Kinetics of the Reaction of OH Radicals with Hydrogen Iodide Between 225 and 950 K

Abstract

Reaction OH + HI → I + H₂O (1) is an important atmospheric process transforming inactive HI into chemically active iodine atoms. In the present work, the reaction kinetics have been studied in a discharge fast-flow reactor coupled with an electron impact ionization mass spectrometer at nearly 2 Torr total pressure of helium and over a wide temperature range, T = 225–950 K. The reaction rate constant was determined both by a relative rate method (with the OH + Br₂ reaction as a reference) and by absolute measurements carried out under pseudo-first order conditions by monitoring the OH consumption kinetics in excess of hydrogen iodide. U-shaped temperature dependence was observed for the reaction rate constant, negative at low temperatures and positive at high temperatures. Recommended expression over the 225–950 K temperature range: k₁ = 1.13 × 10⁻¹¹ exp(354/T) + 6.93 × 10⁻¹¹ exp(−1010/T) cm³ molecule⁻¹ s⁻¹ or in the form of a modified Arrhenius expression, k₁ = 4.2 × 10⁻¹² × (T/298)^1.36 exp(666/T) cm³ molecule⁻¹ s⁻¹, with a total estimated uncertainty of 15% at all temperatures. The rate constant data obtained in this study are compared with the results of previous experimental works.

1. Introduction

Hydrogen iodide (HI) is the atmospheric reservoir of iodine atoms, which are believed to initiate ozone destruction cycles both in the stratosphere [1] and troposphere [2]. The reaction of HI with OH radicals

OH + HI → H₂O + I

(1)

is an important atmospheric process transforming inactive HI into chemically active iodine atoms.

The previous kinetic studies [3,4,5,6,7,8] of Reaction (1) can be divided into three groups. Early measurements [3,4] reported the reaction rate constant close to 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ at room temperature. In the next three studies [5,6,8], the reaction rate constant (k₁) was found to be three times higher (≈3 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹). Finally, Campuzano-Jost and Crowley [7] reported the reaction rate constant that was twice as high (≈6.5 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ at room temperature). The dependence of the reaction rate constant on temperature was investigated only in two studies: by Campuzano-Jost and Crowley [7] in the temperature range T = 246–353 K and by Khamaganov et al. [8] at two temperatures, 298 and 370 K.

The significant uncertainty regarding the reaction rate constant is also reflected in the divergent recommendations from different expert groups on the evaluation of kinetic data for atmospheric modeling. The current recommendation of the NASA Panel for Data Evaluation is k₁ = 3 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ (with an uncertainty factor of 2) at T = 298 K [9]. At the same time, IUPAC Task Group on Atmospheric Chemical Kinetic Data Evaluation recommends negative temperature dependence of k₁ = 1.6 × 10⁻¹¹ exp(440/T) cm³ molecule⁻¹ s⁻¹ at T = 240–360 K with the value of (7.0 ± 2.1) × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ at T = 298 K [10].

In our opinion, and as previously stated [8], the most likely reason for the discrepancy in the k₁ measurements is a problem in determining the absolute concentrations of HI in the reaction zone, given that hydrogen iodide can be lost during storage of prepared mixtures and also upon its introduction into a reactor. In the present work, we report the measurements of the rate constant of Reaction (1) in an extended temperature range (T = 225–950 K) using both the absolute method (with in situ measurements of the concentration of HI directly in the reactor) and the relative rate method where knowledge of [HI] is not required for calculation of k₁.

The high-temperature kinetic data from the present work are of potential interest for modeling the chemistry of hot near-source volcanic plumes, which are known to contain iodine species (HI and I) [11,12].

2. Materials and Methods

The rate constant of Reaction (1) was measured at a total pressure of 2 Torr of helium in a conventional discharge fast-flow reactor combined with a modulated molecular beam mass spectrometer (Balzers, QMG 420, Balzers Aktiengesellschaft, Liechtenstein) with electron impact ionization operating at 30 eV energy [13,14]. The chemical composition of the reactive system was monitored by sampling the gas-phase molecules from the flow reactor and detecting them by mass spectrometry. The reaction time was determined by the distance between the tip of the movable injector and the sampling cone of the mass spectrometer (Figure 1) and by the linear flow velocity in the reactor (900–3100 cm s⁻¹). The use of two different flow reactors allowed for a wide temperature range for kinetic measurements. The low-temperature reactor, used at T = 225–340 K, consisted of a Pyrex tube (45 cm long and 2.4 cm internal diameter, coated with halocarbon wax), with a jacket for circulating a thermostated liquid (ethanol). The second flow reactor used at high temperatures (330–950 K) consisted of an electrically heated uncoated quartz tube (45 cm long and 2.5 cm i.d.) with water-cooled attachments (Figure 1) [13].

