Experimental Study of the Mesospheric Removal of NF3 by Neutral

Article pubs.acs.org/JPCA

Experimental Study of the Mesospheric Removal of NF3 by Neutral Meteoric Metals and Lyman‑α Radiation Anna Totterdill,† J.C. Gómez Martín,† Tamás Kovács,† Wuhu Feng,†,‡ and John M.C. Plane*,† †

School of Chemistry, University of Leeds, Leeds LS2 9JT, United Kingdom National Centre for Atmospheric Science, School of Earth and Environment, University of Leeds, Leeds LS2 9JT, United Kingdom

‡

ABSTRACT: NF3 is a potent anthropogenic greenhouse gas with increasing industrial usage. It is characterized by a large global warming potential due in part to its large atmospheric lifetime. The estimated lifetime of about 550 years means that potential mesospheric destruction processes of NF3 should also be considered. The reactions of NF3 with the neutral metal atoms Na, K, Mg and Fe, which are produced by meteoric ablation in the upper mesosphere, were therefore studied. The observed non-Arrhenius temperature dependences of the reactions between about 190 and 800 K are interpreted using quantum chemistry calculations of the relevant potential energy surfaces. The NF3 absorption cross section at the prominent Lyman-α solar emission line (121.6 nm) was determined to be (1.59 ± 0.10) × 10−18 cm2 molecule−1 (at 300 K). In the mesosphere above 60 km, Lyman-α photolysis is the dominant removal process of NF3; the reactions with K and Na are 1−2 orders of magnitude slower. However, the atmospheric lifetime of NF3 is largely controlled by reaction with O(1D) and photolysis at wavelengths shorter than 190 nm; these processes dominate below 60 km. of cosmic dust which enters the atmosphere daily.9 The most abundant of these metals which have been measured are Fe, Na, Mg, and K, which occur in layers between 80 and 105 km with peak concentrations of approximately 10 000, 5 000, 3 000, and 80 cm−3, respectively.10 To quantify the potential role of metal atoms in the atmospheric lifetime of NF3, the rate constants of the following reactions need to be measured (we use the convention throughout that k1 refers to the rate coefficient for reaction R1, etc.):

1. INTRODUCTION Perfluorinated compounds (PFCs) such as NF3 are potent greenhouse gases characterized by large global warming potentials (GWPs). These result from a combination of strong infrared absorptions in the atmospheric window and very long atmospheric lifetimes (τ). Generally, species defined as longlived have atmospheric lifetimes τ ≥ 5 years, but because of their remarkable inertness, PFCs can possess extraordinarily long lifetimes ranging from many hundreds to several thousands of years.1,2 NF3 is commonly recognized as an extremely potent greenhouse gas with reported 100 year GWPs between 10,800 and 17,0002−4 and a projected atmospheric lifetime of ∼550 years.5,6 NF3 is being used increasingly in industry as a replacement for banned PFCs which were utilized in processes such as chemical cleaning and circuit etching. Consequently, interest in NF3 has been sparked by a substantial increase in its atmospheric mixing ratio. Recent measurements4 yielded a 2008 mean global tropospheric concentration of 0.45 ppt increasing at a rate of 0.053 ppt yr−1 to give a present day estimated concentration of 0.72 ppt assuming no change in emission rate, which agrees well with a modeling simulation.2 The same group conducted studies on archived air samples and collected ice cores2 and found undetectable levels of NF3 prior to 1975, indicating the major source to be anthropogenic and the consequent rise in concentration to be due to the increased usage in industry. Species for which τ ≥ 300 years can be predominantly destroyed in the mesosphere.1,7 It is well-acknowledged that the major sinks of NF3 are photolysis by UV radiation and reaction with O(1D).6,8 In this paper we consider a number of additional mesospheric removal processes whose omission could potentially result in an overestimation of the atmospheric lifetime of NF3. First, the potential reactions between NF3 and the metal atoms that are released in this region by the ablation of 20−100 t © 2014 American Chemical Society

Δr H(298K) = − 223 kJ mol−1

Na + NF3 → NaF + NF2

(R1) K + NF3 → KF + NF2

−1

Δr H(298K) = − 235 kJ mol

Mg + NF3 → MgF + NF2

(R2)

