Atmospheric chemistry of Z- and E-CF₃CH=CHCF₃

Abstract

The atmospheric fates of Z- and E-CF₃CH=CHCF₃ have been studied, investigating the kinetics and the products of the reactions of the two compounds with Cl atoms, OH radicals, OD radicals, and O₃. FTIR smog chamber experiments measured: k(Cl + Z-CF₃CH=CHCF₃) = (2.59 ± 0.47) × 10⁻¹¹, k(Cl + E-CF₃CH=CHCF₃) = (1.36 ± 0.27) × 10⁻¹¹, k(OH + Z-CF₃CH=CHCF₃) = (4.21 ± 0.62) × 10⁻¹³, k(OH + E-CF₃CH=CHCF₃) = (1.72 ± 0.42) × 10⁻¹³, k(OD + Z-CF₃CH=CHCF₃) = (6.94 ± 1.25) × 10⁻¹³, k(OD + E-CF₃CH=CHCF₃) = (5.61 ± 0.98) × 10⁻¹³, k(O₃ + Z-CF₃CH=CHCF₃) = (6.25 ± 0.70) × 10⁻²², and k(O₃ + E-CF₃CH=CHCF₃) = (4.14 ± 0.42) × 10⁻²² cm³ molecule⁻¹ s⁻¹ in 700 Torr of air/N₂/O₂ diluents at 296 ± 2 K. E-CF₃CH=CHCF₃ reacts with Cl atoms to give CF₃CHClC(O)CF₃ in a yield indistinguishable from 100%. Z-CF₃CH=CHCF₃ reacts with Cl atoms to give (95 ± 10)% CF₃CHClC(O)CF₃ and (7 ± 1)% E-CF₃CH=CHCF₃. CF₃CHClC(O)CF₃ reacts with Cl atoms to give the secondary product CF₃C(O)Cl in a yield indistinguishable from 100%, with the observed co-products C(O)F₂ and CF₃O₃CF₃. The main atmospheric fate for Z- and E-CF₃CH=CHCF₃ is reaction with OH radicals. The atmospheric lifetimes of Z- and E-CF₃CH=CHCF₃ are estimated as 27 and 67 days, respectively. IR absorption cross sections are reported and the global warming potentials (GWPs) of Z- and E-CF₃CH=CHCF₃ for the 100 year time horizon are calculated to be GWP₁₀₀ = 2 and 7, respectively. This study provides a comprehensive description of the atmospheric fate and impact of Z- and E-CF₃CH=CHCF₃.

---

1. Introduction

Chlorofluorocarbons (CFCs) are well-known for their ability to destroy stratospheric ozone and their potency as greenhouse gases.^1–3 They have had several uses, for instance as refrigerants, solvents, in foam blowing and in electronic cleaning.⁴ Having recognized their environmental and atmospheric impacts they were phased out and replaced by hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs). Generally, the HCFCs and HFCs are more environmentally benign in that the HCFCs and HFCs have shorter atmospheric lifetimes than the CFCs, and that the HFCs do not contain chlorine substituents. However, they are both still long-lived greenhouse gases.

Hydrofluoroolefins (HFOs) constitutes a recent class of alternative replacement compounds to the CFCs, HCFCs, and HFCs. They are more reactive in the atmosphere and thereby have a smaller impact on the environment. It is important to know the fate of these compounds before they enter large-scale production and are potentially released to the environment. Z- and E-CF₃CH=CHCF₃ (1,1,1,4,4,4-hexafluoro-2-butene, HFO-1336mzzm) belong to this class of proposed HFOs. CF₃CH=CHCF₃ has been proposed to be used for foam blowing with an improvement of the energy efficiency of the process and the Z-isomer has been shown to have beneficial properties as a substitute for CF₃CH₂CHF₂ (1,1,1,3,3-pentafluoropropane, HFC-245fa) as a refrigerant in organic Rankine cycles.^5,6 The photochemical reactor setup at the Copenhagen Center for Atmospheric Research (CCAR) was used to study the atmospheric chemistry of the Z- and E-isomers of CF₃CH=CHCF₃. The reactions of the two isomers with Cl atoms, OH radicals, OD radicals, and O₃ were studied, investigating the kinetics and the products of the reactions with Cl atoms in order to assess the fates of the compounds in the atmosphere. Only one previous study of the Z-isomer by Baasandorj et al.⁷ exists in the literature. OH radicals, and to a lesser degree, Cl atoms, and O₃ are atmospheric oxidants that initiates the atmospheric removal of Z- and E-CF₃CH=CHCF₃. The kinetics of the reaction of Z-CF₃CH=CHCF₃ with OD radicals was also investigated in the aforementioned study by Baasandorj et al., so OD radical experiments have been included here to be able to compare the present study to their work.⁷ We present here the first determination of the Cl atom and O₃ initiated chemistry of Z-CF₃CH=CHCF₃ and the first determination of the atmospheric chemistry of E-CF₃CH=CHCF₃. The findings from the present study are discussed with respect to the atmospheric chemistry of HFOs.

2. Methodology

2.1 Experimental methods

The CCAR photoreactor is a 101.4 L quartz reactor surrounded by 8 UVA (Osram Eversun L100/79 with the main emission peak at 368 nm) or UVB (Waldmann F85/100 UV6, with a wavelength range of 280–360 nm) lamps, and 16 UVC lamps that are used to initiate the experiments. The reactor is interfaced with a Bruker IFS 66 v/s FTIR spectrometer.⁸ All spectra were obtained using 32 co-added interferograms with a spectral resolution of 0.25 cm⁻¹ and optical pathlengths of 45.10, 52.67, and 55.55 meters. All experiments were performed at 296 ± 2 K and at a total pressure of 700 Torr air/N₂/O₂ diluent. The experiments were performed with Cl atoms, OH radicals, OD radicals or O₃. Cl atoms were produced by photolysis of Cl₂:

Cl₂ + hν(UVA or UVB) → 2Cl (1)

OH radicals were produced by the photolysis of (CH₃)₂CHONO (isopropyl nitrite) or CH₃ONO (methyl nitrite) followed by reaction with O₂ forming HO₂:

(CH₃)₂CHONO + hν(UVC) → (CH₃)₂CHO˙ + NO (2)

(CH₃)₂CHO˙ + O₂ → (CH₃)₂CO + HO₂ (3)

CH₃ONO + hν(UVC) → CH₃O˙ + NO (4)

CH₃O˙ + O₂ → H₂CO + HO₂ (5)

The HO₂ formed from (CH₃)₂CHONO or CH₃ONO reacts with NO to give OH radicals:

HO₂ + NO → OH + NO₂ (6)

OD radicals were produced by photolysis of CD₃ONO (deuterated methyl nitrite):

CD₃ONO + hν(UVC) → CD₃O + NO (7)

CD₃O + O₂ → CD₂O + DO₂ (8)

DO₂ + NO → OD + NO₂ (9)

O₃ was produced using a commercial ozone generator from O₃-Technology (Dielectric barrier discharge; model AC-20). The O₃ was pre-concentrated on a silica gel trap to reduce the amount of O₂ introduced to the chamber.

The relative rate method is a well-known method for measuring rate coefficients of gas phase reactions. The reactions of the reactants (Z- and E-CF₃CH=CHCF₃) with the oxidants are measured relative to the reactions of reference compounds with the oxidants:

Cl/OH/OD + Reactant → Products (10)

Cl/OH/OD + Reference → Products (11)

Plotting the loss of reactant versus the loss of reference the following equation is used:where [Reactant]_t0, [Reactant]_t, [Reference]_t0 and [Reference]_t are the concentrations of the reactant and the reference at times t₀ and t. k_Reactant and k_Reference are the rate coefficients of the reactant and the reference reactions, respectively. The slope of the fit of the experimental data to eqn (I) provides the rate coefficient ratio k_Reactant/k_Reference. For the experiments with Cl atoms C₂H₂, C₂H₄, C₂H₆, and CH₃C(O)CH₃ were used as references. For the experiments with OH radicals C₃H₈ and C₂H₆ were used as references. For the experiments with OD radicals C₂H₆ and n-C₄H₁₀ were used as references.

An absolute rate method was used in the O₃ kinetic experiments. Here the loss of the reactant is monitored over time with an excess of O₃:

The pseudo first order rate coefficients (k^pseudo 1st) obtained from the individual experiments are plotted against the varying O₃ concentrations giving the rate coefficient k₁₂ of the reaction as the slope of the line fitted to the data:

O₃ + Reactant → Products (12)

All reagents except CD₃ONO, CH₃ONO, and (CH₃)₂CHONO were obtained from commercial sources. Z- and E-CF₃CH=CHCF₃ were produced by SynQuest Laboratories at quoted purities of ≤100%. Z-CF₃CH=CHCF₃ was supplied to us by Honeywell. Ultrapure N₂ (≥99.999%), ultrapure O₂ (≥99.995%), synthetic air, Z-CF₃CH=CHCF₃, and E-CF₃CH=CHCF₃ were used as received. The Z-CF₃CH=CHCF₃ sample was devoid of any impurities as analyzed by FTIR. The sample of E-CF₃CH=CHCF₃ contained an impurity of 1.8% Z-CF₃CH=CHCF₃, which was taken into account in the analysis of the IR spectra. CD₃ONO, CH₃ONO, and (CH₃)₂CHONO were produced by the dropwise addition of cold H₂SO₄ (sulfuric acid) to a mixture of NaNO₂ (sodium nitrite) and CD₃OH (deuterated methanol), CH₃OH (methanol) or (CH₃)₂CHOH (isopropanol), respectively. The produced nitrites were trapped using either an ice bath or an isopropanol/dry ice bath. They were stored cold and in the dark. CD₃ONO, CH₃ONO, and (CH₃)₂CHONO were subjected to freeze–pump–thaw cycling before use. The CD₃ONO, CH₃ONO, and (CH₃)₂CHONO samples were devoid of impurities as analyzed by FTIR.