OH radicals were produced in the movable injector through the fast reaction of hydrogen atoms with NO₂ ([NO₂] = (2–5) × 10¹³ molecule cm⁻³)

H + NO₂ → OH + NO

(2)

k₂ = (1.47 ± 0.26) × 10⁻¹⁰ cm³ molecule⁻¹ s⁻¹ (T = 195–2000) [15]. Two methods were used to generate H atoms: dissociation of H₂, diluted in He, in a microwave discharge, and reaction of fluorine atoms with H₂:

F + H₂ → H + HF

(3)

k₃ = 1.24 × 10⁻¹⁰ exp(−507/T) cm³ molecule⁻¹ s⁻¹ (T = 220–960 K) [16]. In the latter case, fluorine atoms were generated in a microwave discharge of trace amounts of F₂ in He. In order to reduce F atom reactions with a glass surface inside the microwave cavity, a ceramic (Al₂O₃) tube was inserted in this part of the injector. Mass spectrometric analysis showed that more than 95% of F₂ was dissociated in the microwave discharge.

OH radicals were monitored at m/z = 96/98 (HOBr⁺) after being transformed into HOBr with an excess of Br₂ ([Br₂] = (4–7) × 10¹³ molecule cm⁻³), added 5 cm upstream of the sampling cone (Figure 1):

OH + Br₂ → HOBr + Br

(4)

k₄ = 2.16 × 10⁻¹¹ exp(207/T) cm³ molecule⁻¹ s⁻¹ (T = 220–950 K) [17]. The chemical conversion of OH to HOBr by an excess of Br₂ was also used for the measurements of the absolute concentrations of the radicals, with the concentration of OH being determined from the fraction of Br₂ consumed: [OH] = [HOBr] = Δ[Br₂].

In some experiments, I₂ was used instead of Br₂ and the OH radicals converted to HOI in Reaction (5) were detected at m/z = 144 (HOI⁺):

OH + I₂ → I + HOI

(5)

k₅ = (2.1 ± 0.3) × 10⁻¹⁰ cm³ molecule⁻¹ s⁻¹ (T = 240–350 K) [10]. I₂ was introduced into the reactor by flowing helium through a column containing iodine crystals.

Gaseous HI was prepared by drying a 57% aqueous solution on phosphorus pentoxide and purified by low-temperature distillation to remove I₂ impurities. The monometrically prepared mixtures of HI in helium (2–5%) were stored in a blackened bulb at room temperature. Analysis of the mixtures by mass spectrometry showed that the content of the I₂ impurity was less than 0.2%. Furthermore, no degradation of the HI was observed during several days of storage of the mixtures.

The absolute concentrations of all stable species (Br₂, NO₂, F₂, H₂, HI) in the reactor were calculated from their flow rates from monometrically prepared mixtures. For the determination of HI concentration in the reactor, two additional methods were used, which employed a chemical titration of the same concentration of H atoms successively with excess Br₂, HI and I₂:

H + Br₂ → HBr + Br

(6)

k₆ = 7.06 × 10⁻¹¹ (T/298)^0.88 exp(182/T) cm³ molecule⁻¹ s⁻¹ (T = 220–950 K) [18],

H + HI → H₂ + I

(7)

k₇ = 7.4 × 10⁻¹¹ exp(−290/T) cm³ molecule⁻¹ s⁻¹ (T = 250–370 K) [19].

H + I₂ → HI + I

(8)

k₈ = 6.6 × 10⁻¹⁰ exp(−20/T) cm³ molecule⁻¹ s⁻¹ (T = 250–420 K) [19].