Δr H(298K) = − 209 kJ mol−1 (R3)

Fe + NF3 → FeF + NF2

−1

Δr H(298K) = −193 kJ mol

(R4)

The significant exothermicities of these reactions are calculated from tabulated bond strengths.11 Here we present a kinetic study of reactions R1−R4 using both the pulsed laser photolysis−laser-induced fluorescence (PLP-LIF) and fast flow tube (FFT) techniques. The second NF3 loss process we examine is photolysis by Lyman-α photons (121.6 nm), which does not appear to have been studied previously. The solar Lyman-α emission line is very intense (more than 2 orders of magnitude stronger than surrounding VUV emission). Furthermore, it coincides with a “hole” in the Schumann−Runge absorption continuum of O2 Received: March 26, 2014 Revised: May 17, 2014 Published: May 19, 2014 4120

dx.doi.org/10.1021/jp503003e | J. Phys. Chem. A 2014, 118, 4120−4129

The Journal of Physical Chemistry A

Article

Figure 1. Schematic diagram of the pulsed laser photolysis−laser-induced fluorescence apparatus employed to study the reaction of NF3 with metal atoms.

Table 1. Transitions Used for Laser-Induced Fluorescence Detection of Na, K, Mg, and Fe metal atom

λ Nd:YAG (nm)a

laser dye

λpeak filter (nm)b

λ transition (nm)

transition

Na K Mg Fe

532 355 532 355

rhodamine 610 exalite 404 rhodamine 610 coumarin 503

589 (5) 765 (6) 285 (5) 250 (5)

589.0 766.5 285.2d 248.3d

32P3/2−32S1/2 42P3/2−42S1/2c 31P1−31S0 x5F05−a5D4

Dye laser pump wavelength. bInterference filter peak transmission; fwhm is in parentheses. cK(52P3/2−42S1/2) transition pumped at 404.4 nm and observed nonresonantly at 766.5 nm. dFrequency-doubling crystal employed. a

calibrated mass flow controllers (MKS, models 179A and 1259C). Reactions R1−R4 were initiated by multiphoton photolysis of the corresponding metal-atom precursor at 193 nm using a loosely focused excimer laser beam (Lambda-Physik, model Compex 102; pulse energy, 28−60 mJ; repetition rate, 5 Hz) passing through a quartz window. For reaction R1, sodium iodide (NaI) was placed in the side arm and heated to a temperature of 560 ± 5 K. A flow of N2 (150−250 sccm) carried the vapor into the cell. For reaction R2, potassium iodide (KI) was heated in the side arm to 530 ± 2 K, and for reaction R3 magnesium acetylacetonate (Mg(C5H7O2)2 or MgAcAc) was heated to between 403 and 453 K and maintained to within 5 K throughout the experiment. In the case of these three reactions, counterpropagating excimer and dye laser beams were employed, as shown in Figure 1. In contrast, for reaction R4 the dye and excimer laser beams were arranged to be orthogonal. Powdered ferrocene (Fe (C5H5)2) was placed in a sealed round-bottomed flask kept at 295 K, giving an equilibrium vapor pressure of 0.006 Torr. A small N2 flow (50−150 sccm) passed through the flask, carrying the ferrocene vapor into the stainless steel reaction chamber. The metal atoms were probed by the dye laser beam (Continuum Minilite II Nd:YAG pumped Sirah CBR-G-30 dye laser), using the LIF detection schemes in Table 1. The resulting LIF signal was measured by a photomultiplier tube after passing through an interference filter centered at the appropriate wavelength and recorded using a digital oscilloscope (LeCroy,

and so penetrates to below 80 km in the atmosphere.10 Owing to the negative vertical electron affinity of NF3 of about −2.1 eV (see below), dissociative electron attachment of thermal electrons, which is important in the case of other fluorinated species such as SF6,7 should not be a significant mesospheric loss process for NF3.