To test for unwanted loss of reagents due to photolysis, dark reactions and heterogeneous reactions, control experiments were performed to investigate these processes. Samples of Z- and E-CF₃CH=CHCF₃ in the chamber were irradiated with UV with no observable loss. Reactant/product mixtures obtained after UV irradiation were allowed to stand in the dark for 30 minutes. No loss or changes were observed, indicating that loss processes due to dark reactions or heterogeneous reactions are not a complication in these experiments. Unless otherwise stated, the quoted uncertainties are two standard deviations from the least-squares fits and our estimated uncertainty of the analysis (typically ±1% of initial concentrations).

2.2 Computational methods

The calculations were carried out with a fourth generation composite method referred to as G4MP2⁹ with the GAUSSIAN 09, suite of programs.¹⁰ G4MP2 theory is approximating a large basis set CCSD(T) single point calculation on a B3LYP/6-31G(2df,p) geometry and is incorporating a so-called higher level correction that is derived by a fit to the experimental values in the G3/05 test set with 454 experimental values.¹¹ The average absolute derivation from the experimental test set values is 1.04 kcal mol⁻¹, which places the G4MP2 results well within chemical accuracy of 10 kJ mol⁻¹. The transition structures for the reactions reported in this work have been confirmed in each case by the calculation of vibrational frequencies (one imaginary frequency) and an intrinsic reaction coordinate analysis. Relative free energies stated within the text correspond to G4MP2 values at 298.15 K. The calculated total energies are available as ESI† (Table S1), which also includes the B3LYP/6-31G(2df,p) optimized geometries in the form of GAUSSIAN archive entries (Table S2, ESI†).

The energies of deuterated species were evaluated using the Freq=readiso keyword with an input from the B3LYP/6-31G(2df,p) frequency calculations. The frequencies were also scaled by 0.9854 as for the G4MP2 calculations. The free energy contributions were obtained in this manner together with the relevant partition functions for the evaluation of relative rate coefficients.

Calculations of theoretical IR spectra used in the product studies were performed with the GAUSSIAN 09, suite of programs,¹² using B3LYP/6-31+G(d,p) and ωB97XD/cc-pVTZ frequency calculations. Optimized geometries and IR spectra can be found in the ESI† (Tables S3–S11 and Fig. S1–S9).

3. Results and discussion

3.1 Relative rate study of Z- and E-CF₃CH=CHCF₃ + Cl

The rates of reactions (13) and (14) were measured relative to reactions (15–17) and (15), (17), and (18), respectively. The initial reaction mixtures were 1.56–2.61 mTorr Z- or E-CF₃CH=CHCF₃, 72.1–83.3 mTorr Cl₂, and 10.4–10.5 mTorr C₂H₂, 5.21 mTorr C₂H₄, 7.82 mTorr C₂H₆ or 10.0–10.4 mTorr CH₃C(O)CH₃ in 700 Torr air. The mixtures were subjected to a total of 25–76 seconds of UV irradiation.

Cl + Z-CF₃CH=CHCF₃ → CF₃CHClCHCF₃ (13)

Cl + E-CF₃CH=CHCF₃ → CF₃CHClCHCF₃ (14)

Cl + C₂H₂ → Products (15)

Cl + C₂H₄ → Products (16)

Cl + CH₃C(O)CH₃ → Products (17)

Cl + C₂H₆ → Products (18)

Fig. 1 shows the loss of Z- and E-CF₃CH=CHCF₃versus the loss of the reference compounds. Linear least squares analyses of the data for k₁₃ in Fig. 1 gives the rate coefficient ratios k₁₃/k₁₅ = 0.56 ± 0.06, k₁₃/k₁₆ = 0.28 ± 0.03, and k₁₃/k₁₇ = 11.0 ± 1.21. Using k₁₅ = (5.07 ± 0.34) × 10⁻¹¹,¹³k₁₆ = (9.29 ± 0.51) × 10⁻¹¹,¹³ and k₁₇ = (2.10 ± 0.15) × 10^−12 14 gives k₁₃ = (2.85 ± 0.29) × 10⁻¹¹, (2.61 ± 0.26) × 10⁻¹¹, and (2.31 ± 0.25) × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹, respectively. The values for k₁₃ are identical within the ranges of uncertainty. Using a similar approach three rate coefficient values for k₁₄ were obtained. A summary of all the rate coefficient ratios and individual rate coefficient determinations in this work is shown in Table 1. The three values for k₁₄ are also identical within the ranges of uncertainty (see Table 1). We choose to quote final values for k₁₃ and k₁₄ as the averages of the three determinations with uncertainties that encompass the extremes of the individual determinations. Hence, k₁₃ = (2.59 ± 0.47) × 10⁻¹¹ and k₁₄ = (1.36 ± 0.27) × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹.

Fig. 1 Panel A: Loss of Z-CF₃CH=CHCF₃versus the loss of C₂H₂ (circles), C₂H₄ (squares), and CH₃C(O)CH₃ (triangles) in the presence of Cl atoms. Panel B: Loss of E-CF₃CH=CHCF₃versus the loss of C₂H₂ (circles), C₂H₆ (squares), and CH₃C(O)CH₃ (triangles) in the presence of Cl atoms.

Table 1 Rate coefficient ratios, reference rate coefficient values, and the individual determinations of the rate coefficients of the reactions of Z-CF₃CH=CHCF₃ (Z) and E-CF₃CH=CHCF₃ (E) with Cl atoms, OH radicals, and OD radicals

Reaction	Reference	k Reactant/kReference	k Reference (cm^3 molecule^−1 s^−1)	k Reactant (cm^3 molecule^−1 s^−1)
a Wallington et al.^13 b Atkinson et al.^14 c Grenier.^36 d Paraskevopoulos and Nip.^37
Z + Cl	C2H2	0.56 ± 0.06	(5.07 ± 0.34) × 10^−11^a	(2.85 ± 0.29) × 10^−11
Z + Cl	C2H4	0.28 ± 0.03	(9.29 ± 0.51) × 10^−11^a	(2.61 ± 0.26) × 10^−11
Z + Cl	CH3C(O)CH3	11.0 ± 1.21	(2.1 ± 0.15) × 10^−12^b	(2.31 ± 0.25) × 10^−11
Z + OH	C3H8	0.38 ± 0.06	(1.1 ± 0.08) × 10^−12^b	(4.14 ± 0.40) × 10^−13
Z + OH	C2H6	1.78 ± 0.18	(2.4 ± 0.08) × 10^−13^b	(4.27 ± 0.44) × 10^−13
Z + OD	C2H6	2.70 ± 0.29	(2.74 ± 0.27) × 10^−13^c	(7.39 ± 0.80) × 10^−13
Z + OD	n-C4H10	0.23 ± 0.03	(2.76 ± 0.22) × 10^−12^d	(6.49 ± 0.75) × 10^−13
E + Cl	C2H2	0.28 ± 0.03	(5.07 ± 0.34) × 10^−11^a	(1.40 ± 0.14) × 10^−11
E + Cl	C2H6	0.25 ± 0.03	(5.9 ± 0.06) × 10^−11^b	(1.48 ± 0.15) × 10^−11
E + Cl	CH3C(O)CH3	5.80 ± 0.58	(2.1 ± 0.15) × 10^−12^b	(1.22 ± 0.12) × 10^−11
E + OH	C3H8	0.14 ± 0.02	(1.1 ± 0.08) × 10^−12^b	(1.52 ± 0.19) × 10^−13
E + OH	C2H6	0.80 ± 0.10	(2.4 ± 0.08) × 10^−13^b	(1.91 ± 0.23) × 10^−13
E + OD	C2H6	2.05 ± 0.26	(2.74 ± 0.27) × 10^−13^c	(5.63 ± 0.71) × 10^−13
E + OD	n-C4H10	0.20 ± 0.02	(2.76 ± 0.22) × 10^−12^d	(5.59 ± 0.68) × 10^−13

---

This is the first determination of k₁₃ and k₁₄. The value of k₁₄ is approximately half of k₁₃. The magnitudes of k₁₃ and k₁₄ are consistent with expectations based on the reactivities of similar HFOs such as E-CF₃CH=CHF, Z- and E-CF₃CF=CHF, and CF₃CF=CF₂, which have the rate coefficient values (4.64 ± 0.59) × 10⁻¹¹,¹⁵ (4.36 ± 0.48) × 10⁻¹¹,¹⁶ (5.00 ± 0.56) × 10⁻¹¹,¹⁶ and (2.7 ± 0.3) × 10^−11 17 cm³ molecule⁻¹ s⁻¹, respectively.