The consumed fractions of Br₂ and HI in Reactions (6) and (7), respectively, and [HI] produced in Reaction (8) were monitored by mass spectrometry. First, it was observed that the concentrations of HI consumed in Reaction (7) and formed in Reaction (8) were identical to within a few percent, as would be expected given that these quantities correspond to the same consumed concentration of H atoms. Secondly, this concentration of HI could be related to the concentration of Br₂, consumed in Reaction (6), which allowed the absolute calibration of HI signals using that of Br₂. Absolute HI concentrations measured by this method always agreed well (within 10%) with those calculated from flow rate measurements from a flask containing the HI/He mixture. The agreement between HI concentrations calculated from the flow measurements and HI concentrations measured independently in the reactor indicated that HI losses during its storage and in the gas handling lines used for the introduction of HI into the reactor were insignificant.

3. Results

3.1. Rate Constant of Reaction (1): Absolute Measurements

In the absolute measurements, the rate constant of Reaction (1) was determined under pseudo-first order conditions by monitoring the OH decays in excess of HI. The initial concentration of OH radicals was in the range (1–3) × 10¹¹ molecule cm⁻³; the ranges of HI concentrations at different temperatures are shown in

Table 1. Absolute measurements of the rate constant of Reaction (1): experimental conditions and results.

T (K)	[HI] a	k1 b	OH Source	OH Detection c	Reactor Surface d
230	0.76–9.94	5.58	F + H2 + NO2	HOBr	HW
240	0.79–13.2	4.64	F + H2 + NO2	HOBr	HW
247	0.6–12.8	4.97	F + H2 + NO2	HOBr	HW
255	0.69–12.1	4.70	H + NO2	HOBr	HW
260	0.59–11.2	4.59	F + H2 + NO2	HOBr	HW
275	0.62–11.8	4.51	H + NO2	HOBr	HW
280	0.72–13.0	4.09	F + H2 + NO2	HOBr	HW
298	0.65–12.6	3.87	F + H2 + NO2	HOBr	HW
315	0.69–13.1	3.94	F + H2 + NO2	HOBr	HW
330	0.93–12.6	3.77	H + NO2	HOBr	Q
335	0.53–11.7	3.73	H + NO2	HOI	Q
340	0.68–14.3	3.64	F + H2 + NO2	HOBr	HW
360	0.65–12.8	3.49	H + NO2	HOBr	Q
380	0.82–13.6	3.24	H + NO2	HOBr	Q
410	0.59–13.1	3.33	H + NO2	HOBr	Q
450	0.63–15.4	3.12	H + NO2	HOBr	Q
465	1.07–17.3	3.08	H + NO2	HOBr	Q
500	1.07–16.1	3.21	H + NO2	HOI	Q
540	0.72–9.00	3.14	H + NO2	HOBr	Q
580	0.63–15.2	3.26	H + NO2	HOBr	Q
630	0.87–13.7	3.37	H + NO2	HOBr	Q
680	0.89–13.2	3.50	H + NO2	HOBr	Q
720	0.39–6.86	3.53	H + NO2	HOBr	Q
775	0.52–9.85	3.82	H + NO2	HOI	Q
840	0.42–12.5	3.89	H + NO2	HOI	Q
950	0.62–6.50	3.90	H + NO2	HOI	Q

^a Units of 10¹² molecules cm⁻³. ^b Units of 10⁻¹¹ cm³ molecule⁻¹ s⁻¹; statistical 2σ uncertainty is ≤3%, total estimated uncertainty is 15%. ^c See text for OH detection. ^d HW: halocarbon wax; Q: uncoated quartz.

.

In most experiments, OH radicals were detected as HOBr, with Br₂ being added at the end of the reactor (see Section 2). However, at high reactor temperatures, a drop in the HI signal was observed upon addition of Br₂. Most likely, this is due to a chain reaction:

OH + Br₂ → HOBr + Br

Br + HI → HBr + I

(9)

k₉ ≈ 1.0 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ (T = 298 K) [20].

I + Br₂ → IBr + Br

(10)

Reaction (10) is relatively slow at room temperature (k₁₀ = 1.65 × 10⁻¹³ cm³ molecule⁻¹ s⁻¹ [21]); however, it can be much faster at elevated temperatures, where the only available literature data is k₁₀ = 5.7 × 10⁻¹³ cm³ molecule⁻¹ s⁻¹ at T = 450 K [22]. Although the change in HI concentration in the OH detection zone did not affect the measured OH kinetics and could be taken into account when calculating the actual HI concentration in the reaction zone, we have carried out additional k₁ measurements using I₂ instead of Br₂ and detecting OH as HOI at high temperatures (see Section 2).

