter-fitting routine of the
FACSIMILE software (MCPA Software Ltd, UK). This parameter-fitting
routine consisted of two phases. During the first phase, a number of
simulation runs with different parameter values were carried out to find
the combination of parameter values that minimized the sum of squared
residuals. The residual, Rij, is defined as
Rij =(νij – uij)/σij,
(9)
where j corresponds to the jth point in time and i to the ith time series;
vij is the observed partial pressure of HFPO and uij is the corresponding
calculated value; σij is a weighting error. In the fitting procedure, a
constant value of σij (0.032 Pa) was used for all the data. When sa-
tisfactory convergence to a minimum sum of squared residuals had
been achieved, the second phase was started. During the second phase,
the sensitivity matrix that characterized the dependence of Rij on each
parameter at the best parameter values was recomputed. The sensitivity
matrix was used to carry out a statistical analysis of the goodness of fit
and to estimate the variances and covariances of the fitting parameters.
Occasionally some of the final parameter values differed slightly from
the values at the end of the first phase.
Fig. 3. Dependence of the rate constant k1, which was deduced from the data
shown in Fig. 2, on the concentrations of aqueous NaOH and aqueous H2SO4.
where R is the universal gas constant (8.314 J mol−1 −1); Ta is room
K
temperature (298 K); 10−3 is a conversion factor (m3 dm−3); VG is the
gas-phase volume of the closed-circulation reactor (0.962 dm3); and VL
is the volume of the test solutions (0.180 dm3). Values of km here in-
volved only mass transfer resistance in the liquid phase because of the
relatively slow decay rate of gaseous HFPO and its low solubility in
water.
With KH assigned the reported value of 9.5 × 10−9 M Pa−1 [15],
the parameter values were obtained as listed in Table 1. The obtained
values of α were 0.71–0.73, very similar to the expected value of 0.71.
Figs. 1 and 2 show the values of Pt/P0 calculated from the fitting
procedure. The values of the parameters in Table 1 reproduce the time
series of gaseous HFPO partial pressures under different conditions. The
values of km increased with stirring speed. The values of kwater and kb
were 6000 and 1000 times, respectively, the values at 298 K of oxirane
[8]. The value of kwater was half the corresponding value for tri-
chloroethylene oxide at 298 K [10].
The values of kwater and kb were determined by fitting Eqs. (5) and
(6) to all the data with kwater and kb as common parameters. In this
procedure, the ratio (α) of the partial pressure of HFPO just after the
circulation route was changed at 60 min, to P0 was also fitted as a
parameter for each experimental run. The value of α was expected to be
Table 1
Values at 295.9 K of kwater, kb, and km determined by fitting Eqs. (5) and (6) to all the time series of HFPO partial pressures.
a
kwater (s−1
)
(3.68
(1.61
500
0.04) × 10−3
0.07) × 10−1
kb (M−1 s−1
)
a
stirring speed (rpm)
0
0.3
200
1.2
800
7.5
1000
12.5
1200
20.1
km (10−4 dm3 s−1
)
0.0
0.0
3.3
0.1
0.1
0.3
0.7
a
a
Error bounds are 90% confidence intervals based on the fitting procedure.
111

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
Fig. 4. Time series of the residence ra-
tios of gas-phase HFPO (Pt/P0) when
the HFPO-air mixture was allowed to
flow over water or aqueous NaOH test
solutions stirred at 1000 rpm at each
temperature. Red lines indicate values
calculated from the fitting procedure
(For interpretation of the references to
colour in this figure legend, the reader
is referred to the web version of this
article).
2.3. Temperature dependence of the rate constants for hydrolysis of HFPO
Experimental runs similar to those shown in Fig. 2 were carried out
for deionized water and 10–50 mM aqueous NaOH stirred at 1000 rpm
and at temperatures of 278.7 K, 288.5 K, and 307.3 K (Fig. 4). The
partial pressure of HFPO decreased with time according to first-order
kinetics (Eq. (3)) in each experimental run.
In a way similar to that described in Section 2.2, Eqs. (5) and (6)
were fit to all the time series of HFPO partial pressures at each tem-
perature, with kwater, kb, and km as common parameters at each tem-
perature. Results of this procedure are shown in Table 2. The value of α
was also fitted as a parameter in each experimental run. The values of α
obtained by this procedure were almost equal to the expected value of
0.71. In this fitting, the temperature-dependence of KH (M Pa−1) was
assumed to be represented by Eq. (10) [15].
Fig. 5. Temperature-dependence of kwater and kb. The two y axes correspond to
data points with the same color. Lines represent regression Eqs. (11) and (12).
−1
KH(T) = 8.8 × 10−9 exp[3000 × (T
– 1/298.15)]
(10)
Values of Pt/P0 calculated with this fitting procedure are shown in
Fig. 4. The values of the parameters in Table 2 reproduced the time
series of gaseous HFPO partial pressures at each temperature.