2. EXPERIMENTAL SECTION 2.1. Pulsed Laser Photolysis−Laser-Induced Fluorescence. The apparatus shown in Figure 1 consists of a cylindrical stainless steel reaction cell with four orthogonal horizontal arms and a fifth vertical side arm (radius, 0.8 cm; length, 12 cm). Different metal-atom precursors were placed in a temperaturecontrolled stainless steel boat located in one of the horizontal side arms and heated. The resulting vapor was entrained into a carrier gas flow (N2) passing over the boat and transported into the reaction chamber where it was mixed with larger flows of NF3 and N2. Brewster-angled quartz windows were fitted to the ends of the other three horizontal arms to admit the laser beams into the center of the chamber for photolysis and optical detection. The chamber was placed within a thermally insulated container which can be operated as a furnace, reaching temperatures up to 1100 K, or filled with dry ice for low-temperature measurements. The temperature was monitored by a K-type thermocouple, placed at the center of the reaction chamber. The reactor pressure was measured with calibrated capacitance manometers (MKS Baratron, model 226A) and controlled by a needle valve on the exit line to the pump. Flows were controlled using 4121



Article

Figure 2. Schematic diagram of the experimental setup for Lyman-α absorption measurements.

nm) via an interference filter with peak transmission at 121.6 and 20 nm fwhm (Princeton Research, type 122-N). Experiments were carried out at room temperature (295 ± 2 K). Materials. Reactant gas mixtures for the experiments were prepared on all-glass vacuum lines. The gases N2 (99.9999%, BOC), NF3 (99.99%, BOC) and He (99.9999%, BOC) were used without further purification. The metal-atom precursors sodium (98% Sigma-Aldrich), sodium iodide (Sigma-Aldrich 98%), potassium iodide (Sigma-Aldrich 99%), magnesium acetyl acetonate (Sigma- Aldrich 98%), and ferrocene (Sigma-Aldrich 98%) were purified under vacuum (heating where appropriate) for at least an hour before kinetic experiments commenced.

LT262). The transient LIF signal from the metal atom was monitored as a function of time by varying the delay between the excimer and the probing laser pulse using a delay generator controlled by a customized Lab-View program. A typical timeresolved LIF profile consisted of 100 delay steps and resulted from the average of five individual delay scans, in all of which the signal for a particular delay was the average of three laser shots. 2.2. Fast Flow Tube. The FFT apparatus used to study reaction R1 (Na + NF3) has been described in detail previously;12 therefore, only a brief description is provided below. The apparatus consists of a 37.5 mm ID, 1 m long stainless steel tube made up of cross-pieces and nipple sections connected by conflat flanges and sealed with copper gaskets. An aluminum oxide crucible housed within a tungsten basket heater containing pure sodium chips was located within the flow tube, downstream of the bath gas inlet, and was heated to approximately 500 K. The resulting Na vapor was then entrained in the flow of carrier gas N2 and carried downstream where it mixed with varying ratios of NF3 in N2. The total flow rate was varied between 2500 and 3500 sccm at total pressures between 1.2 and 4.0 Torr; the ensuing flow velocities were between 0.7 and 31 m s−1 and Reynold’s numbers Na > Mg > Fe, which is the inverse order of the metal atom ionization potentials: K (4.34 eV) < Na (5.14 eV) < Mg (7.64 eV) < Fe (7.90 eV). At first glance, this might suggest that a classic electron transfer (“harpoon”) mechanism governs these reactions. However, in spite of all four reactions being highly exothermic, even the K and Na reactions are about 2 orders of magnitude slower than their collision frequencies at 300 K. These reactions therefore do not appear to proceed by the classical harpoon mechanism. The reason for this is that the vertical electron affinity of NF3 is significantly negative. Figure 8 shows the potential energy curves

Figure 7. Arrhenius plots for the reactions of NF3 with Na (◇), K (■), Mg (▲), and Fe (○). The measurements of k1(Na + NF3) with the fast flow are indicated with a cross through the diamond.