3.2 Relative rate study of Z- and E-CF₃CH=CHCF₃ + OH

The rate of reaction (19) and (20) were measured relative to reactions (21) and (22). The reaction mixtures consisted of 1.60–1.98 mTorr Z-CF₃CH=CHCF₃ or 2.08 mTorr E-CF₃CH=CHCF₃, 7.19–7.40 mTorr C₃H₈ or 7.30 mTorr C₂H₆, and 73.0 mTorr (CH₃)₂CHONO or 157–224 mTorr CH₃ONO in 700 Torr of air. The mixtures were subjected to a total of 265–3180 seconds of UV irradiation.

OH + Z-CF₃CH=CHCF₃ → Products (19)

OH + E-CF₃CH=CHCF₃ → Products (20)

OH + C₃H₈ → Products (21)

OH + C₂H₆ → Products (22)

Fig. 2 shows the loss of Z- and E-CF₃CH=CHCF₃versus the loss of reference compounds C₃H₈ and C₂H₆ in the presence of OH radicals. Rate coefficient ratios obtained from linear least squares analyses of the data in Fig. 2 are shown in Table 1. Using the rate coefficient values of k₂₁ = (1.1 ± 0.08) × 10⁻¹² and k₂₂ = (2.4 ± 0.08) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹,¹⁴ gives two values for each k₁₉ and k₂₀, which are identical within the ranges of uncertainty (Table 1). We choose to quote final values for k₁₉ and k₂₀ as the average of the two determinations with uncertainties that encompass the extremes of the individual determinations. Hence, k₁₉ = (4.21 ± 0.62) × 10⁻¹³ and k₂₀ = (1.72 ± 0.42) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹. Baasandorj et al.⁷ reported a value for k₁₉ of (4.91 ± 0.50) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹, which is in reasonable agreement with the value determined here. The value of k₂₀ is approximately 2/5 the size of k₁₉. The values for k₁₉ and k₂₀ determined in the present work are consistent with expectations based on the OH radical reactivities of similar HFOs such as Z- and E-CF₃CF=CHF and Z- and E-CF₃CH=CHCl, which have rate coefficients of (1.22 ± 0.14) × 10⁻¹²,¹⁶ (2.15 ± 0.23) × 10⁻¹²,¹⁶ (8.45 ± 1.52) × 10⁻¹³,¹⁸ and (3.61 ± 0.37) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹,¹⁸ respectively.

Fig. 2 Panel A: Loss of Z-CF₃CH=CHCF₃versus the loss of C₃H₈ (circles) and C₂H₆ (triangles) in the presence of OH radicals. Panel B: Loss of E-CF₃CH=CHCF₃versus the loss of C₃H₈ (circles) and C₂H₆ (triangles) in the presence of OH radicals.

3.3 Relative rate study of Z- and E-CF₃CH=CHCF₃ + OD

The rate of reactions (23) and (24) were measured relative to reactions (25) and (26). The initial reaction mixtures consisted of 1.68–2.08 mTorr Z- or E-CF₃CH=CHCF₃, 6.67–8.34 mTorr C₂H₆ or 7.30–8.24 mTorr n-C₄H₁₀, and 169–230.4 mTorr CD₃ONO in 700 Torr air. The mixtures were subject to a total of 860–1080 seconds of UV irradiation.

OD + Z-CF₃CH=CHCF₃ → Products (23)

OD + E-CF₃CH=CHCF₃ → Products (24)

OD + C₂H₆ → Products (25)

OD + n-C₄H₁₀ → Products (26)

Fig. 3 shows the loss of Z- and E-CF₃CH=CHCF₃versus the loss of the two reference compounds C₂H₆ and n-C₄H₁₀ in the presence of OD radicals. Linear regression of the data gives the rate coefficient ratios and values of k₂₃ and k₂₄ shown in Table 1. The two values for k₂₃ are identical within the ranges of uncertainty. The same is the case for k₂₄. We choose to quote final values for k₂₃ and k₂₄ as the averages of the two determinations with uncertainties that encompass the extremes of the individual determinations. Hence, k₂₃ = (6.94 ± 1.25) × 10⁻¹³ and k₂₄ = (5.61 ± 0.98) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹. The value of k₂₃ determined here is slightly higher than that reported by Baasandorj et al. ((5.73 ± 0.50) × 10⁻¹³ cm³ molecule⁻¹ s^−1 7), still they are in agreement within the ranges of uncertainty.

Fig. 3 Panel A: Loss of Z-CF₃CH=CHCF₃versus the loss of C₂H₆ (circles) and n-C₄H₁₀ (squares) in the presence of OD radicals. Panel B: Loss of E-CF₃CH=CHCF₃versus the loss of C₂H₆ (circles) and n-C₄H₁₀ (squares) in the presence of OD radicals.

No other studies of OD radicals with HFOs are available in the literature making it difficult to compare the values of k₂₃ and k₂₄ to other similar compounds. The value of k₂₄ is 4/5 the size of k₂₃. This was unexpected, since the reactions of E-CF₃CH=CHCF₃ with the other oxidants in this study are significantly lower than the corresponding reactions of Z-CF₃CH=CHCF₃.

3.4 Absolute rate study of Z- and E-CF₃CH=CHCF₃ + O₃

Reaction mixtures consisted of 1.53–1.78 mTorr Z-CF₃CH=CHCF₃ or 2.08–2.81 mTorr E-CF₃CH=CHCF₃, 0.73–6.41 Torr O₃, and 0–10.4 mTorr c-C₆H₁₂ (cyclohexane) in 700 Torr air or 140 Torr O₂ made up to 700 Torr with N₂. Reaction (27) and (28) were monitored over time.

O₃ + Z-CF₃CH=CHCF₃ → Products (27)

O₃ + E-CF₃CH=CHCF₃ → Products (28)

Ozonolysis can be a source of OH radicals. To avoid complications due to the reaction of Z- and E-CF₃CH=CHCF₃ + OH, c-C₆H₁₂ was added to the reaction mixture as an OH radical scavenger. In the absence of c-C₆H₁₂ the loss of Z-CF₃CH=CHCF₃ + O₃ was found to be approximately 15% higher than when c-C₆H₁₂ was present. Over the ratios of [c-C₆H₁₂]/[Z- or E-CF₃CH=CHCF₃] = 1.8–6.8 no difference in the O₃ reaction rate was observed. Fig. 4 and Fig. S1 in the ESI† show plots of the pseudo first order rate coefficients for Z- and E-CF₃CH=CHCF₃versus O₃ concentration, with the former obtained from the slopes of the degradation of Z- and E-CF₃CH=CHCF₃versus time at different O₃ concentrations (see insets in Fig. 4 and Fig. S1, ESI†). The slopes of the plots in Fig. 4 and Fig. S1 (ESI†) give the O₃ rate coefficients for Z- and E-CF₃CH=CHCF₃, respectively. Hence, k₂₇ = (6.25 ± 0.70) × 10⁻²² and k₂₈ = (4.14 ± 0.42) × 10⁻²² cm³ molecule⁻¹ s⁻¹.

Fig. 4 Pseudo first-order rate coefficients of the reaction O₃ + Z-CF₃CH=CHCF₃versus O₃ concentration. The inset shows the loss of Z-CF₃CH=CHCF₃versus time at different O₃ concentrations. The symbols indicate different O₃ partial pressures: 0.73 Torr (white circles), 2.01 Torr (gray circles), 2.91 Torr (black circles), 4.39 Torr (white triangles), 5.71 Torr (gray triangles), and 6.41 Torr (black triangles). See text for details.

Baasandorj et al.⁷ reported an upper limit for the reaction of Z-CF₃CH=CHCF₃ + O₃ of k₂₇ < 6 × 10⁻²¹ cm³ molecule⁻¹ s⁻¹, which is in agreement with the rate coefficient determined in the present work. The measured rate coefficient for reaction (28) is the slowest rate coefficient ever determined using the photochemical reactor at CCAR. Prior to this study the slowest rate coefficient for a similar compound reacting with ozone determined is of the reaction of Z-CF₃CF=CHF + O₃, which is determined to be (1.45 ± 0.15) × 10⁻²¹ cm³ molecule⁻¹ s⁻¹.¹⁶ The value of k₂₈ is approximately 2/3 that of k₂₇.

3.5 Reactivity trends

As mentioned earlier, one previous study⁷ exists in the literature on the reactivity of Z-CF₃CH=CHCF₃, but no studies have ever been conducted involving E-CF₃CH=CHCF₃. Furthermore, the kinetic study of Baasandorj et al. was limited to the reactions of OH radicals, OD radicals and O₃. The rate coefficients reported by Baasandorj et al.⁷ are all in agreement with our results, although Baasandorj et al.⁷ also conclude that k_OD = k_OH for Z-CF₃CH=CHCF₃ within their reported uncertainties. As seen from Table 1, this is contrary to our findings. We find that the rate coefficient for the reaction with OD radicals is larger than that for the reaction with OH radicals for both compounds. Using computational methods, we investigate this further in Section 3.6.