The concentrations of hydrogen iodide and OH radicals in the reactor were monitored simultaneously as a function of reaction time. The consumption of HI was found to be insignificant (less than 15%), and the mean concentration of HI along the reaction zone was used in the calculations of k₁. Figure 2 displays typical examples of the OH kinetics observed in the presence of different concentrations of HI in the reactor: [OH] = [OH]₀ × exp(−k₁′ × t), where k₁′ = k₁ × [HI] + k_w is the pseudo-first-order rate constant and k_w represents the heterogeneous loss rate of OH radicals. The values of k₁′ were determined from the slopes of the straight lines in Figure 2. All the pseudo-first-order rate constants were corrected for axial and radial diffusion of OH radicals [23] with the diffusion coefficient of OH in He calculated as D₀ = 660 × (T/298)^1.85 Torr cm⁻² s⁻¹ [24]. The diffusion corrections accounted for <10% of the observed k′ values.

In Figure 3, the pseudo-first order rate constants, k₁′ = k₁ × [HI] + k_w, are traced against the concentration of HI. A linear least-squares fit of the k₁′ data as a function of [HI] at each temperature provides the rate constant of Reaction (1). All the results of absolute measurements of k₁ are summarized in Table 1. The estimated total uncertainty on the individual measurements of k₁ is approximately 15% and includes statistical error (≤3%) and those on the measurements of the flows, pressure, temperature, and absolute concentration of HI.

During the measurements at low temperatures, a phenomenon of degradation of the reactor surface was observed, which manifested itself as an increase in the rate of OH loss during the treatment of the reactor by the reaction. This effect is most likely due to contamination of the reactor surface by products of Reaction (1), as a result of the recombination of iodine atoms (in the gas phase or heterogeneous), leading to the formation of I₂ and its adsorption on the reactor wall. It was observed that the presence of HF in the reaction system “stabilizes” the reactor surface, thereby reducing the rate of heterogeneous loss of OH radicals and ensuring reproducibility of measurements. For this reason, in low-temperature experiments, Reaction (3) between F-atoms and H₂ was used as a source of H atoms needed for the generation of OH radicals by Reaction (2). As can be seen in Table 1, the results obtained using two sources of H atoms, as well as two methods for detecting OH radicals (HOBr vs. HOI), are in good agreement.

3.2. Rate Constant of Reaction (1): Relative Rate Measurements

Despite the apparently satisfactory results obtained for k₁ at low temperatures, for more certainty we carried out relative rate measurements (RRM1), which consisted of a rapid titration of OH in the gas phase with a mixture of HI and Br₂ (Reactions (1) and (4), respectively) and the measurements of HOBr yield as a function of the [HI]/[Br₂] ratio:

$$\left[\mathrm{H}\mathrm{O}\mathrm{B}\mathrm{r}\right]={\displaystyle \frac{k_4\left[{\mathrm{B}\mathrm{r}}_2\right]}{k_4\left[{\mathrm{B}\mathrm{r}}_2\right]+k_1\left[\mathrm{HI}\right]+k_w}}\times [\mathrm{O}\mathrm{H}{]}_0,$$

or after rearrangement,

$${\displaystyle \frac{{\left[\mathrm{O}\mathrm{H}\right]}_0}{\left[\mathrm{H}\mathrm{O}\mathrm{B}\mathrm{r}\right]}}-1={\displaystyle \frac{k_1\left[\mathrm{H}\mathrm{I}\right]}{k_4\left[{\mathrm{B}\mathrm{r}}_2\right]}}+{\displaystyle \frac{k_w}{k_4{[\mathrm{B}\mathrm{r}}_2]}}$$

At a constant concentration of Br₂, the second term in above equation is constant and k₁/k₄ can be determined as the slope of the linear dependence of ([OH]₀/[HOBr] − 1) versus [HI]/[Br₂] ratio. Initial concentration of OH was recorded as [HOBr]₀ in the absence of HI in the reactor, [OH]₀ = [HOBr]₀ ≈ 3 × 10¹¹ molecule cm⁻³. Reaction time was (0.011–0.015) s. Typical examples of the experimental data are shown in Figure 4.

Final values of k₁ (

Table 2. Relative measurements of the rate constant of Reaction (1): experimental conditions and results.