Reactions R1 and R2 exhibit evidence for non-Arrhenius behavior; therefore, the rate coefficients were expressed as the sum of two Arrhenius terms which capture the high- and lowtemperature dependence of k1 and k2: k1(Na + NF3, 197−622 K) = (6.0 ± 4.1) × 10−10 exp( −(18.6 ± 3.8)/RT ) + (2.3 ± 1.4) × 10−11 exp( −(4.9 ± 1.2)/RT ) cm 3 molecule−1 s−1 k 2(K + NF3, 210−626 K) = (16.0 ± 5.4) × 10−10 exp( −(19.1 ± 1.5)/RT ) + (1.3 ± 0.3) × 10−11 exp( −(7.2 ± 0.5)/RT ) cm 3 molecule−1 s−1 These expressions were obtained by fitting the low-temperature rate coefficients (T < 500 K) to a single Arrhenius term and then fitting the residual (i.e., high-temperature component) to a second Arrhenius term. The reason for doing this sequential fit of two Arrhenius terms to the data is that the alternative procedure of fitting four parameters simultaneously produces physically unreasonable pre-exponential factors with very large uncertainties (this arises because of the limited number of data points and their uncertainties). As we discuss below, separating the hightemperature Arrhenius term yields some physical insight into the cause of the non-Arrhenius behavior. Reactions R3 and R4 have been fitted to a single-term Arrhenius expression:

Figure 8. Potential energy curves for NF3 and NF3− as a function of NF2−F distance, calculated at the B3LYP/6-311+g(2d) level. The neutral NF2−F curve (solid line) and ionic NF2−F− curve (short dashed line) are relaxed scans along the reaction coordinate, i.e., the minimum energy paths of the respective potential energy surfaces. The points on the ionic NF2−F− curve (long dashed line) are calculated at the geometry of the neutral NF2−F curve, thus providing the vertical electron affinity.

for NF3 and NF3− as a function of the NF2−F distance. These curves were calculated at the B3LYP/6-311+G(2d) level of theory. The curve for neutral NF2−F (solid line) is a relaxed scan along the minimum path of the potential energy surface (PES). The ionic NF2−F− curve (long dashed line) is calculated at the corresponding neutral geometry, so that the vertical separation between these curves is the vertical electron affinity at each N−F separation. In contrast, the NF2−F− curve (short dashed line) is a

k 3(Mg + NF3, 312−693 K) = (9.2 ± 0.5) × 10−10 exp( −(32.5 ± 0.4)/RT ) cm 3 molecule−1 s−1 4125



Article

Figure 9. Potential energy surfaces for Na, K, Mg, and Fe + NF3, calculated at the MP2/6-311+g(2d) level of theory. The contour labels indicate the energy in kilojoules per mole. These surfaces are for the case where the metal atom attack is collinear with one of the N−F bonds. The scan is therefore along the N−F−M (M = metal atom) linear coordinate, where rN−F and rM−F are varied. Note that the geometry of the NF2 moiety is frozen, which means that the surface does not represent the lowest possible energy path from reactants to products.

(M = metal atom) linear coordinate, where rN−F and rM−F are varied. Note that the geometry of the NF2 moiety is frozen in these scans, which means that the surface does not represent the lowest possible energy path from reactants to products. The PESs are calculated at the MP2/6-311+g(2d) level of theory, where the Møller−Plesset correlation energy correction is used to account for the switch from the covalent nature of the surface in the entrance channel to the ionic exit channel when the metal fluoride has formed. At each point on the PES a new initial guess for the Hartree−Fock wave function was generated. The four PESs exhibit late barriers in their exit channels. The barriers increase in the order K (15 kJ mol−1), Na (20 kJ mol−1), Mg (35 kJ mol−1), and Fe (36 kJ mol−1), which is the same order in which the rate coefficients decrease (Figure 7). Reactions with late barriers tend to be activated by vibrational excitation in a reactant bond corresponding to the reaction coordinate.21 In this case, sufficient excitation of the NF2−F stretching mode (ν3 =