A summary of the final rate coefficients obtained for all the kinetic experiments is shown in Table 2 along with the rate coefficient values determined previously by Baasandorj et al. as well as rate coefficients of the reactions for other structurally similar HFOs. The Z-isomer has reaction rate coefficients that are greater than those of the E-isomer in the reactions with all the oxidants studied here. This can be explained by the structural differences of the Z- and E-isomers, where more strain is released from the Z-isomer upon addition of the oxidants to the double bond, making it more favorable. Some trends can be observed for the HFOs presented in Table 2. Firstly, the rate coefficients of the reactions with Cl atoms determined here are of the same order as the ones determined for other similar HFOs, but with smaller reaction rate coefficients than the partially fluorinated HFOs, indicating a decrease in reactivity when exchanging a halogen atom (Cl or F) with a CF₃ group. For CF₃CF=CHF and CF₃CH=CHCl the reactivities of the Z- and E-isomers towards Cl atoms are identical within the ranges of uncertainty. In this study the reactivity of the Z-isomer of CF₃CH=CHCF₃ has been found to be greater than that of the E-isomer. Secondly, the reactions of CF₃CF=CF₂, Z- and E-CF₃CF=CHF with OH radicals proceed with rate coefficients that are one order of magnitude larger than those determined for the other HFOs. This could indicate an increasing effect on the reactivity towards OH radicals by having fluorine on both carbons on the double bond. Additional computational studies could aid the interpretation of this observation. The length of the perfluorinated chain may not be of great importance since the rate coefficient of E-(CF₃)₂CFCH=CHF + OH is approximately half the size of the one of E-CF₃CH=CHF + OH and identical to the one of E-CF₃CH=CHCl + OH. Thirdly, the HFOs listed from literature have reactivities towards O₃ that are 1–2 orders of magnitude faster than the O₃ reactivities of Z- and E-CF₃CH=CHCF₃. This could be due to differences in reactivity of fluorinated propenes and butenes towards O₃. Additional studies on other fluorinated butenes would be needed to verify this.

Table 2 Final rate coefficients for the reactions of Z- and E-CF₃CH=CHCF₃ with Cl atoms, OH radicals, OD radicals, and O₃ as well as the available literature data for Z-CF₃CH=CHCF₃ and structurally similar HFOs. All units are cm³ molecule⁻¹ s⁻¹

Compound	k Cl (10^−11)	k OH (10^−13)	k OD (10^−13)	k O3 (10^−22)
a Baasandorj et al.^7 b Andersen et al.^18 c Gierczak et al.^38 d Sulbaek Andersen et al.^19 e Hurley et al.^16 f Søndergaard et al.^15 g Mashino et al.^17 h Papadimitriou and Burkholder.^39
Z-CF3CH=CHCF3	2.59 ± 0.47	4.21 ± 0.62	6.94 ± 1.25	6.25 ± 0.70
	—	4.91 ± 0.50^a	5.73 ± 0.50^a	<60^a
E-CF3CH=CHCF3	1.36 ± 0.27	1.72 ± 0.42	5.61 ± 0.98	4.14 ± 0.42
Z-CF3CH=CHCl	6.26 ± 0.84^b	8.45 ± 1.52^b	—	15.3 ± 1.20^b
	—	9.46 ± 0.85^c	—	—
E-CF3CH=CHCl	5.22 ± 0.72^d	3.61 ± 0.37^b	—	14.6 ± 1.20^d
	—	3.76 ± 0.35^c	—	—
Z-CF3CF=CHF	4.36 ± 0.48^e	12.2 ± 1.4^e	—	14.5 ± 1.5^e
E-CF3CF=CHF	5.00 ± 0.56^e	21.5 ± 2.3^e	—	198 ± 15^e
E-CF3CH=CHF	4.64 ± 0.59^f	9.25 ± 1.72^f	—	28.1 ± 2.1^f
CF3CF=CF2	2.7 ± 0.3^g	24 ± 3^g	—	<30^g
E-(CF3)2CFCH=CHF	—	3.26 ± 0.26^h	—	—

---

It is interesting to compare the reactivities of Z- and E-CF₃CH=CHCF₃ to Z-and E-CF₃CH=CHCl. For both pairs of isomers it can be seen in Table 2 that both the Cl atom, the OH radical and the O₃ rate coefficients obtained for the Z-isomers are faster than the ones obtained for the E-isomers.^18,19 All the rate coefficients obtained for Z- and E-CF₃CH=CHCl are larger than for Z- and E-CF₃CH=CHCF₃ even though they have the same number of halogen/CF₃ substituents. This can be rationalized by the fact that the total number of fluorine atoms, and associated electron-withdrawing effect, in Z- and E-CF₃CH=CHCF₃ is double that found in Z- and E-CF₃CH=CHCl. Furthermore, greater steric hindrance is associated with the CF₃ group than that from the Cl atom. In the case of CF₃CF=CHF, the E-isomer has a greater reactivity than the Z-isomer towards Cl atoms, OH radicals and O₃, so no general Z- versus E-isomer trend can be assessed from the data in Table 2. Further experimental and computational studies of general trends in HFO reactivity would be of interest, but beyond the scope of the present study.

3.6 Computational study of the reactivity of OH and OD radicals

The calculated geometries of the reactant alkenes, Z- and E-CF₃CH=CHCF₃, are shown in Fig. 5 together with the geometries of the reaction complexes between the OH radical and the alkenes, the product alkyl radicals, and the connecting transition states. It is worth noting that the transition state in each case closely resembles the OH–alkene complex. We also note that the Z-isomer gives rise to the transition state resembling the complex closest, which is most likely due to the larger strain release in this case, which gives an energetic bonus earlier during the progress of the addition. The free energy changes relative to the separated reactants (alkene + OH and alkene + OD) are shown in Table 3. The formation of a complex between the E-isomer and the OH radical is endothermic by approximately 5.2 kJ mol⁻¹ and there is a barrier for formation of the alkyl radical of 22.6 kJ mol⁻¹. The overall reaction is exothermic by −98.1 kJ mol⁻¹.

Fig. 5 Structures of optimized geometries of 1: E-CF₃CH=CHCF₃, 2: Z-CF₃CH=CHCF₃, 3: complex of OH + E-CF₃CH=CHCF₃, 4: alkyl radical product formed from OH + E-CF₃CH=CHCF₃, 5: complex of OH + Z-CF₃CH=CHCF₃, 6: alkyl radical product formed from OH + Z-CF₃CH=CHCF₃, and transition states (TS) connecting 3 → 4 and 5 → 6. The highlighted bond lengths are in units of Ångstrøm. See text for details.

Table 3 Free energy changes associated with the formation of CF₃CH=CHCF₃–OH/OD complexes and the formation of alkyl radicals from the complexes as well as the free energy changes associated with the formation of alkyl radicals from the reactions with Cl atoms^a

	ΔrG (complex)	ΔrG^‡ (barrier)	ΔrG (alkyl radical)
a G4MP2 values in kJ mol^−1 at 298.15 K.
E-CF3CH=CHCF3 + OH	5.2	22.6	−98.1
E-CF3CH=CHCF3 + OD	4.9	22.1	−99.6
Z-CF3CH=CHCF3 + OH	2.7	7.5	−112.5
Z-CF3CH=CHCF3 + OD	2.5	7.3	−114.1
E-CF3CH=CHCF3 + Cl	—	—	−57
Z-CF3CH=CHCF3 + Cl	—	—	−60

---

The OD equivalent gives rise to a slightly more weakly bound complex (by 0.3 kJ mol⁻¹) and a barrier, which is slightly lower (0.5 kJ mol⁻¹). The formation of the alkyl radical on the other hand is more exothermic by 1.5 kJ mol⁻¹. In the case of the Z-isomer the barriers and the endothermicities of formation are significantly reduced. The barrier for the formation of the alkyl radical from the RCF₃–HO adduct is 7.5 kJ mol⁻¹ whereas the adduct is 2.7 kJ mol⁻¹ less stable than the separated reactants. The alkyl radical is more stable than the reactants by −112.5 kJ mol⁻¹ and the increased stability reflects release of strain from the Z-isomer. The effect of the deuterium substitution is slightly reduced compared to the E-isomer. The barrier for the OD addition is lower by 0.2 kJ mol⁻¹. The values are summarized in Table 3. The lower barrier in the case of OD is attributed to the generally weaker hydrogen bonds that are in play when a hydrogen is replaced by a deuterium.

The relative barrier difference for OH versus OD addition is in qualitative agreement with experimental finding that OD addition proceeds faster than OH addition in the sense that the barrier for OD addition is smaller for both Z- and E-CF₃CH=CHCF₃. The experimental result that the effect is smaller in the case of the Z-isomer is also in agreement with the calculations, because the calculated barrier is less impacted by isotope substitution in the Z-isomer case. At a first glance, the barrier differences seem small compared to the experimental relative rate coefficients. To assess the impact on the rate coefficients we evaluated the ratios of k_OD/k_OH for both Z- and E-CF₃CH=CHCF₃ from the partition functions, Q, and barrier differences using the equation:

where TS is the transition state, complex is the OH/OD–alkene complexes, subscripts on the partition functions denote either the OH or OD radical reaction, and ΔΔG^‡ is the difference between the barrier heights to form the alkyl radical products. The k_OD/k_OH ratio is found to be 1.8 for E-CF₃CH=CHCF₃ and 1.1 for Z-CF₃CH=CHCF₃. This trend is consistent with the experimental results, showing a larger difference between the OD and OH rate coefficients for E-CF₃CH=CHCF₃ than for Z-CF₃CH=CHCF₃. The calculated total energies and optimized geometries can be found in the ESI† (Tables S1 and S2).

