Temperature dependent kinetics of the gas-phase reaction of OH radicals with EMS

Using an improved smog chamber system, the temperature dependence of OH radical reaction with EMS was investigated over the temperature range of 297–346 K at 1.01×105 Pa pressure of air. The Arrhenius expression of the reaction was obtained for the first time. The mechanism of the reaction was also investigated.

A large number of volatile sulfur compounds are emitted into the atmosphere by natural and anthropogenic activities. These sulfur-containing compounds, including SO 2 , COS, H 2 S, alkyl sulfides and alkyl thiols play an important role in both the tropospheric and stratospheric budget of atmospheric gases. It is reported that there are certain relationships between organosulfur compounds and acidity precipitation in remote areas [1]. Sulfur-containing compounds, especially dimethyl sulfide (DMS, CH 3 SCH 3 ), are believed to be directly relevant to the formation of sulfur aerosol and cloud in marine air, and thus is potentially important in regulating cloud optical properties and climate on a global scale [2][3][4].
Among sulfur-containing compounds, reduced sulfur compounds including alkyl sulfide and alkyl thiol, especially DMS, have been a subject of intense scientific interest for many years. Articles on the atmospheric oxidation of various types of reduced sulfur compounds are given by Atkinson and Carter [5], Yin et al. [6], Atkinson [7,8], Tyndall and Ravishankara [9], Barnes el al. [10], and Wine et al. [11]. In these articles, the kinetics, products and mechanisms of atmospheric oxidations of the major reduced sulfur compounds in the atmosphere were studied and the importance of reduced sulfur compounds (especially DMS) in the atmosphere was discussed. In 2006, Barnes et al. [12] reviewed the studies on dimethyl sulfide (DMS), dimethyl sulfoxide (DMSO) and their oxidation in the atmosphere, which emphasize the importance of sulfur-containing compounds further.
The dominant fate of reduced sulfur compounds in the atmosphere is chemical transformation, including reactions with ozone, OH radicals (under day-time conditions) and NO 3 radicals (under night-time conditions). In our previous studies, we investigated the gas phase reaction of alkyl sulfide with ozone using our smog chamber system [13][14][15]. The results reveal that in the atmosphere alkyl sulfides are mainly scavenged by the reaction with OH radicals and NO 3 radicals. In the present work, to test our improved smog chamber system and expand our understanding of the atmospheric fate of alkyl sulfide, we conducted relative rate measurements of ethyl methyl sulfide (EMS, C 2 H 5 SCH 3 ) with OH radicals over the temperature range of 297-346 K at 1.01×10 5 Pa pressure of air.

Apparatus
The smog chamber used here is similar to that we reported in previous publication [14]. A few improvements were made for better determination. Hence just a brief description is provided here. The reactor used is actually a 100L FEP Teflon film bag housed in a light-tight box made of iron. At the two ends of the reactor, there is an inlet and an outlet made of Teflon, respectively, facilitating the introduction of reactants and sampling. The interlayer of the box is filled with basalt fiber to maintain the temperature. The chamber is equipped with a self-made system for temperature regulation within ±0.5 K from ambient temperature to 400 K.
The reaction chamber is also equipped with a gas chromatograph with flame ionization detection (GC/FID) (GC6820, Agilent Technologies) for detection of reactants and products in the reactor. A stainless steel gas valve coupled with a 0.5-mL steel loop was fixed on the GC for gas sampling. During the experiments, to avoid the great absorption of EMS on the steel tube of the valve, the steel tube was changed to Teflon tube, and the valve was heated to 430 K. The gas samples were collected from the chamber by a windtight syringe and injected into the GC coupled with a 15m DB-1 capillary column (0.53 mm i.d., 1.5 µm phase).