Fig. 5 shows the values of kwater and kb determined by this proce-
dure (Tables 1 and 2) on a logarithmic scale against the inverse of
temperature. The plot for kwater is linear, and the values of kwater can be
expressed by the Arrhenius equation:
gave values of kb at 298.2 K and ΔEb: kb(T0) = (1.9
0.5) × 10−1
M−1 20 kJ mol−1. The large error bounds for ΔEb
s
−1; ΔEb = 61
suggest that ΔEb might vary in the temperature range 278.7–307.3 K.
−1
kb(T) = kb(T0) × exp[− ΔEb /R × (T
– T0−1)]
(12)
−1
kwater(T) = kwater(T0) × exp[− ΔEwater /R × (T
– T0−1)],
(11)
2.4. Degradation products from hydrolysis in deionized water, aqueous
H2SO4, and aqueous NaOH
where T0 is 298.2 K and ΔEwater (kJ mol−1) is the activation energy for
the pH-independent hydrolysis of HFPO. From the regression line fit to
the data in Fig. 5, the value of kwater at 298.2 K was determined to be
In the deionized water test solutions exposed to gas mixtures con-
taining HFPO, F− was detected as a degradation product via ion
chromatography. The chromatogram shows the F− peak at a retention
time of about 9 min and suggests that another species was produced
with a retention time of about 21 min (Fig. 6a). The amount of F− was
(4.2
52
0.4) × 10−3 −1, and the value of ΔEwater was estimated to be
s
7 kJ mol−1. The error bounds are the 95% confidence intervals
based on the regression equation.
The regression equation for the plot of kb with Eq. (12) likewise
Table 2
Values of kwater, kb, and km determined for the test solutions stirred at 1000 rpm at each temperature by fitting Eqs. (5) and (6) to all the time series of HFPO
partial pressures.
a
a
a
temperature (K)
kwater (s−1
)
kb (M−1 s−1
)
km (10−4 dm3 s−1
)
278.7
288.5
307.3
(0.84
(1.93
(8.17
0.01) × 10−3
0.03) × 10−3
0.14) × 10−3
(0.48
(0.66
(4.18
0.03) × 10−1
0.05) × 10−1
0.28) × 10−1
6.3
12.4
16.0
0.3
0.9
0.3
a
Error bounds are 90% confidence intervals based on the fitting procedure.
112

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
hydrolyzed in water mainly via Eq. (1). The peak in the ion chroma-
togram at ∼21 min might have been CF3C(OH)2COOH, which would
have dissociated into CF3C(O−)2COO− in the aqueous KOH solutions
used as an eluent in the ion chromatographic analysis.
In the aqueous H2SO4 test solutions (Fig. 7b), F− was detected as a
degradation product of HFPO with a yield of 2.6
0.7 per HFPO de-
graded (in accord with Eq. (1)), similar to the yield of F− in deionized
water. The extent of formation of CF3C(OH)2COOH was unclear be-
cause the ion chromatogram peak attributable to CF3C(OH)2COOH was
2−
overlapped by a large SO4
peak (Fig. 6b).
In the aqueous NaOH test solutions (Fig. 7b), F− was detected as a
degradation product of HFPO, and the amount of F− was proportional
to the decrease in the amount of HFPO. The yield was almost three
(2.7
0.1) F− per HFPO, similar to the yields observed in deionized
water and aqueous H2SO4. In contrast, the ion chromatogram peak
attributable to CF3C(OH)2COOH did not appear in ion chromatograms
of aqueous NaOH test solutions (Fig. 6c). Instead, CF3C(O)O− was de-
tected as an ion chromatogram peak at a retention time of ∼14 min.
The concentrations of CF3C(O)O− were determined by ion-exclusion
chromatography; the yield was about 1 (0.92
0.02) CF3C(O)O− per
HFPO (Fig. 7b). Furthermore, carbon monoxide was detected as a
gaseous product (Fig. 8b). Neither CF3C(O)O− nor carbon monoxide
was observed in the deionized water and aqueous H2SO4 test solutions.
Fig. 9 plots the partial pressure of carbon monoxide (PCO in Pa)
versus the decrease of the partial pressure of HFPO (−ΔPHFPO in Pa) for
all the experimental runs in aqueous NaOH test solutions at 295.9 K.
The increases of the partial pressures of carbon monoxide were pro-
portional to the decreases of the partial pressures of HFPO; the ratio of
the former to the latter (m) for each concentration of NaOH was re-
gressed in accord with Eq. (13).
Fig. 6. Ion chromatograms of test solutions exposed to an HFPO-air mixture for
about 9 h at 295.9 K with stirring at 1000 rpm: the test solutions were deionized
water (panel a), a 10 mM aqueous H2SO4 solution (panel b), and a 10 mM
aqueous NaOH solution (panel c). The test solution in panel b had been neu-
tralized with an aqueous NaOH solution before ion chromatographic analysis.
The values of m were almost constant (ca. 0.87), independent of
concentrations of NaOH and stirring speeds of the test solutions.
2.9
0.1 times the decrease in the amount of HFPO (Fig. 7a). Error
bounds are 95% confidence intervals based on the regression line fit
that passes through the origin to the corresponding data points, and so
are in this section. The signal intensity of the ion chromatogram peak at
∼21 min was also proportional to the decrease in the amount of HFPO.