relaxed scan along the reaction coordinate, which shows that the adiabatic electron affinity is positive (a detailed exploration of the geometry-dependent electron affinity of NF3 has been published recently by Matsuura and Ohoyama20 in a study of the reaction of NF3 with metastable Kr atoms). The vertical electron affinity is −2.1 eV at the equilibrium geometry of NF3; thus, electron transfer is unfavorable. In fact, the electron affinity becomes favorable (i.e., positive) only when the NF2−F bond has stretched from its equilibrium length of 1.38 Å in neutral NF3 to more than 1.6 Å. A second point to note in Figure 8 is that the neutral and ionic curves cross about 0.2 eV above the minimum of neutral NF3, so this amount of internal excitation should result in dissociative electron attachment leading to NF2 + F− when an electron transfer from a metal atom occurs. Figure 9 shows PESs for the four reactions. These surfaces illustrate the case where the metal atom attack is collinear with one of the N−F bonds. The scan is therefore along the N−F−M 4126



Article

907 cm−1)22 will make the electron affinity positive and allow the reaction to proceed via the longer-range harpoon mechanism.20 Inspection of the two Arrhenius terms in the expressions for k1 and k2 listed above shows that the difference between the activation energies in the two terms is (13.7 ± 5.0) kJ mol−1 for k1 and (11.9 ± 2.0) kJ mol−1 for k2. These differences correspond (within error) to one ν3 quantum (10.8 kJ mol−1). This suggests that the first term in each expression contains the probability of ν3 excitation accompanied by the large pre-exponential factor expected for a long-range electron transfer. Note that the Boltzmann population of NF3 (v3 > 0) increases from 1.3% at 300 K to 15.5% at 700 K. 4.2. Lyman-α Absorption Cross Section. Figure 10 illustrates the NF3 absorption cross section between 121.6 and

will be very slow at a typical temperature of 180 K in the upper mesosphere. The room-temperature rate coefficient for O(1D) + NF3 has been reported by several groups,8,26−28 ranging between (1.20 ± 0.25) × 10−11 cm3 molecule−1 s−127 and k = (2.55 ± 0.38) × 10−11 cm3 molecule−1 s−1.26 The reactive channels (rather than quenching) account for most of the O(1D) loss, with branching ratios between (0.83 ± 0.25)27 and (0.99 ± 0.01).28 Zhao et al.28 observed a weak negative temperature dependence, whereas Dillon et al.8 reported a temperature-independent value of (2.0 ± 0.3) × 10−11 cm3 molecule−1 s−1. When these two data sets are compared, there is little evidence of temperature dependence between 199−356 K. We therefore use here the result from Dillon et al.8 and assume 100% reactive loss of NF3. In the mesosphere above 80 km, NF3 will encounter relatively high concentrations of O(3P) and H atoms.10 Although the reaction with H is thermodynamically favorable ΔHr(298 K) = −326 kJmol−1

NF3 + H → HF + NF2

(R5) −1

11

this reaction has an energy barrier of 59 kJ mol at the CBSQB3 level of theory and so will be too slow at mesospheric temperature to affect the NF3 lifetime. Similarly, the reaction11 ΔHr(298 K) = 29 kJmol−1

NF3 + O → FO + NF2

(R6) −1

has a barrier of 148 kJ mol (CBS-QB3 level), and so can be discounted. The NF3 photolysis cross section over the atmospherically relevant spectral range is required for calculating the NF3 photolysis rate. The experimental cross sections illustrated in Figure 10 were therefore used with the following linear interpolation between 121.6 and 126.6 nm where there is no experimental data:

Figure 10. NF3 absorption cross section (left-hand ordinate) between 121.6 and 200 nm: Lyman-α from the present study (●), extrapolation (○), La Paglia and Duncan24 (■), and Papadimitriou et al.23 (□). The solid line is the solar spectral irradiance (right-hand ordinate). Note the intense Lyman-α emission line at 121.6 nm.

σ(121.6−126.6 nm) = 5.22 × 10−19λ + 6.18 × 10−17 cm 2 molecule−1

200 nm (the atmospherically significant range), obtained by combining the results of the present study with previous work.6,23,24 Dillon et al.6 reported NF3 cross sections between 184 and 226 nm. Their values are in excellent agreement with a more recent study by Papadimitriou et al.23 who measured the cross section over a wider wavelength range (185−250 nm). A comparison of the Lyman-α cross section at 121.6 nm from the present study with the measurements of La Paglia and Duncan24 suggests that the cross section decreases below 126.6 nm. Because of the lack of data between 121.6 and 126.6 nm, a linear extrapolation was applied to estimate the cross sections over this range (Figure 10). Note that the bond energy of NF2−F is 241 kJ mol−1,25 which corresponds to a photon threshold of 496 nm. We therefore assume that absorption at wavelengths below 200 nm leads to photodissociation of NF3.