3.7 Product study of E-CF₃CH=CHCF₃ + Cl

The initial reaction mixtures for the product study of the reaction E-CF₃CH=CHCF₃ + Cl consisted of 2.61–4.07 mTorr E-CF₃CH=CHCF₃ and 73.2–89.0 mTorr Cl₂ in 700 Torr air/N₂/O₂ diluent. The mixtures were subjected to a total of 33–52 seconds of UV irradiation. The reaction proceeds with the addition of a Cl atom to the double bond creating an alkyl radical, that will react with O₂ forming a peroxy radical (RO₂):

Cl + E-CF₃CH=CHCF₃ → CF₃CHClCHCF₃ (14)

CF₃CHClCHCF₃ + O₂ → CF₃CHClCH(O₂)CF₃ (29)

The peroxy radical can react with other peroxy radicals, itself or NO (if present), to form an alkoxy radical (RO):

CF₃CHClCH(O₂)CF₃ + RO₂ → CF₃CHClCH(O˙)CF₃ + RO + O₂ (30)

The alkoxy radical can then react with O₂ to give a ketone or decompose via C–C bond scission to give a CF₃CHCl radical and CF₃CHO. CF₃CHCl will react with O₂ giving CF₃C(O)Cl:

CF₃CHClCH(O˙)CF₃ + O₂ → CF₃CHClC(O)CF₃ + HO₂ (31a)

CF₃CHClCH(O˙)CF₃ + M → CF₃CHCl + CF₃CHO + M (31b)

Fig. 6 shows spectra of a mixture of 2.61 mTorr E-CF₃CH=CHCF₃ and 73.2 mTorr Cl₂ in 700 Torr air, before (panel A) and after (panel B) 24 seconds of UV irradiation. 59% of E-CF₃CH=CHCF₃ was consumed in the irradiation. Panel E is a residual spectrum obtained by subtracting all remaining E-CF₃CH=CHCF₃ from panel B (panel B – 0.41 × panel A). IR features present in panel E are assigned to CF₃CHClC(O)CF₃, as the only observed product. While we do not have sample of CF₃CHClC(O)CF₃ with which to obtain a genuine reference spectrum, a calculated spectrum of the compound is shown in panel F. This and other spectra were calculated using the GAUSSIAN 09 suite of programs, revision D.01 at the B3LYP/6-31+G(d,p) level (see below).¹² No formation of CF₃CHO (panel C) or CF₃C(O)Cl (panel D) was observed in these experiments. Upper limits for the yields of CF₃CHO and CF₃C(O)Cl were determined as 1% for both compounds. Therefore, we conclude that reaction (31b) is not significant in the Cl atom initiated degradation. Fig. 7 shows the formation of CF₃CHClC(O)CF₃ as a function of the loss of E-CF₃CH=CHCF₃ in the presence of Cl atoms.

Fig. 6 Panel A: IR spectrum of 2.61 mTorr E-CF₃CH=CHCF₃ and 73.2 mTorr Cl₂ before UV irradiation, panel B: spectrum of the reaction mixture after 24 seconds UV irradiation, panel C: reference spectrum of CF₃CHO, panel D: reference spectrum of CF₃C(O)Cl, panel E: residual spectrum (panel B – 0.41 × panel A) assigned to CF₃CHClC(O)CF₃, and panel F: calculated IR spectrum of CF₃CHClC(O)CF₃ at the B3LYP/6-31+G(d,p) level. See text for details.

Fig. 7 Formation of CF₃CHClC(O)CF₃versus the loss of E-CF₃CH=CHCF₃ in the presence of Cl atoms. The shades indicate the O₂ partial pressure: 16 Torr (white), 140 Torr (gray), and 700 Torr (black) in a total of 700 Torr made up with air or N₂.

Calculated spectra of E-CF₃CH=CHCF₃ and several conformers of CF₃CHClC(O)CF₃ were calculated at the B3LYP/6-31+G(d,p) level. One conformer of CF₃CHClC(O)CF₃ was calculated at the ωB97XD/cc-pVTZ level being in agreement with the other calculations. The optimized geometries and IR spectra are shown in the ESI† (Tables S3–S8 and Fig. S2–S7).

3.8 Product study of Z-CF₃CH=CHCF₃ + Cl

A product study of the reaction of Z-CF₃CH=CHCF₃ + Cl was performed using initial reaction mixtures of 1.12–1.91 mTorr Z-CF₃CH=CHCF₃, 75.4–87.3 mTorr Cl₂, and 0 or 10.4 mTorr NO in 700 Torr air or O₂ diluent. The mixtures were subjected to a total of 17–29 seconds of UV irradiation. The reaction of Z-CF₃CH=CHCF₃ + Cl is initiated by the addition of Cl atoms to the double bond of Z-CF₃CH=CHCF₃ creating an alkyl radical as for E-CF₃CH=CHCF₃ in reaction (14).

Cl + Z-CF₃CH=CHCF₃ → CF₃CHClCHCF₃ (13)

After addition of the Cl atom, the Z/E isometry is lost, free rotation of the C–C bond is possible, and the following reactions in the degradation of the Z-isomer will be the same as observed for the E-isomer in reactions (29)–(32).

Fig. 8 shows spectra of a reaction mixture of 1.91 mTorr Z-CF₃CH=CHCF₃ and 75.4 mTorr Cl₂ in 700 Torr air, before (panel A) and after (panel B) 17 seconds of UV irradiation. 62% of Z-CF₃CH=CHCF₃ was consumed in the irradiation. Panel C shows a reference spectrum of E-CF₃CH=CHCF₃ and panel D is a residual spectrum obtained by subtracting all remaining features of Z-CF₃CH=CHCF₃ (0.38 × panel A) and E-CF₃CH=CHCF₃ (0.03 × panel C) from panel B. This spectrum is assigned to CF₃CHClC(O)CF₃. Panel E shows the residual spectrum of CF₃CHClC(O)CF₃ obtained from the product study of E-CF₃CH=CHCF₃.

Fig. 8 Panel A: A mixture of 1.91 mTorr Z-CF₃CH=CHCF₃ and 75.4 mTorr Cl₂ in 700 Torr air before UV irradiation, panel B: the reaction mixture after 17 seconds UV irradiation, panel C: reference spectrum of E-CF₃CH=CHCF₃, panel D: residual spectrum obtained from subtracting features of Z- and E-CF₃CH=CHCF₃ (panel B – 0.38 × panel A – 0.03 × panel C) assigned to CF₃CHClC(O)CF₃, and panel E: spectrum of CF₃CHClC(O)CF₃ obtained from the product study of E-CF₃CH=CHCF₃ in the presence of Cl atoms. See text for details.

Fig. 9 shows the formation of CF₃CHClC(O)CF₃ and E-CF₃CH=CHCF₃versus the loss of Z-CF₃CH=CHCF₃. The shades of the symbols indicate different O₂ partial pressures and the presence or absence of NO. No effect of varying the O₂ partial pressure was observed. The formation of E-CF₃CH=CHCF₃ is evidence of isomerization. After a Cl atom has added to the double bond in Z-CF₃CH=CHCF₃ (reaction (13)), the center C–C bond has free rotation and the double bond will reform to the more stable E-conformation in competition with reaction (29) expelling a Cl atom:

CF₃CHClCHCF₃ → E-CF₃CH=CHCF₃ + Cl (33)

Fig. 9 Formation of CF₃CHClC(O)CF₃ (circles, left y-axis) and E-CF₃CH=CHCF₃ (triangles, right y-axis) versus the loss of Z-CF₃CH=CHCF₃ in the presence of Cl atoms in a total of 700 Torr air in the absence (gray) and presence (white) of NO or 700 Torr O₂ in the absence of NO (black). The lines are non-linear fits to the data. See text for details.

As discussed in Section 3.7 E-CF₃CH=CHCF₃ reacts with Cl to give 100% CF₃CHClC(O)CF₃. No evidence of isomerization was observed in the degradation of E-CF₃CH=CHCF₃. During the course of the reaction of Z-CF₃CH=CHCF₃ with Cl atoms the apparent yield of E-CF₃CH=CHCF₃ decreases and the concentration of CF₃CHClC(O)CF₃ increases in a non-linear way. Fits to the data are based on reactions (34a), (34b) and (14):

Cl + Z-CF₃CH=CHCF₃ → αE-CF₃CH=CHCF₃ + Cl (34a)

Here α and β are the initial yields of E-CF₃CH=CHCF₃ and CF₃CHClC(O)CF₃, respectively. A method described by Meagher et al.,²⁰ is used to fit the E-CF₃CH=CHCF₃ and CF₃CHClC(O)CF₃ data. The equation describing the fractional conversion of Z- to E-CF₃CH=CHCF₃ is:Here k₃₄ and k₁₄ are the rate coefficients of the reactions of Cl + Z- and E-CF₃CH=CHCF₃ determined above, where k₃₄ = k₁₃, and x is the conversion of Z-CF₃CH=CHCF₃, defined as:In this fit, the k₁₄/k₁₃ ratio is known, so this value is fixed to be:This gives an initial yield of E-CF₃CH=CHCF₃ of α = (6.9 ± 0.8)%. The fit to the CF₃CHClC(O)CF₃ data uses eqn (IV) as a modification to a linear fit, accounting for reaction (14), describing the conversion of Z-CF₃CH=CHCF₃ to CF₃CHClC(O)CF₃:Fitting eqn (VII) to the CF₃CHClC(O)CF₃ data with the known values of k₁₄/k₁₃ and α, the initial yield of CF₃CHClC(O)CF₃ is β = (95 ± 10)% in the absence of NO. Combining the initial yields of E-CF₃CH=CHCF₃ and CF₃CHClC(O)CF₃ mass balance is obtained. It is clear that at atmospheric pressure, the dominant pathway of the reaction of Cl + Z-CF₃CH=CHCF₃ is viareaction (34b).