No gaseous products such as carbon monoxide were observed (Fig. 8a).
These results are consistent with the scenario that HFPO was
PCO = m × (−ΔPHFPO
)
(13)
These results clearly indicate that hydrolysis of HFPO in aqueous
NaOH test solutions involved much CeC bond cleavage. Furthermore,
the proportionality of the relationship between −ΔPHFPO and PCO
(Fig. 9) suggests that hydrolysis of HFPO was a rate-determining step
Fig. 7. The amount of degradation products versus the decrease in the amount of HFPO for each experimental run in deionized water stirred at 200–1200 rpm (panel
a) and in aqueous H2SO4 stirred at 1000 rpm or aqueous NaOH solutions stirred at 800–1200 rpm (panel b). All the experimental runs were performed at 295.9 K. The
dashed lines have slopes of 1 or 3, as indicated.
113

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
Fig. 8. IR spectra of HFPO-air mixtures exposed to deionized water (panel a) and an aqueous NaOH solution (50 mM, panel b) stirred at 1000 rpm and at 295.9 K. In
panels a and b, (e) is the difference between the spectrum in (d) and 0.67 or 0.50 times the spectrum in (c), respectively.
for formation of carbon monoxide. In the aqueous NaOH test solutions,
hydrolysis of HFPO proceeded primarily via Eq. (14) instead of Eq. (1):
proceed primarily via Eq. (1) in the environment as discussed in Section
3.3.
CF3CF(eOe)CF2 + OH− + H2O → CF3COO− + CO + 3F− + 3H+
3. Discussion
(14)
A similar experiment was performed for a 4 × 10−2 mM aqueous
NaOH test solution stirred at 1000 rpm at 295.9 K to examine how
hydrolysis of HFPO (Eqs. (1) or (14)) depended on lower concentration
of OH− (Fig. S1). Scissioning of CeC bonds in HFPO produced carbon
monoxide for ca. 4 h after the HFPO-air mixture began to flow over the
aqueous NaOH test solution, but it did not in the following duration;
scheme for hydrolysis of HFPO was expected to change from Eq. (14) to
(1). This change probably arose from the decrease in OH− concentra-
tion with hydrolysis via Eq. (14). Four mole of OH− was consumed per
a mole of HFPO reacted according to Eq. (14). Fig. 10 plots the re-
lationship between −ΔPHFPO and PCO. Fig. 10 also shows the calculated
concentration of OH− involving the aforementioned decrease in the test
3.1. Potential degradation scheme of HFPO in water and aqueous NaOH
Degradation products resulting from the hydrolysis of HFPO dif-
fered distinctly between reactions carried out in deionized water and
aqueous NaOH. This difference might seem strange, because the reac-
tion rates for the hydrolysis of HFPO differed by less than a factor of
two (Fig. 3) between the deionized water and the aqueous NaOH test
solutions. A scheme to reconcile the large difference in the degradation
products with the comparatively small change of reaction rates may
include dissociation of an intermediate 5 in the aqueous NaOH test
solutions (Scheme 1).
In both deionized water and aqueous NaOH, hydrolysis of HFPO 1
(Scheme 1) could be initiated via attack by H2O or OH− as a nucleo-
philic reagent at the CF3–CF– position (secondary position) [16] fol-
lowed by formation of an intermediate of CF3C(O)C(O)F 3 by analogy
with the mechanism proposed for hydrolysis of trichloroethylene oxide
solution ([OH−
]calc). It indicates that hydrolysis of HFPO proceeded via
Eq. (1) in the OH− concentration range of less than ca. 2 × 10−5 M.
Because the OH− concentration range of less than 2 × 10−5 M corre-
sponds to less than ca. 9 of pH, hydrolysis of HFPO is expected to
Fig. 9. Increase in partial pressure of carbon monoxide versus decrease of the partial pressure of HFPO for each experimental run in aqueous NaOH stirred at 800,
1000, and 1200 rpm at 295.9 K. Orange dashed lines represent the regression using Eq. (13); orange numbers indicate the slope obtained with the regression.
114

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
CF3C(OH)2C(O)F 5 was dissociated to CF3C(OH)(O−)C(O)F 6, followed
by formation of CF3C(O)O− 7 and CO 8. The intermediate 5 was en-
tirely hydrated to 9 in deionized water, whereas dissociation of 5 to 6
probably hindered hydration of 5 to 9 in aqueous NaOH. Scheme 1 can
thus account for the experimental result that hydrolytic degradation
products of HFPO differed substantially in deionized water and aqueous
NaOH, despite the similarity of the rates of hydrolysis. Scheme 1 can
also explain why the yield of carbon monoxide was almost constant (ca.
0.87), independent of the concentration of NaOH in aqueous NaOH.
High yields of 7 and 8 in aqueous NaOH (ca. 0.9) suggest that the
branching ratio for hydration of 3 to 5 (β; 0 ≤ β ≤1) was high: β > ca.
0.9.