(E6)

The solar photon flux as a function of wavelength (i.e., the actinic flux) is taken from a box model version of the threedimensional chemical transport model SLIMCAT.29,30 The photolysis rate J(NF3) is calculated using a scheme31 which employs a four-dimensional look-up table as a function of pressure from the surface to 10−5 hPa, temperature, column ozone, and zenith angle. The Na and O(1D) profiles are from a recent implementation of Na chemistry in the WACCM chemistry-climate model,32 and the profile of atomic K was measured at 54°N.33 The NF3 removal rates are plotted as a function of altitude in Figure 11, for the conditions of noon at 40°N latitude in January (50° solar zenith angle). The photolysis rates for solar radiation over the complete spectral range, and for the Lyman-α line only, are both illustrated in Figure 11. This shows that photolysis by Lyman-α is negligible below 50 km but becomes dominant above 66 km and contributes 75% of the total photolysis rate in the upper mesosphere (80−90 km). Photolysis is the dominant removal process throughout the stratosphere and mesosphere. Removal by O(1D) is nearly as important as photolysis in the stratosphere where the concentration of O(1D) is high because of the ozone layer. However, O(1D) removal becomes less important in the mesosphere and is overtaken by Na removal above 80 km. The metals clearly play a secondary role in this model scenario. However, the Na and K reactions will be the only

5. ATMOSPHERIC IMPLICATIONS To assess the impact of these results on the lifetime of NF3 in the atmosphere, we now compare the removal rates of the metal atom reactions and Lyman-α photolysis from the present study with previous work on removal by reaction of NF3 with O(1D) and photolysis at longer wavelengths. Note that because of the negative vertical electron affinity of NF3 (Figure 8), electron attachment of thermal electrons above 50 km should be a negligible loss process. We also do not consider further the reactions of Mg and Fe with NF3 because their rate coefficients 4127



■

removal processes at high latitudes during winter when photolysis rates become very small (note that the metal atom concentrations are highest in the winter mesosphere during polar night32). Whether this impacts on the overall atmospheric lifetime of NF3 depends on the circulation of NF3 from the tropical tropopause, where it will enter the stratosphere, to the upper mesosphere.10