In the presence of NO the yield of CF₃CHClC(O)CF₃ is slightly decreased compared to the experiments in the absence of NO. The yield of E-CF₃CH=CHCF₃ is not affected by the presence of NO, so eqn (VII) can be used to fit the CF₃CHClC(O)CF₃ data using the known values of k₁₄/k₁₃ and α. This gives an initial yield of CF₃CHClC(O)CF₃ of β = (81 ± 8)% in the presence of NO. The formation of ClNO and ClNO₂ was observed in these experiments. A minor product with an IR absorption peak at 1699 cm⁻¹ was observed, which, assuming conserved mass balance, could have a yield of 12%. This compound was not positively identified, but could be due to the formation of an organic nitrate compound.

While we did not perform any comprehensive OH radical initiated product studies of Z- and E-CF₃CH=CHCF₃, Baasandorj et al. identified some products of the reaction of Z-CF₃CH=CHCF₃ with OH radicals.⁷ They observed C(O)F₂ and CF₃CHO as the main F-containing products, suggesting that C–C scission of the central carbon bond could be a more significant pathway in the atmospheric degradation initiated by OH radical than what we have been able to observe above in the Cl atom initiated reaction.

Calculated IR spectra of two conformers of Z-CF₃CH=CHCF₃ at the B3LYP/6-31+G(d,p) level are supplied in the ESI† (Tables S9, S10 and Fig. S8, S9) and show good agreement with the experimental spectrum.

3.9 Isomerization of Z- and E-CF₃CH=CHCF₃

Experiments were performed to investigate any potential isomerization in the reactions of Z-CF₃CH=CHCF₃ with OH or OD radicals giving E-CF₃CH=CHCF₃ as observed for the reaction of Z-CF₃CH=CHCF₃ with Cl atoms. Reaction mixtures consisting of 1.21–1.37 mTorr Z-CF₃CH=CHCF₃ and 43.9 mTorr CH₃ONO or 42.9 mTorr CD₃ONO in 700 Torr air diluent were subjected to a total of 110–210 seconds of UV irradiation. No formation of E-CF₃CH=CHCF₃ was observed in these experiments. Calculations were performed of the energy of the initial alkyl radical formed from the addition of Cl atoms to Z-CF₃CH=CHCF₃. The alkyl radical energy is approximately twice that of the alkyl radical formed with OH or OD radicals (see Table 3), making it more likely that the reaction can proceed in the reverse direction eliminating the Cl atom and reforming CF₃CH=CHCF₃. The E-isomer has less strain than the Z-isomer thus making it more likely to be formed. A similar calculation was performed of the alkyl radical formed from the reaction of Cl + E-CF₃CH=CHCF₃. The energy of this initial alkyl radical is slightly smaller than that from the Z-isomer (see Table 3), but in the experiments, no isomerization was observed. The optimized structures of both initial alkyl radicals formed from Cl atoms are shown in Fig. S11 in the ESI.† Comparing the barriers of formation of the alkyl radicals from the OH/OD radical additions, the E-isomer barriers are more than double those of the Z-isomer, making the barriers the potential limiting factor of isomerization. Similar barriers can be expected from the Cl atom additions to Z- and E-CF₃CH=CHCF₃, explaining why no isomerization was observed for the reaction Cl + E-CF₃CH=CHCF₃. It would be expected that E-CF₃CH=CHCF₃ will not isomerize in the reactions with OH or OD radicals for the same reasons as discussed above. No additional experiments were performed with OH/OD + E-CF₃CH=CHCF₃ besides the ones used for determining the rate coefficients.

3.10 Product study of CF₃CHClC(O)CF₃ + Cl

To further investigate the oxidation of the primary product, CF₃CHClC(O)CF₃, experiments were conducted in which all of the starting material, Z-CF₃CH=CHCF₃, was allowed to react with Cl to generate CF₃CHClC(O)CF₃. Only then, the loss of CF₃CHClC(O)CF₃ and the subsequent formation of products was monitored, normalizing the products to the initial concentration of CF₃CHClC(O)CF₃. All E-CF₃CH=CHCF₃ had also reacted at this point. The initial reaction mixtures were 1.12–1.91 mTorr Z-CF₃CH=CHCF₃ and 75.4–87.3 mTorr Cl₂ in 700 Torr air or O₂. The mixtures were subjected to a total of 3150–5400 seconds of UV irradiation after all Z-CF₃CH=CHCF₃ was converted to CF₃CHClC(O)CF₃. Three products are observed in the degradation of CF₃CHClC(O)CF₃: CF₃C(O)Cl, C(O)F₂, and CF₃O₃CF₃. Fig. 10 shows IR spectra recorded before (panel A) and after (panel B) UV irradiation of a mixture of CF₃CHClC(O)CF₃ and Cl₂. Panel C shows the resulting spectrum after subtracting all features of CF₃CHClC(O)CF₃ in panel A from panel B (panel B – 0.56 × panel A). Panels D and E show reference spectra of C(O)F₂ and CF₃C(O)Cl, respectively. The residual obtained when subtracting features of C(O)F₂ and CF₃C(O)Cl from panel C is shown in panel F. Panel G shows a reference spectrum of CF₃O₃CF₃. The residual spectrum is assigned to CF₃O₃CF₃, and quantified as such.²¹

Fig. 10 Panel A: IR spectrum of CF₃CHClC(O)CF₃ + Cl₂ before UV irradiation, panel B: spectrum of the reaction mixture after 25 minutes of UV irradiation, panel C: the resulting spectrum of subtracting the remaining features of CF₃CHClC(O)CF₃ from panel B (panel B – 0.56 × panel A), panel D: reference spectrum of C(O)F₂, panel E: reference spectrum of CF₃C(O)Cl, panel F: residual spectrum, when subtracting C(O)F₂ and CF₃C(O)Cl from panel C, and panel G: reference spectrum of CF₃O₃CF₃. The residual spectrum in panel F is assigned to CF₃O₃CF₃. See text for details.

The degradation of CF₃CHClC(O)CF₃ is initiated by Cl atoms by abstraction of the hydrogen atom forming an alkyl radical that reacts with O₂ forming a peroxy radical:

CF₃CHClC(O)CF₃ + Cl → CF₃CClC(O)CF₃ + HCl (35)

CF₃CClC(O)CF₃ + O₂ → CF₃C(O₂)ClC(O)CF₃ (36)

The peroxy radical will then react with itself or NO (if present) or another peroxy radical forming an alkoxy radical:

CF₃C(O₂)ClC(O)CF₃ + RO₂ → CF₃C(O˙)ClC(O)CF₃ + RO + O₂ (37)

The alkoxy radical formed will break by C–C scission to give CF₃C(O)Cl and a C(O)CF₃ radical that will either decompose to give CO and a CF₃ radical or react with O₂ to give a peroxy radical:

CF₃C(O˙)ClC(O)CF₃ → CF₃C(O)Cl + C(O)CF₃ (38)

C(O)CF₃ → CO + CF₃ (39)

C(O)CF₃ + O₂ → (O₂)C(O)CF₃ (40)

The alkyl radical, C(O)CF₃, will predominantly react with O₂ (97% under laboratory conditions, 99.5% in the atmosphere) viareaction (40) and the remainder (3% or 0.5%) will decompose viareaction (39).²² The peroxy radical formed in reaction (40) will then react with itself or another peroxy radical giving an alkoxy radical, which decomposes to give a CF₃ radical and CO₂:²²

(O₂)C(O)CF₃ + RO₂ → (O)C(O)CF₃ + RO + O₂ (41)

(O)C(O)CF₃ → CO₂ + CF₃ (42)

The CF₃ radical will react with O₂ forming a peroxy radical that can react with another peroxy radical to give an alkoxy radical and O₂:

CF₃ + O₂ → CF₃O₂ (43)

CF₃O₂ + RO₂ → CF₃O + RO + O₂ (44)

The formed alkoxy radical CF₃O can abstract a hydrogen atom from a hydrocarbon (RH) forming an alcohol. This alcohol will eliminate HF forming C(O)F₂:

CF₃O + RH → CF₃OH + R (45)

CF₃OH → C(O)F₂ + HF (46)

The peroxy radical CF₃O₂ and the alkoxy radical CF₃O can also react with each other forming the trioxide CF₃O₃CF₃ (bis-(trifluoromethyl)trioxide):

CF₃O + CF₃O₂ + M → CF₃O₃CF₃ + M (47)