T (K)	[HI] a	[Br2] a	k1/k4 b	k1 c	Method d
225	0.11–4.11	2.23	1.046 ± 0.030	5.67	RRM1
240	0.08–3.97	1.93	0.957 ± 0.035	4.90	RRM1
247	0.14–3.99	1.52	0.948 ± 0.025	4.73	RRM1
263	0.13–3.78	1.94	0.911 ± 0.020	4.32	RRM1
265	0.09–3.99	2.03	0.917 ± 0.012	4.33	RRM1
290	0.08–4.4	1.95	0.900 ± 0.013	3.97	RRM1
300	0.03	0.02	0.878 ± 0.026	3.80	RRM2

^a Units of 10¹³ molecule cm⁻³. ^b statistical 2σ uncertainty is given. ^c Units of 10⁻¹¹ cm³ molecule⁻¹ s⁻¹; estimated total uncertainty is 20% and includes that on the reference reaction. ^d RRM1: OH titration with HI/Br₂ mixture; RRM2: HI and Br₂ consumption (see text).

) were calculated with k₄ = 2.16 × 10⁻¹¹ exp(207/T) cm³ molecule⁻¹ s⁻¹ which seems to be well established [17].

It should be noted that the relative rate method applied above does not require the measurement of absolute concentrations of OH radicals, minimizes the possible impact of secondary and heterogeneous reactions, but still relies on the measurement of [HI]. To definitively dispel any doubts about the possible influence of uncertainty in the measurement of absolute HI concentrations on current measurements of k₁, we carried out additional experiments at room temperature using a “conventional” relative rate method (still with Reaction (4) as a reference), which does not require measuring the absolute concentrations of the species involved (RRM2).

The experiments consisted of the monitoring of the consumption of HI and Br₂, simultaneously present in the reactor, in reactions with OH. The relative consumptions of HI and reference compound, Br₂, are defined by the rate constants of their reactions with OH:

$$\mathrm{l}\mathrm{n}{\displaystyle \frac{[\mathrm{H}\mathrm{I}{]}_0}{\left[\mathrm{H}\mathrm{I}\right]}}={\displaystyle \frac{k_1}{k_4}}\times \mathrm{l}\mathrm{n}{\displaystyle \frac{[{\mathrm{B}\mathrm{r}}_2{]}_0}{\left[{\mathrm{B}\mathrm{r}}_2\right]}}$$

where the expressions under the logarithm are the ratios of the compound concentration in the absence of OH radicals to that in the presence of OH radicals for a given reaction time. Initial concentrations of HI and Br₂ were 3 × 10¹¹ and 2 × 10¹¹ molecule cm⁻³, respectively, and that of OH radicals was varied, being ≤1.5 × 10¹² molecule cm⁻³. Resulting dependence of ln([HI]₀/[HI]) on ln([Br2]₀/[Br]) is shown in Figure 5.

The slope of the straight line in Figure 5 represents the k₁/k₄ ratio. The variations in the reaction time (as shown in Figure 5) had no impact on the results. Along with the good linearity of the data, this observation seems to indicate a negligible role of the possible secondary reactions, which could lead to additional consumption of the reactants, HI and Br₂. In fact, the only secondary reaction that could influence the measurements (at room temperature) is reaction (9) of Br atoms (formed in Reaction (4)) with HI, with a rate constant of about 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ at T = 298 K [20]. However, given the relatively low concentrations of Br₂ used in the experiments, its potential impact is expected to be negligible.

4. Discussion

The results of the current measurements of k₁ are presented in Figure 6 along with those from previous studies. The present measurements confirm the negative temperature dependence of the rate constant observed previously by Campuzano-Jost and Crowley [7] and Khamaganov et al. [8] in the temperature ranges 246–353 K and 298–370 K, respectively. As for the absolute values of k₁, at room temperature, our data are somewhat higher (by a factor of 1.15–1.4) than those reported in refs. [5,6,8], although the agreement can be considered acceptable, taking into account the experimental uncertainties of the measurements. However, the difference between our measurements and those of Campuzano-Jost and Crowley [7] at room temperature is significantly greater (our data are lower by a factor of 1.6). We have difficulty establishing the reason for this important divergence and are led to agree with the authors of the most recent study of Reaction (1) [8], who analyzed the experiments of Campuzano-Jost and Crowley [7] and concluded that there is no obvious explanation for a large difference between their results and those of the references [5,6,8].