5. CONCLUSIONS This study has examined the potential loss processes of NF3 in the mesosphere. The reactions of NF3 with Na, K, Mg, and Fe atoms, which occur in the mesosphere as a result of meteoric ablation, were studied over a range of temperatures. Although quite exothermic, the reactions were found to be much slower than their collision frequencies (particularly Mg and Fe + NF3). This appears to arise from the negative vertical electron affinity of NF3. Vibrational excitation along the F2N−F coordinate leads to a positive electron affinity, which may cause the observed nonArrhenius behavior in the Na and K reactions at higher temperatures, and is consistent with the late barriers on the potential energy surfaces of all four reactions. Nevertheless, at the low temperatures of the mesosphere, these reactions are not important routes for NF3 removal. The Lyman-α absorption cross section of NF3 was also measured, and this turns out to be the dominant removal process for the gas if it survives long enough to reach altitudes above 60 km. AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: 44-113-3438044. School of Chemistry, University of Leeds, Woodhouse Lane, Leeds LS2 9JT, UK. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Morris, R. A.; Viggiano, A. A.; Arnold, S. T.; Paulson, J. F. Chemistry of Atmospheric Ions Reacting with Fully Fluorinated Compounds. Int. J. Mass Spectrom. Ion Processes 1995, 149−150, 287− 298. (2) Arnold, T.; Harth, C. M.; Mühle, J.; Manning, A. J.; Salameh, P. K.; Kim, J.; Ivy, D. J.; Steele, L. P.; Petrenko, V. V.; Severinghaus, J. P.; et al. Nitrogen Trifluoride Global Emissions Estimated from Updated Atmospheric Measurements. Proc. Nat. Acad. Sci. U.S.A. 2012, 110, 2029−2034. (3) Robson, J. L.; Gohar, L. K.; Hurley, M. D.; Shine, K. P.; Wallington, T. J. Revised IR Spectrum, Radiative Efficiency and Global Warming Potential of Nitrogen Trifluoride. Geophys. Res. Lett. 2006, 33, L10817. (4) Weiss, R. F.; Muhle, J.; Salameh, P. K.; Harth, C. M. Nitrogen Trifluoride in the Global Atmosphere. Geophys. Res. Lett. 2008, 35, L20821. (5) Fthenakis, V.; Clark, D. O.; Moalem, M.; Chandler, P.; Ridgeway, R. G.; Hulbert, F. E.; Cooper, D. B.; Maroulis, P. J. Life-Cycle Nitrogen Trifluoride Emissions from Photovoltaics. Environ. Sci. Technol. 2010, 44, 8750−8757. (6) Dillon, T. J.; Horowitz, A.; Crowley, J. N. Cross-Sections and Quantum Yields for the Atmospheric Photolysis of the Potent Greenhouse Gas Nitrogen Trifluoride. Atmos. Environ. 2010, 44, 1186−1191. (7) Ravishankara, A. R.; Solomon, S.; Turnipseed, A. A.; Warren, R. F. Atmospheric Lifetimes of Long-Lived Halogenated Species. Science 1993, 259, 194−199. (8) Dillon, T. J.; Vereecken, L.; Horowitz, A.; Khamaganov, V.; Crowley, J. N.; Lelieveld, J. Removal of the Potent Greenhouse Gas NF3 by Reactions with the Atmospheric Oxidants O(1D), OH and O3. Phys. Chem. Chem. Phys. 2011, 13, 18600−18608. (9) Vondrak, T.; Plane, J. M. C.; Broadley, S.; Janches, D. A Chemical Model of Meteoric Ablation. Atmos. Chem. Phys. 2008, 8, 7015−7032. (10) Plane, J. M. C. Atmospheric Chemistry of Meteoric Metals. Chem. Rev. (Washington, DC, U.S.) 2003, 103, 4963−4984. (11) Lide, D. R. Handbook of Physics and Chemistry, 87th ed.; CRC Press: Boca Raton, FL, 2006 (12) Plane, J. M. C.; Whalley, C. L. A New Model for Magnesium Chemistry in the Upper Atmosphere. J. Phys. Chem. A 2012, 116, 6240− 6252. (13) Lu, H. C.; Chen, H. K.; Chen, H. F.; Cheng, B. M.; Ogilvie, J. F. Absorption Cross Section of Molecular Oxygen in the Transition E 3Σ−u v = 0 − X 3Σ−g v = 0 at 38 K. Astron. Astrophys. 2010, 520, A19. (14) Alkemade, U.; Homann, K. H. Gas-phase Dimerization of Sodium Halides and Formation of Mixed Sodium Phenyl Iodide. Bunsen-Ges. Phys. Chem., Ber. 1989, 93, 434−439. (15) Vileno, E.; LeClair, M. K.; Suib, S. L.; Cutlip, M. B.; Galasso, F. S.; Hardwick, S. J. Thermal Decomposition of NF3 with Various Oxides. Chem. Mater. 1996, 8, 1217−1221. (16) Howard, C. J. Kinetic Measurements Using Flow Tubes. J. Phys. Chem. 1979, 83, 3−9. (17) Broadley, S. L.; Vondrak, T.; Plane, J. M. C. A Kinetic Study of the Reactions of Ca+ Ions with O3, O2, N2, CO2 and H2O. Phys. Chem. Chem. Phys. 2007, 9, 4357−4369. (18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A. et al. Gaussian 09, revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (19) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson, G. A. A Complete Basis Set Model Chemistry. VI. Use of Density Functional Geometries and Frequencies. J. Chem. Phys. 1999, 110, 2822−2827. (20) Matsuura, Y.; Ohoyama, H. Collision-Induced Harpooning Observed in the Excimer Formation in the Orientated NF3 + Orientated Kr*(3P2, MJ = 2) Reaction. J. Phys. Chem. A 2011, 115, 4583−4591. (21) Smith, I. W. M. Kinetics and Dynamics of Elementary Gas Reactions; Butterworths: London, 1980. (22) Shimanouchi, T. Tables of Molecular Vibrational Frequencies. In Nat. Stand. Ref. Data Ser.; National Bureau of Standards: Washington, DC, 1972; Vol. 39; pp 164.