Fig. 11 shows the formation of products versus the loss of CF₃CHClC(O)CF₃. The solid lines through the CF₃C(O)Cl data are linear least squares analyses of the data. The slopes give yields of CF₃C(O)Cl of (78 ± 8)% in the experiments where UVB lamps were used for Cl₂ photolysis and (102 ± 10)% when using UVA lamps to photolyze Cl₂. The dotted lines through the C(O)F₂ and CF₃O₃CF₃ data are trend lines to serve as visual aids for inspection of the data. As a measure of the combined yield of CF₃ radicals formed in the degradation of CF₃CHClC(O)CF₃, diamonds show the sum of the fractional formation of C(O)F₂ and twice the fractional formation of CF₃O₃CF₃. As described above the CF₃ radicals will eventually give either C(O)F₂ or CF₃O₃CF₃. CF₃O₃CF₃ will react on the walls of the chamber and form two C(O)F₂ molecules. This is evidenced in Fig. 11, by the decrease in the slopes of formation of CF₃O₃CF₃ and the increase in the slopes of the formation of C(O)F₂ during the course of the experiments. The solid lines through the diamonds in Fig. 11 are linear regressions giving yields of CF₃ radicals of (95 ± 10)% and (122 ± 13)% for the experiments using UVA and UVB lamps, respectively. UVB lights are expected to photolyze CF₃CHClC(O)CF₃, as has been observed for similar fluorinated ketones.^23,24 The products observed in the experiments using UVB lamps are therefore a combination of the reaction with Cl atoms and of the photolysis. The experiment using UVA lamps gives products of the reaction with Cl atoms only. Here, the yield of the reaction is indistinguishable from 100% of CF₃C(O)Cl and the co-formed CF₃ radical (observed as C(O)F₂ and CF₃O₃CF₃). In the experiments using UVB lamps, (78 ± 8)% of the loss of CF₃CHClC(O)CF₃ is via reaction with Cl atoms observed as CF₃C(O)Cl and the remaining is loss by photolysis. The photolysis of CF₃CHClC(O)CF₃ gives an alkoxy radical that will break into two radicals; CF₃CH and C(O)CF₃:

CF₃CHClC(O)CF₃ + hν(UVB) → CF₃CHC(O)CF₃ + Cl (48)

CF₃CHC(O)CF₃ → CF₃CH + C(O)CF₃ (49)

The CF₃CH radical will react with O₂ forming a peroxy radical and then proceed to react with another peroxy radical to form CF₃CHO:

CF₃CH + O₂ → CF₃CHO₂ (50)

CF₃CHO₂ + RO₂ → CF₃CHO + RO + O₂ (51)

CF₃CHO has a fast reaction with Cl giving a C(O)CF₃ radical:^25,26

CF₃CHO + Cl → C(O)CF₃ + HCl (52)

The C(O)CF₃ radical decomposition will follow the reaction pathway outlined in reactions (39)–(47) yielding a CF₃ radical that will react to give either C(O)F₂ or CF₃O₃CF₃. The total observed formation of CF₃ radicals in the experiments with UVB lamps is 122% (as mentioned above). Of this, 78% is the co-product of CF₃C(O)Cl from the Cl atom reaction, leaving 44% CF₃ radicals formed from the photolysis of CF₃CHClC(O)CF₃. The photolysis of CF₃CHClC(O)CF₃ and the reaction with Cl atoms balances the mass in the experiments using UVB lamps. No change in the product yields was observed with an O₂ partial pressure varying between 140 and 700 Torr.

Fig. 11 Panel A: Formation of CF₃C(O)Cl (circles) versus the loss of CF₃CHClC(O)CF₃ in the presence of Cl atoms in a total of 700 Torr air using UVA (white) or UVB (gray) lamps and 700 Torr O₂ using UVB lamps (black). Panel B: C(O)F₂ (triangles), and CF₃O₃CF₃ (squares) versus the loss of CF₃CHClC(O)CF₃ in the presence of Cl atoms in a total of 700 Torr air using UVA (white) or UVB (gray) lamps and 700 Torr O₂ using UVB lamps (black). The diamond symbols are the sum of CF₃ radicals formed (2 × CF₃O₃CF₃ + C(O)F₂). The solid lines are linear fits to the data and the dotted lines are trend lines through the data to ease visual inspection of the data, see text for details.

3.11 Atmospheric lifetimes of Z- and E-CF₃CH=CHCF₃

Compounds such as Z- and E-CF₃CH=CHCF₃ can leave the atmosphere via photolysis, wet and dry deposition or reaction with atmospheric oxidants: NO₃ radicals, O₃, OH radicals and Cl atoms. Z- and E-CF₃CH=CHCF₃ does not absorb light in wavelengths <200 nm, so photolysis in the troposphere is not important.²⁷ Compounds such as Z- and E-CF₃CH=CHCF₃ are likely to be in the gas phase rather than aqueous phase, so wet deposition is not likely to be an important atmospheric sink. The volatility of both compounds will render dry deposition unlikely as an atmospheric removal mechanism. The reaction with NO₃ radicals and O₃ is too slow to be of significance when compared to that of reaction with OH radicals and Cl atoms.²⁸ Even though the rate coefficients for the reactions of Z- and E-CF₃CH=CHCF₃ with Cl atoms are two orders of magnitude faster than those of the reactions with OH radicals, the reaction with OH radicals is the main sink for Z- and E-CF₃CH=CHCF₃. The global average atmospheric concentration of Cl atoms is generally low; approximately [Cl] = 1 × 10³ atom cm⁻³,²⁹ which is 3 orders of magnitude lower than the average global of OH radical concentration of [OH] = 1 × 10⁶ molecule cm⁻³.³⁰ Locally Cl atom concentrations can be significantly higher, e.g. [Cl] = 1.8 × 10⁴ atom cm⁻³ in the marine boundary layer, so in some cases the Cl atom the reactions can compete with the OH radical reaction.³¹ The rate coefficients of the reactions with OH radicals determined here are at a temperature of 296 ± 2 K, but the appropriate temperature used to estimate atmospheric lifetimes based on OH radical reactions is 272 K.³² This needs to be considered when estimating the atmospheric lifetimes of the two compounds. Baasandorj et al. performed their study of the kinetics of the reaction of Z-CF₃CH=CHCF₃ + OH radicals at temperatures of 212–374 K and found an Arrhenius expression for temperatures <300 K giving k₁₉ = (5.19 ± 0.53) × 10⁻¹³ cm³ molecule⁻¹ s⁻¹ at 272 K, showing a slight temperature dependence.⁷ This value is indistinguishable from our value for k₁₉ determined at here at 296 K ± 2 K. It is assumed that the reaction of E-CF₃CH=CHCF₃ + OH will have a similar temperature dependency. Hence, we use the rate coefficients determined in this study at 296 ± 2 K to estimate atmospheric lifetimes of Z- and E-CF₃CH=CHCF₃ of 27 and 67 days, respectively. Baasandorj et al. estimated the lifetime of Z-CF₃CH=CHCF₃ as 22 days, in reasonable agreement with our present estimate.⁷

3.12 IR spectra, radiative efficiencies, and global warming potentials of Z- and E-CF₃CH=CHCF₃

The IR spectra in absorption cross section σ of Z- and E-CF₃CH=CHCF₃ are shown in Fig. 12. The integrated absorption cross sections of the two spectra (550–2000 cm⁻¹) have been determined to be (2.51 ± 0.13) × 10⁻¹⁶ and (2.96 ± 0.15) × 10⁻¹⁶ cm molecule⁻¹, for Z- and E-CF₃CH=CHCF₃, respectively.

Fig. 12 IR spectra of 1.30 mTorr Z-CF₃CH=CHCF₃ (panel A) and 2.25 mTorr E-CF₃CH=CHCF₃ (panel B). The insets show the linearity of the absorbance.

Using the method described by Pinnock et al.³³ the radiative efficiencies for Z- and E-CF₃CH=CHCF₃ were found to be 0.334 and 0.300 W m⁻² ppb⁻¹, respectively. The value for Z-CF₃CH=CHCF₃ is in agreement with the value estimated by Baasandorj et al. of 0.38 W m⁻² ppb⁻¹ using the same method.⁷ This method assumes that the gases are well-mixed in the atmosphere. For gases such as Z- and E-CF₃CH=CHCF₃ this is not the case since their atmospheric lifetime is not long enough for vertical mixing. Therefore we employ a correction factor, f(τ), dependent on the atmospheric lifetime of the compounds, τ, as described by Hodnebrog et al.,³⁴

where a, b, c, and d are constants with values of 2.962, 0.9312, 2.994, and 0.9302, respectively. The correction factors were calculated to be f(τ) = 0.21 and 0.38, for Z- and E-CF₃CH=CHCF₃ respectively. This gives effective values of the radiative efficiencies for Z- and E-CF₃CH=CHCF₃ of 0.069 and 0.113 W m⁻² ppb⁻¹, respectively. The radiative efficiencies and atmospheric lifetimes are used in the calculation of the global warming potentials (GWPs) of Z- and E-CF₃CH=CHCF₃ using the following equation,where F_x is the radiative efficiency of x (x = Z- or E-CF₃CH=CHCF₃), τ_x is the atmospheric lifetime of x with an exponential decay, t′ is the time horizon (typically 20, 100, and 500 years), F_CO2 is the radiative efficiency of CO₂, and R(t) is a response function that describes the decay of an instantaneous pulse of CO₂. The GWP is calculated relative to CO₂ mass for mass, and the denominator in eqn (IX) is the absolute global warming potential (AGWP) values of CO₂. These have been determined to be AGWP(CO₂) = 2.49 × 10⁻¹⁴, 9.17 × 10⁻¹⁴, and 32.2 × 10⁻¹⁴ W year m⁻² kg⁻¹, (= 0.196, 0.722, and 2.534 W year m⁻² ppb⁻¹) for 20, 100, and 500 year time horizons.³⁵ Thus we estimate the GWPs for Z-CF₃CH=CHCF₃ to be 6, 2, and 0 for time horizons of 20, 100, and 500 years, respectively. The GWPs for E-CF₃CH=CHCF₃ are estimated as 26, 7, and 2 for time horizons of 20, 100, and 500 years, respectively.