Possible reasons for the discrepancies between different measurements of k₁ have been discussed in previous studies, most extensively by Campuzano-Jost and Crowley [7], and will not be repeated here. It can be noted that the main reason appears to be uncertainty regarding the concentration of HI in the reactor (directly affecting the measured value of the rate constant), associated with possible losses of HI during its storage and in the gas delivery lines used for introduction of HI into the reactor. In this regard, it is worth noting that this study is the only one in which HI was monitored and its concentration measured directly in the reaction zone. Furthermore, the results of the absolute measurements of k₁ at room temperature agree well with those by the relative rate method (RRM2), which does not involve the measurement of the absolute concentration of HI.

Campuzano-Jost and Crowley [7], discussing the results of flow tube measurements by Lancar et al. [6], mentioned that the vibrationally excited OH radicals formed upon OH generation through H + NO₂ reaction are not efficiently quenched by 1 Torr of He and may possibly lead to a reduction in the rate coefficient for Reaction (1). In current experiments as well, the reaction of H atoms with NO₂ was used for OH generation at a total pressure of 2 Torr of He. However, the impact of the vibrationally excited OH on the measurements of k₁ seems unlikely, since the excited OH(v) are rapidly deactivated in reaction with NO₂ in the OH-source zone (movable injector), i.e., before entering the main reactor. The rate constant of (6.4 ± 0.3) × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹ at T = 298 K [25] was reported for OH(v) + NO₂ relaxation reaction. In addition, OH(v) can also be deactivated upon collisions with the wall of the movable injector.

The present measurements, carried out over an extended temperature range, revealed a U-shaped temperature dependence of the reaction rate constant: k₁ decreases with temperature at low temperatures and has a positive temperature dependence at high temperatures. The solid line in Figure 6 represents a fit of the present experimental k₁ data to the sum of two exponential functions:

k₁ = 1.13 × 10⁻¹¹ exp(354/T) + 6.93 × 10⁻¹¹ exp(−1010/T) cm³ molecule⁻¹ s⁻¹

This expression reproduces all experimental data to within 6% and can be recommended from the present work in the temperature range T = (225–950) K with an estimated systematic independent of temperature uncertainty of 15%. Alternatively, the experimental data can also be described by a modified Arrhenius expression:

k₁ = 4.2 × 10⁻¹² × (T/298)^1.36 exp(666/T) cm³ molecule⁻¹ s⁻¹

Both expressions give almost identical results for k₁ in the temperature range of the study.

Interestingly, a similar temperature behavior was previously observed for the rate constant of the analogous reaction of OH radicals with HBr in the same temperature range (dashed line in Figure 6) [14]. Theoretical calculations showed that at low temperatures, the OH + HBr reaction proceeds via the formation of a collisional complex involving H atom tunneling, while at higher temperatures, it proceeds predominantly via direct H atom abstraction [26]. A similar reaction mechanism can be hypothesized for the title reaction, but its theoretical confirmation would be beneficial.

5. Conclusions

In this study, a low-pressure discharge flow reactor coupled with modulated molecular beam mass spectrometry was used to measure the rate constant (k₁) of the reaction of OH radicals with hydrogen iodide, an important atmospheric process that converts the reservoir species HI into reactive iodine atoms.

As a result of absolute and relative measurements carried out in a wide temperature range from 225 to 950 K, it was found that the reaction rate constant has a U-shaped temperature dependence, that is, it decreases with temperature at low temperatures and increases at high temperatures. Recommended expression over the 225–950 K temperature range: k₁ = 1.13 × 10⁻¹¹ exp(354/T) + 6.93 × 10⁻¹¹ exp(−1010/T) cm³ molecule⁻¹ s⁻¹ or in the form of a modified Arrhenius expression, k₁ = 4.2 × 10⁻¹² × (T/298)^1.36 exp(666/T) cm³ molecule⁻¹ s⁻¹, with a total estimated uncertainty of 15% at all temperatures.

The present kinetic measurements, realized for the first time over an extended temperature range, provide an experimental basis for theoretical developments and can be recommended for use in modeling atmospheric chemistry over a wide temperature range: from the high temperatures characteristic of hot volcanic plumes near emission sources to the lowest temperatures in the Earth’s atmosphere.