The observed degradation products nearly satisfied material bal-
ances for carbon and fluorine (Section 2.4); however, a small amount of
mass was missing (less than ca. 10% of the mass). A reason for this
missing mass may be formation of CF3CF2C(O)O− 14 through hydro-
lysis of CF3CF2C(O)F 13, during which an intermediate 12 underwent
an intramolecular rearrangement. This hypothetical scheme is based on
analogy with the formation of CCl3C(O)O− in the hydrolysis of tri-
chloroethylene oxide [10,11].
Fig. 10. Increase in partial pressure of carbon monoxide (left axis) and the
concentration of OH− calculated (right blue axis) versus decrease of the partial
pressure of HFPO for the experimental run in 4 × 10−2 mM aqueous NaOH
stirred at 1000 rpm at 295.9 K. The two y axes correspond to data points or a
line with the same color (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article).
3.2. Dependence of the rate constants for hydrolysis of HFPO on the
assumed values of KH
The values of the determined parameters (Tables 1 and 2) may
depend on the assumed value of KH. This dependence can be estimated
as follows. Under the experimental conditions, the partial pressure of
HFPO obeyed Eq. (3). Eq. (3) can also be written as:
dPt / dt = − k1 Pt
(3´)
Combining this Eq. (3′) with Eqs. (5) and (7) gives Eq. (15).
KH′ = Ct / Pt = KH – k1 / (a1 km) = KH − 10−3 VG k1 / (R Ta km)
(15)
Fig. S2 shows the time series of KH′ values for each experimental
run, where the values of Ct and Pt obtained by the fitting procedure
were used to calculate the values of KH′. The values of KH′ were con-
stant, except during the initial period (ca. a few tens of seconds) in each
run. This result implies that a quasi-equilibrium state was established
for mass transfer of HFPO between the gas and liquid. In that case, Eq.
(16) applies because of the requirement for material balance of HFPO.
−10−3 VG (R Ta)−1 dPt / dt = kr KH′ Pt VL
(16)
Substituting Eqs. (3′) and (15) into Eq. (16) gives Eq. (17).
10−3 k1 VG/ (R T ) = kr VL × (KH − 103 k1 VG/(R T km))
(17)
a
a
Rearranging Eq. (17) yields Eq. (18).
Scheme 1. Possible reaction scheme for hydrolysis of HFPO in deionized water
and aqueous NaOH.
1 / k1 = 10−3 VG / (KH R Ta) × (1/VL × (kwater + kb [OH−])−1
+
km−1) = 10−3 VG/(R Ta VL) × (KH kwater + KH kb [OH−])−1 + 10−3
VG/(R Ta) × (KH km)−1
(18)
[10,11]. This intermediate 3 was hydrated to either CF3C(O)C(OH)2F 4
or CF3C(OH)2C(O)F 5. The –C(OH)2F group in the intermediate 4 could
easily be transformed into –C(O)O−. No carbon monoxide 8 was
therefore produced from 4. Carbon monoxide could, however, be pro-
duced from the hydrated intermediate 5.
The degradation scheme of intermediate 5 may differ in deionized
water and aqueous NaOH. The acidic character of the OH group in
fluorinated alcohols is well known; for example, the pKa of 1,1,1-tri-
fluoroethanol is 12.8 [16]. Intermediate 5 was thus dissociated into
CF3C(OH)(O−)C(O)F 6 in aqueous NaOH but undissociated in deio-
nized water. Unless dissociated into 6, CF3C(OH)2C(O)F 5 would be
further hydrated to CF3C(OH)2C(OH)2F 9, followed by formation of
CF3C(OH)2C(O)O− 11. In contrast, in aqueous NaOH, most of
This equation implies that the fitting procedure yielded the products
KH kwater, KH kb, and KH km —not values of kwater, kb, and km in-
dividually— from the time series of partial pressures of HFPO at dif-
ferent concentrations of OH− and stirring speeds. Accordingly, the
values of kwater determined by the fitting procedure were inversely
proportional to the assumed value of KH, whereas the product kwater
×
KH obtained by the fitting was independent of the assumed values of KH
and would depend only on temperature. Likewise, the determined va-
lues of kb were inversely proportional to the assumed value of KH, but
the product of kb × KH was constant, as was the product of each value
of km and KH. Eq. (19) thus gave the value of the product kwater × KH
(s−1 M Pa−1) from Eqs. (10) and (11), regardless of the assumed value
of KH.
115

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
the semi-infinite approximation used to derive Eq. (24) is that the eddy-
hydrolysis scale depth (He) was less than the ocean mixed layer depth,
where He was defined as follows:
−1
kwater(T) KH(T) = 3.7 × 10−11 × exp[−3300 × (T
– 1/298.2)]
(19)
1/2
He = (DO / khyd
)
(25)
With KH assigned the reported values of (1.2–1.7) × 10−8 M Pa−1
in the temperature range 280–290 K (Eq. (10)) [15], the values of kwater
were calculated from Eq. (19) to be (1.1–2.3) × 10−3 s−1. The lifetime
of HFPO in water was then estimated to be 430–910 s in the tem-
perature range 280–290 K. Virtually all the HFPO dissolved in the ocean
was therefore assumed to be hydrolyzed in the ocean mixed layer: the
calculated values of He (1.3–1.9 m) were much less than the assumed
ocean mixed layer depth (80 m). Eq. (24) was therefore applicable to
HFPO. With KH assigned the value calculated with Eq. (10) [15], the
3.3. Estimate of tropospheric lifetime of HFPO with respect to hydrolysis in
clouds and uptake by the ocean
HFPO was assumed to be a well-mixed species in the troposphere
because almost no atmospheric sink has been reported in a peer-re-
viewed journal. Based on that assumption, the tropospheric lifetime of
HFPO with respect to hydrolysis in clouds, τcloud, and that with respect
to uptake by the ocean, τocean, were roughly estimated.