4. Conclusions and atmospheric impact

The present work provides a comprehensive description of the atmospheric chemistry and fate of Z- and E-CF₃CH=CHCF₃. The kinetics of the reactions of Z- and E-CF₃CH=CHCF₃ with Cl atoms, OH radicals, OD radicals, and O₃ have been determined. This is the first kinetic study of E-CF₃CH=CHCF₃ and the first determination of the rate coefficients of the reactions of Z-CF₃CH=CHCF₃ with Cl atoms and O₃. Using the obtained rate coefficients for the reactions with OH radicals we find that both Z- and E-CF₃CH=CHCF₃ have short atmospheric lifetimes of 27 and 67 days, respectively. GWP values for both compounds are small; 2 and 7 for the 100 year time horizon for Z- and E-CF₃CH=CHCF₃, respectively. The Cl atom initiated degradation mechanism of Z- and E-CF₃CH=CHCF₃ is summarized in Fig. 13. For both Z- and E-CF₃CH=CHCF₃ the primary degradation product is CF₃CHClC(O)CF₃. E-CF₃CH=CHCF₃ gives CF₃CHClC(O)CF₃ in a yield indistinguishable from 100% and Z-CF₃CH=CHCF₃ gives (95 ± 10)% CF₃CHClC(O)CF₃ and (7 ± 1)% E-CF₃CH=CHCF₃. CF₃CHClC(O)CF₃ reacts with Cl atoms to give the secondary product CF₃C(O)Cl in a yield indistinguishable from 100%, co-formed with a CF₃ radical observed as C(O)F₂ and CF₃O₃CF₃. Calculations were performed to investigate differences in the reactivities of Z- and E-CF₃CH=CHCF₃ towards OH and OD radicals and the observed isomerization of Z-CF₃CH=CHCF₃ giving E-CF₃CH=CHCF₃ in the experiments with Cl atoms. Energies of a pre-reaction complex, transition states and the formed alkyl radical were calculated for both alkenes for the reactions with OH and OD radicals as well as the energies of the alkyl radicals of the reaction of both alkenes with Cl atoms. The calculated energies are in good agreement with the experimental observations and in agreement with the observed reactivity trends. Based on the short atmospheric lifetimes, the small GWP values and the non-toxic products of the Cl atom initiated degradation, the atmospheric impact of both Z- and E-CF₃CH=CHCF₃ is negligible. Further studies of the OH radical initiated products are needed to assess the atmospheric impact of the products formed by atmospheric degradation.

Fig. 13 Cl atom initiated degradation mechanism of Z- and E-CF₃CH=CHCF₃. The observed products are indicated by the boxes.

Acknowledgements

The authors thank Rajiv R. Singh (Honeywell) for supplying us with a sample of Z-CF₃CH=CHCF₃. We thank Mogens Brøndsted Nielsen and Martyn Jevric (University of Copenhagen) for supplying samples of CD₃ONO and (CH₃)₂CHONO and for aid in the synthesis of CH₃ONO. Thanks are due to Timothy J. Wallington (Ford Motor Company) for kindly providing the absorption cross section of CF₃O₃CF₃.

References

1. M. J. Molina and F. S. Rowland, Nature, 1974, 249, 810–812.
2. F. S. Rowland, Ambio, 1990, 19, 281–292.
3. J. C. Farman, B. G. Gardiner and J. D. Shanklin, Nature, 1985, 315, 207–210.
4. T. Midgley and A. L. Henne, Ind. Eng. Chem., 1930, 22, 542–545.
5. DuPont, DuPont Develops HFO Foam Expansion Agent with Low Global Warming Potential, Additives for Polymers, 2014, 6, 2.
6. F. Molés, J. Navarro-Esbrí, B. Peris, A. Mota-Babiloni, Á. Barragán-Cervera and K. Kontomaris, Appl. Therm. Eng., 2014, 71, 204–212.
7. M. Baasandorj, A. R. Ravishankara and J. B. Burkholder, J. Phys. Chem. A, 2011, 115, 10539–10549.
8. E. J. K. Nilsson, C. Eskebjerg and M. S. Johnson, Atmos. Environ., 2009, 43, 3029–3033.
9. L. A. Curtiss, P. C. Redfern and K. Raghavachari, J. Chem. Phys., 2007, 127, 124105.
10. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, E. B. J. J. Heyd, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision A.02, Gaussian Inc., Wallingford CT, 2009.
11. L. A. Curtiss, P. C. Redfern and K. Raghavachari, Gaussian-4 Theory, J. Chem. Phys., 2007, 126, 084108.
12. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, E. B. J. J. Heyd, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision D.01, Gaussian Inc., Wallingford CT, 2009.
13. T. J. Wallington, J. M. Andino, I. M. Lorkovic, E. W. Kaiser and G. Marston, J. Phys. Chem., 1990, 94, 3644–3648.
14. R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F. Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi and J. Troe, Atmos. Chem. Phys., 2006, 6, 3625–4055.
15. R. Søndergaard, O. J. Nielsen, M. D. Hurley, T. J. Wallington and R. Singh, Chem. Phys. Lett., 2007, 443, 199–204.
16. M. D. Hurley, J. C. Ball and T. J. Wallington, J. Phys. Chem. A, 2007, 111, 9789–9795.
17. M. Mashino, Y. Ninomiya, M. Kawasaki, T. J. Wallington and M. D. Hurley, J. Phys. Chem. A, 2000, 104, 7255–7260.
18. L. L. Andersen, F. F. Østerstrøm, M. P. Sulbaek Andersen, O. J. Nielsen and T. J. Wallington, Chem. Phys. Lett., 2015, 639, 289–293.
19. M. P. Sulbaek Andersen, E. J. K. Nilsson, O. J. Nielsen, M. S. Johnson, M. D. Hurley and T. J. Wallington, J. Photochem. Photobiol., A, 2008, 199, 92–97.
20. R. J. Meagher, M. E. McIntosh, M. D. Hurley and T. J. Wallington, Int. J. Chem. Kinet., 1997, 29, 619–625.
21. Absorption cross section obtained from Dr T. J. Wallington, Ford Motor Company, private communication July 2016.
22. T. J. Wallington, M. D. Hurley, O. J. Nielsen and J. Sehested, J. Phys. Chem., 1994, 98, 5686–5694.
23. B. Ballesteros, E. Jiménez, A. Moreno, A. Soto, M. Antiñolo and J. Albaladejo, Chemosphere, 2017, 167, 330–343.
24. Y. Diaz-de-Mera, A. Aranda, A. Notario, A. Rodriguez, D. Rodriguez and I. Bravo, Phys. Chem. Chem. Phys., 2015, 17, 22991–22998.
25. M. P. Sulbaek Andersen, O. J. Nielsen, M. D. Hurley, J. C. Ball, T. J. Wallington, J. E. Stevens, J. W. Martin, D. A. Ellis and S. A. Mabury, J. Phys. Chem. A, 2004, 108, 5189–5196.
26. T. J. Wallington and M. D. Hurley, Int. J. Chem. Kinet., 1993, 25, 819–824.
27. J. G. Calvert, J. N. Pitts Jr., Photochemistry, John Wiley and Sons Inc., New York, 1966.
28. B. J. Finlayson-Pitts, J. N. Pitts Jr., Kinetics and Atmospheric Chemistry, Chemistry of the Upper and Lower Atmosphere, Academic Press, San Diego, 2000, ch. 5, pp. 130–178.
29. B. J. Finlayson-Pitts, J. N. Pitts Jr., Rates and Mechanisms of Gas-Phase Reactions in Irradiated Organic–NO_x–Air Mixtures, Chemistry of the Upper and Lower Atmosphere, Academic Press, San Diego, 2000, ch. 6, pp. 179–263.
30. R. Prinn, J. Huang, R. Weiss, D. Cunnold, P. Fraser, P. Simmonds, A. McCulloch, C. Harth, P. Salameh and S. O'Doherty, Science, 2001, 292, 1882–1888.
31. W. Allan, H. Struthers, D. C. Lowe and S.E. Mikaloff Fletcher, J. Geophys. Res.: Atmos., 2010, 115, 8.
32. C. M. Spivakovsky, J. A. Logan, S. A. Montzka, Y. J. Balkanski, M. Foreman-Fowler, D. B. A. Jones, L. W. Horowitz, A. C. Fusco, C. A. M. Brenninkmeijer, M. J. Prather, S. C. Wofsy and M. B. McElroy, J. Geophys. Res.: Atmos., 2000, 105, 8931–8980.
33. S. Pinnock, M. D. Hurley, K. P. Shine, T. J. Wallington and T. J. Smyth, J. Geophys. Res.: Atmos., 1995, 100, 23227–23238.
34. Ø. Hodnebrog, M. Etminan, J. S. Fuglestvedt, G. Marston, G. Myhre, C. J. Nielsen, K. P. Shine and T. J. Wallington, Rev. Geophys., 2013, 51, 300–378.
35. IPCC , Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013, p. 1535.
36. N. R. Greiner, J. Chem. Phys., 1968, 48, 1413.
37. G. Paraskevopoulos and W. S. Nip, Can. J. Chem., 1980, 58, 2146–2149.
38. T. Gierczak, M. Baasandorj and J. B. Burkholder, J. Phys. Chem. A, 2014, 118, 11015–11025.
39. V. C. Papadimitriou and J. B. Burkholder, J. Phys. Chem. A, 2016, 120, 6618–6628.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cp07234h

---