Figure 11. First-order removal rates of NF3 by different processes (solid line, full spectral range; solid line with symbols, Lyman-α only; short dash−dot, reaction with O(1D); dashed line, reaction with Na; dashed− dotted line, reaction with K).

■

Article

ACKNOWLEDGMENTS

This work was part of the MAPLE project (PI: Professor Martyn Chipperfield) funded by Research Grant NE/J008621/1 from the UK Natural Environment Research Council, which also provided a studentship for A.T. 4128



Article

(23) Papadimitriou, V. C.; McGillen, M. R.; Fleming, E. L.; Jackman, C. H.; Burkholder, J. B. NF3: UV Absorption Spectrum Temperature Dependence and the Atmospheric and Climate Forcing Implications. Geophys. Res. Lett. 2013, 40, 440−445. (24) La Paglia, S. R.; Duncan, A. B. F. Vacuum Ultraviolet Absorption Spectrum and Dipole Moment of Nitrogen Trifluoride. J. Chem. Phys. 1961, 34, 1003−1007. (25) Christophorou, L. G.; Olthoff, A. J. K.; Rao, M. V. V. S. Electron Interactions with CF4. J. Phys. Chem. Ref. Data 1996, 25, 1341−1388. (26) Baasandorj, M.; Hall, B. D.; Burkholder, J. B. Rate Coefficients for the Reaction of O(1D) with the Atmospherically Long-lived Greenhouse Gases NF3, SF3CF3, CHF3, C2F6, c-C3F8. Atmos. Chem. Phys. 2012, 12, 11753−11764. (27) Sorokin, V. I.; Gritsan, N. P.; Chichinin, A. I. Collisions of O(1D) with HF, F2, XeF2, NF3, and CF4: Deactivation and Reaction. J. Chem. Phys. 1998, 108, 8995−9003. (28) Zhao, Z.; Laine, P. L.; Nicovich, J. M.; Wine, P. H. Reactive and Nonreactive Quenching of O(1D) by SO2F2, NF3, and SF5CF3. Proc. Nat. Acad. Sci. U.S.A. 2010, 107, 6610−6615. (29) Chipperfield, M. P. New Version of the TOMCAT/SLIMCAT Off-Line Chemical Transport Model: Intercomparison of Stratospheric Tracer Experiments. Q. J. R. Meteorol. Soc. 2006, 132, 1179−1203. (30) Feng, W.; Chipperfield, M. P.; Davies, S.; Sen, B.; Toon, G.; Blavier, J. F.; Webster, C. R.; Volk, C. M.; Ulanovsky, A.; Ravegnani, F.; et al. Three-dimensional Model Study of the Arctic Ozone Loss in 2002/ 2003 and Comparison with 1999/2000 and 2003/2004. Atmos. Chem. Phys. 2005, 5, 139−152. (31) Lary, D. J.; Pyle, J. A. Diffuse Radiation, Twilight and Photochemistry. J. Atmos. Chem. 1991, 13, 373−392. (32) Marsh, D. R.; Janches, D.; Feng, W.; Plane, J. M. C. A Global Model of Meteoric Sodium. J. Geophys. Res. 2013, 118, 11,442−11,452. (33) Eska, V.; von Zahn, U.; Plane, J. M. C. The Terrestrial Potassium Layer (75−110 km) between 71° S and 54° N: Observations and Modeling. J. Geophys. Res. 1999, 104, 17173−17186.

4129


Experimental Study of the Mesospheric Removal of NF3 by Neutral

Recommend Documents