First, τcloud was estimated as follows. The assumed values of KH and
the values of khyd determined here suggest that hydrolysis in cloud
droplets proceeds to the equilibrium predicted by the Henry's law, that
is, without mass-transport limitation [17]. Accordingly, removal rates
second term in parenthesis in Eq. (24), (DO khyd)
−1/2, was calculated
with Eq. (19) to be (3.3–4.8) × 102 s m−1. In this calculation, khyd was
assumed to be kwater because kb [OH−] is negligible compared to kwater
of gaseous HFPO with respect to hydrolysis in cloud droplets, Rcloud
,
in the pH range of seawater. This calculation shows that the value of
−1/2
were described by Eqs. (20) and (21).
(DO khyd
)
for HFPO is an order of magnitude smaller than the first
term in parentheses in Eq. (24), rs. Values of τocean for HFPO are
therefore insensitive to khyd and depend primarily on the values of rs
and KH. The thin film resistance for mass transport across the air-sea
interface is therefore the primary control on the uptake of HFPO by the
ocean.
Rcloud = khyd-a PHFPO
(20)
(21)
khyd-a = khyd × 103 KH RT (νL/νG) / (1 + 103 KH RT (νL/νG))
where PHFPO is the partial pressure of HFPO in the atmosphere; R is the
universal gas constant; 103 is a conversion factor (dm3 m−3); and νL and
νG are the liquid volume and gas volume, respectively, in clouds. The
ratio νL/νG is the liquid water content of clouds expressed as a di-
mensionless volume fraction and is typically 3 × 10−7 to 1 × 10−6
[17]. Because 103 KHRTνL /νG ≪ 1 due to the low solubility of HFPO in
water, Eq. (21) is approximated by Eq. (22):
It was beyond the scope of the present study to determine values of
rs and KH, but by assuming that the solubility of HFPO in seawater was
about 80% of its solubility in freshwater because of salting-out effects
and that the temperature of the ocean was 286 K, τocean was calculated
to be ∼300 years from Eq. (24) with a KH of 1.3 × 10−8 M Pa−1 and rs
of 1.7 × 104 s m−1 based on values reported in the literature [15,19].
This estimate may be substantially in error because of the assumption of
single values for rs and solubility; rs is sensitive to local conditions such
as wind speed, and solubility is sensitive to temperature. Nevertheless,
this calculation suggests that the hydrolysis of HFPO occurs not in
clouds but in the ocean if no processes other than hydrolysis remove
HFPO from the atmosphere. The degradation products of HFPO in the
ocean are expected to be CF3C(OH)2COOH and F− in the pH range of
seawater (ca. 8) as shown in Fig. 10 (Section 2.4). Environmental fate of
CF3C(OH)2COOH is an issue necessary for future investigation.
khyd-a = khyd × 103 KH RT (νL/νG)
(22)
The value of τcloud is roughly estimated as follows:
−1
τcloud = khyd-a
/ f,
(23)
where f is the fraction of time that air spends within liquid water clouds
in the lower half of the troposphere [18]. From Eq. (19), the product
kwater × KH was calculated to be (1.2–1.8) × 10−11 s−1 M Pa−1 in the
temperature range 270–280 K. By assuming a νL/νG ratio of 3 × 10−7
and f of 0.15 [18], τcloud was estimated to be ∼17,000–26,000 years.
Removal via hydrolysis in clouds is therefore not expected to be a
substantial sink of HFPO.
4. Conclusions
Second, uptake of HFPO by the ocean was evaluated with the use of
a two-layer diffusion model on the assumption that diffusive transport
occurred in a serial manner across two layers of the surface water, a
surface layer with a thickness of ∼40 μm above an ∼80 m-thick, wave-
mixed, turbulent layer characterized by an eddy diffusion coefficient DO
[19]. The lifetime of HFPO in the ocean, τocean, was roughly estimated
as follows [20,21]:
In deionized water, HFPO was hydrolyzed via Eq. (1) and trans-
formed into CF3C(OH)2COOH and F−. In contrast, in 10–50 mM aqu-
eous NaOH solutions, HFPO was hydrolyzed mainly through Eq. (14):
cleavage of the CeC bond occurred, and HFPO was transformed into
CF3C(O)O− and CO along with F−. As shown in Scheme 1, the fact that
an intermediate CF3C(OH)2C(O)F 5 could be dissociated to CF3C(OH)
(O−)C(O)F 6 only in aqueous NaOH test solutions may explain the
distinctly different degradation products in deionized water and aqu-
eous NaOH.
Hydrolysis rates of HFPO increased with increasing concentrations
of NaOH in aqueous NaOH, but they were almost independent of the
H2SO4 concentration in aqueous H2SO4 in the H2SO4 concentration
range 10–30 mM. Based on reported values of KH (Eq. (10)), the first-
τocean = (rs + (DO khyd
)
−1/2) × HA / (103 KH R T fE) + τmix
(24)
where HA is the scale height of the troposphere (8000 m); 103 is a
conversion factor (dm3 −3); fE is the fractional oceanic coverage of
m
the earth’s surface (0.7); rs is the thin film surface resistance to gas
uptake at the ocean surface; and τmix represents the effective mixing
time within the troposphere (∼45 days). The first term in parenthesis
in Eq. (24) is related to the thin-film resistance to mass transport across
the air-sea interface, and the second term (in parentheses) is related to
the loss of HFPO by hydrolysis during downward diffusion in the ocean
mixed layer. Values of rs depend on turbulence at the air-sea interface
which is generated by some factors such as winds over the ocean [22].
Values of rs vary over time and space. Values of rs and DO were assumed
order rate constant for the pH-independent hydrolysis of HFPO (kwater
at 298.2 K was determined to be (4.2
0.4) × 10−3 −1, and the
7 kJ
)
s
activation energy for this hydrolysis was estimated to be 52
mol−1; the bimolecular rate constant for the hydroxide-catalyzed hy-
drolysis (aqueous reaction with OH−) of HFPO (kb) at 298.2 K was
determined to be (1.9
0.5) × 10−1 M−1 −1. Eq. (19) gave the
s
to be 1.7 × 104 s m−1 and 4 × 10−3 m2
s
−1, respectively [19].
temperature-dependent values of the product KH × kwater, regardless of
the assumed values of KH.
On the basis of Eq. (19), the tropospheric lifetime of HFPO with
respect to hydrolysis in clouds was estimated to exceed 17,000 years;
Use of Eq. (24) assumes that hydrolysis decreases the concentration
of a species to zero during downward transport of the species in the
ocean mixed layer [19,21]. In other words, a necessary condition for
116

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
therefore, removal via hydrolysis in clouds is probably not a substantial
sink of HFPO. The tropospheric lifetime of HFPO with respect to uptake
by the ocean was estimated to be some hundreds of years, although this
estimate may contain a substantial error. Hydrolysis of HFPO should
therefore proceed not in clouds but in the ocean if no processes other
than hydrolysis remove HFPO from the atmosphere. The degradation
products of HFPO in the ocean are expected to be CF3C(OH)2COOH and
F−. Environmental fate of CF3C(OH)2COOH is an issue necessary for
future investigation.
spectral resolution was 0.5 cm−1 with an acquisition of 64 scans.
At the end of the above period, the reaction products in the test solution
were analyzed with an ion chromatograph (Dionex ICS-2100, Thermo
Fisher Scientific K.K., Tokyo, Japan) in which an aqueous KOH solution
was used at a flow rate of 1.0 mL min−1 to elute F− and other ions from the
IonPac AS-20 column (4 mm i.d., 250 mm long) at 308 K. Concentration of
KOH in the eluent was gradually increased from 2.5 to 45 mM according to
a time program (Fig. S4). If necessary, the aqueous NaOH and aqueous
H2SO4 test solutions were neutralized with aqueous H2SO4 and aqueous
NaOH, respectively, for the ion-chromatographic analysis. Ion exclusion
chromatography analysis for determination of trifluoroacetate (CF3C(O)
O−) was also performed with an ion chromatograph (Model 8020, Tosoh
Co., Tokyo, Japan) in which terephthalic acid (10 mM) was used at a flow
rate of 0.6 mL min−1 to elute CF3C(O)O− from a TSK-gel OApak-A column
(7.8 mm i.d., 300 mm long) at 313 K [25].
The partial pressure of HFPO was determined for the observed IR
spectrum from the height of the peak at 1162.2 cm−1. Because this peak
overlapped part of a peak of hexafluoropropene, the absorbance due to
hexafluoropropene was subtracted before determining the partial
pressure of HFPO. In making this subtraction, no reaction of hexa-
fluoropropene was presumed to occur. The subtraction reduced the
peak height at 1162.2 cm−1 by less than 3% in all the IR spectra
measured. A calibration curve of HFPO in the partial pressure range
examined was prepared using gas mixtures with known partial pres-
sures of HFPO in air. The partial pressure of HFPO, PHFPO, was de-
scribed by Eq. (26).
5. Experimental
5.1. Reagents
HFPO (purity, 97%) supplied from Daikin Industries (Osaka, Japan)
was used without further purification. This reagent contained hexa-
fluoropropene as an almost unique impurity; the hexafluoropropene
content was determined to be 3.7% from the absorption intensity of a
band at 1797 cm−1 in the infrared spectrum of this reagent. The ab-
sorption coefficient used for this determination was calculated from the
reported absorption constants of HFPO integrated between 970 and
1850 cm−1 [23]. Carbon monoxide gas (207.3 ppmv in synthetic air)
was purchased from Takachiho Chemical Industrial Co. (Tokyo, Japan).
Standard aqueous solutions of NaOH (1 M) and H2SO4 (1 M) were
purchased from Wako Pure Chemical Industries (Osaka, Japan). Water
was purified with an EMD Millipore (Billerica, MA, USA) Milli-Q Gra-
dient A10 system (> 18 MΩ cm).
PHFPO = 23.40 × h – 3.507 × h2,
(26)
5.2. Closed-circulation reactor experiment
where the units of PHFPO are pascals (Pa), and h is the absorbance
(common logarithm) at 1162.2 cm−1 for a 3-m path length. The pro-
cedure for preparing the calibration curve of HFPO (Fig. S5) is de-
scribed in the Supporting Information.
The partial pressure of carbon monoxide was determined from a
calibration curve prepared using standard gas mixtures of carbon
monoxide in air. The absorption bands of carbon monoxide overlapped
other weak but complex bands that originated from the absorption
bands of HFPO and hexafluoropropene in the wavenumber range
2000–2500 cm−1. A loading factor analysis [26] was then used to de-
termine partial pressures of carbon monoxide. Partial pressures of
carbon monoxide, PCO, were described by Eq. (27).
A closed-circulation reactor was used to monitor the decrease of the
partial pressure of HFPO with time while an HFPO-air mixture flowed
over the test solution. Fig. S3 shows a schematic of the closed-circula-
tion reactor. The reactor has been described in detail before [24], ex-
cept for its cylindrical glass liquid cell (component d in Fig. S3a), and is
described only briefly here.
A test solution such as deionized water (volume, 0.180 dm3) was
introduced into the cylindrical glass liquid cell. The cylindrical glass
liquid cell was composed of three parts: a bottom part (86 mm inner
diameter and 40 mm high), a middle part (inner diameter gradually
decreasing from 86 to 16 mm as the height increased from 40 to
60 mm), and a top part consisting of a tube (16-mm inner diameter and
100-mm length). The bottom part had four baffles (ca. 5-mm length and
ca. 10-mm height) at cylindrically symmetric positions on the inner
wall. The test solution was typically stirred at 1000 revolutions per
minute (rpm) using a polytetrafluoroethylene (PTFE)-coated stirring
bar (8-mm diameter × 40-mm length) and a magnetic stirrer. The cy-
lindrical glass liquid cell was placed in a temperature-controlled water
bath typically at 295.9 K and was connected to the closed-circulation
main system with an Allihn condenser between them. The Allihn con-
denser was cooled to 275.2 K to suppress diffusion of water vapor from
the cylindrical glass liquid cell to the main system.
2
3
PCO = 5.405 × SCO – 0.4877 × SCO + 0.4311 × SCO
,
(27)
where the units of PCO are pascals and SCO is the score of the loading
factor corresponding to the IR spectrum of carbon monoxide at a partial
pressure of 5.44 Pa (Fig. S6). Fig. S7 shows the calibration curve of
carbon monoxide.
Declarations of interest
None.
Funding
The HFPO gas mixture was prepared in a two-step procedure by
using an absolute pressure meter to dilute HFPO with synthetic air. The
initial partial pressure of the HFPO was typically set to ∼20 Pa
(2 × 10−4 atm), whereas the total pressure of the gas mixture was
1 atm. A magnetically driven glass pump was used to circulate the gas
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
mixture through the reactor at a flow rate of 0.7 dm3 min−1
.
Acknowledgement
The experimental procedure was as follows. The HFPO gas mixture
was circulated for 1 h without contacting the test solution (Fig. S3a;
volume, 0.681 dm3); after 1 h, the gas circulation path was changed so
that the gas mixture flowed over the test solution for approximately 9 h
(Fig. S3b; volume except for the test solution, 0.962 dm3). The gas
mixture was analyzed every 10 min with a Fourier Transform Infrared
(FTIR) spectrometer JEOL Winspec 50 (JEOL Co., Tokyo, Japan) using a
White-type multi-reflection cell with an optical path length of 3 m; the
The author thanks Shingo Nakamura (Daikin Industries, Ltd.) for
supplying a sample of HFPO.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.jfluchem.2018.04.013.
117

S. Kutsuna
JournalofFluorineChemistry211(2018)109–118
References
organic compounds in the atmosphere: state of the art and challenges, Chem. Rev.
115 (2015) 3919–3983.
[14] O. Levenspiel, Chemical Reaction Engineering, third ed., John Wiley & Sons, NJ,
USA, 1999.
[15] R. Sander, Compilation of Henty’s law constants (version 4.0) for water as solvent,
Atmos. Chem. Phys. 15 (2015) 4399–4981.
[16] R.D. Chambers, Fluorine in Organic Chemistry, John Wiley & Sons, NJ, USA, 2004.
[17] S.E. Schwartz, Mass-transport considerations pertinent to aqueous-phase reactions
of gases in liquid-water clouds, in: W. Jaeschke (Ed.), Chemistry of Multiphase
Atmospheric Systems, Springer-Verlag, Berlin, 1986, pp. 415–471.
[18] J. Lelieveld, P.J. Crutzen, Influences of cloud photochemical processes on tropo-
spheric ozone, Nature 343 (1990) 227–233.
[19] J.E. Johnson, The lifetime of carbonyl sulfide in the troposphere, Geophys. Res. Let.
8 (1981) 938–940.
[20] W.J. De Bruyn, J.A. Shorter, P. Davidovits, D.R. Worsnop, M.S. Zahniser, C.E. Kolb,
Uptake of haloacetyl and carbonyl halides by water surfaces, Environ. Sci. Technol.
29 (1995) 1179–1185.
[21] P.H. Wine, W.L. Chameides, Possible atmospheric lifetimes and chemical reaction
mechanisms for selected HCFCs, HFCs, CH3CCl3, and their degradation products
against dissolution and/or degradation in seawater and cloudwater World
Meteorological Organization, Global Ozone Research and Monitoring Project,
Report 20, Scientific Assessments of Stratospheric Ozone: 1989 (1990) 272–295.
[22] R. Wanninkhof, Relationship between wind speed and gas exchange over the ocean,
J. Geophys. Res. 97 (1992) 7373–7382.
[23] G. Acerboni, J.A. Beukes, N.R. Jensen, J. Hjorth, G. Myhre, C.J. Nielsen,
J.K. Sundet, Atmospheric degradation and global warming potentials of three
perfluoroalkenes, Atmos. Environ. 35 (2001) 4113–4123.
[1] H. Millauer, W. Schwertfeger, G. Siegemund, Hexafluoropropene oxide—a key
compound in organofluorine chemistry, Angew. Chem. Inter. Ed. Engl. 24 (1985)
161–179.
[2] D. Sianesi, A. Pasetti, F. Tarli, The chemistry of hexafluoropropene epoxide, J. Org.
Chem. 31 (1966) 2312–2316.
[3] D.L. Whalen, Mechanisms of hydrolysis and rearrangements of epoxides, Adv. Phys.
Org. Chem. 40 (2005) 247–298.
[4] W. Mabey, T. Mill, Critical review of hydrolysis of organic compounds in water
under environmental conditions, J. Phys. Chem. Ref. Data 7 (1978) 383–415.
[5] J.N. Brönsted, M. Kilpatrick, M. Kilpatrick, Kinetics studies of ethylene oxides, J.
Am. Chem. Soc. 51 (1929) 428–461.
[6] F.A. Long, J.G. Pritchard, Hydrolysis of substituted ethylene oxides in H2O18 so-
lutions, J. Am. Chem. Soc. 78 (1956) 2663–2667.
[7] J.G. Pritchard, F.A. Long, Kinetics and mechanism of the acid-catalyzed hydrolysis
of substituted ethylene oxides, J. Am. Chem. Soc. 78 (1956) 2667–2670.
[8] R.A. Conway, G.T. Waggy, M.H. Spiegel, R.L. Berglund, Environmental fate and
effects of ethylene oxide, Environ. Sci. Technol. 17 (1983) 107–112.
[9] G.P. Ford, C.T. Smith, Gas-phase hydrolysis of protonated oxirane. Ab initio and
semiempirical molecular orbital calculations, J. Am. Chem. Soc. 109 (1987)
1325–1331.
[10] H. Cai, F.P. Guengerich, Mechanism of aqueous decomposition of trichloroethylene
oxide, J. Am. Chem. Soc. 121 (1999) 11656–11663.
[11] J. Huang, C.S. Yeung, J. Ma, E.R. Gayner, D.L. Phillips, A computational chemistry
investigation of the mechanism of the water-assisted decomposition of tri-
chloroethylene oxide, J. Phys. Chem. A 118 (2014) 1557–1567.
[12] T. Yoshioka, J.A. Krauser, F.P. Guengerich, Tetrachloroethylene oxide: hydrolytic
products and reactions with phosphate and lysine, Chem. Res. Toxicol. 15 (2002)
1096–1105.
[13] B. Nozière, M. Kalberer, M. Claeys, J. Allan, B. D’Anna, S. Decesari, E. Finessi,
M. Glasius, I. Grgić, J.F. Hamilton, T. Hoffmann, Y. Iinuma, M. Jaoui, A. Kahnt,
C.J. Kampf, I. Kourtchev, W. Maenhaut, N. Marsden, S. Saarikoski, J. Schnelle-Kreis,
J.D. Surratt, S. Szidat, R. Szmigielski, A. Wisthaler, The molecular identification of
[24] S. Aoki, S. Kutsuna, K. Murano, Evaluation of the alkaline hydrolysis of HCFC-22
(CHClF2) in a closed-circulation reactor, J. Fluor. Chem. 182 (2016) 127–133.
[25] T. Abe, H. Baba, I. Soloshonok, K. Tanaka, Novel way of separating poly-
fluorocarboxylic acids by ion-exclusion chromatography, J. Chromatogr. A 884
(2000) 93–103.
[26] E.D. Malinowski, Factor Analysis in Chemistry, third ed., John Wiley & Sons, NJ,
USA, 2002.
118