Procedure
The kinetic experiments for determining rate coefficients for the reaction of OH radicals with EMS were performed at a total pressure of 1.01×10 5 Pa synthetic air and six different temperatures (297, 310.5, 316, 326, 336 and 346 K). Cyclohexane, whose rate coefficient for the reaction with OH radicals is well established, was used as the reference compound in this investigation. The photolysis of hydrogen peroxide (H 2 O 2 ) was used as the OH radical source: The starting concentrations (calculated at the working temperature) of EMS and the reference organic compounds were typically in the range of about (0.7-2.6) ×10 15 molecule cm 3 and that of H 2 O 2 was approximately (1.48-2.6) ×10 15 molecule cm 3 . It was observed that for both gaseous and liquid organics, these calculated concentrations agreed within ±10% with the concentrations quantitatively measured by gas chromatography. In a typical experiment, the concentration of EMS and the reference compound in the chamber was measured every 5-8 min, and the experiment lasted for about 2 h.
The rate coefficient for the reaction of EMS with OH radicals was determined using a relative rate method in which the relative disappearance rates of EMS and the reference compound are monitored synchronously in the presence of OH radicals: Given that EMS and the reference are removed solely by the reaction with OH radicals, eq. (I) can be obtained in the following form where C EMS,0 and C ref,0 are the concentrations of EMS and the reference compound, respectively, at the beginning of the experiment; C EMS,t and C ref,t are the corresponding concentrations at time t; k s and k r are the rate coefficients for the reactions of OH radicals with EMS and the reference compound, respectively. The photo-stability of EMS and cyclohexane was established by irradiation of EMS-cyclohexane-air mixtures in the absence of OH radical precursors for a time period as long as that employed in the kinetic experiments, after switching on the lamps. Cyclohexane was not photolysed and its wall losses were within the variation coefficient for duplicate injections. For EMS, wall losses represented by reaction (3) accounted for approximately 7% of its measured decay under irradiation conditions at room temperature and the rate coefficient for wall losses (k wall ) decreased gradually to zero at the highest temperature employed in the experiments.
Thus, incorporating the first-order wall loss process of EMS into eq. (I) leads to eq. (II), In every experiment, to avoid the error caused by the slight difference in EMS wall losses, the wall decay of EMS was measured at the beginning of every experiment by recording 8-10 GC spectra prior to irradiation. According to eq. (II), a plot of ln(C EMS,0 /C EMS,t )  k wall ×t against ln(C ref,0 / C ref,0 )/k r gives straight lines directly yielding k s as the slope.

Reagents
Ethyl methyl sulfide was supplied by Aldrich with a stated purity of 99%. Cyclohexane with purity of 99% was purchased from Alfa Aesar. Ethanol and methanol were supplied by Beijing Beihua Fine Chemcials Co., Ltd with the purity of 99% and 99.5%, respectively. All liquid organics were thoroughly degassed through a series of freeze-pumpthaw cycles at 77 K to remove the impurities. The following gases were supplied by Beijing AP Gas Industry Co., Ltd: nitrogen (99.999%) and oxygen (99.999%). Synthetic air was the mixture of N 2 and O 2 with a volume ratio of 4:1. H 2 O 2 (30 wt%) from Beijing Beihua Fine Chemcials Co., Ltd. was condensed from 30% to about 80% by vacuum distillation.

Kinetics of OH radicals with ethanol
As a testing experiment, the OH rate constant for ethanol was obtained using the apparatus described above. The experiments were conducted at 298 K under 1.01×10 5 Pa and N 2 was used as the buffer gas. The use of cyclohexane (k cyclohexane+OH =6.7×10 12 cm 3 molecule 1 s 1 ) [16] and methanol (k methanol+OH =9.28×10 13 cm 3 molecule 1 s 1 ) [17] as reference compounds resulted in the OH + ethanol rate constant of (3.23±0.29) and (3.49±0.47)×10 12 cm 3 molecule 1 s 1 , respectively. For both reference compounds, the determination was conducted in three different ethanol concentration conditions. The combined plots, which are a modified version of eq. (I) (here in eq. (I), [EMS] is changed to [ethanol]), are shown in Figure 1. For each reference compound, only one plot derived from a typical experiment is listed here. A linear least-squares analysis of the plots gives the rate constants for the OH + ethanol reaction. Incorporating the uncertainties associated with the reference rate  constants used to derive the OH + ethanol rate constant yields a final value for k ethanol+OH of (3.26±0.35)×10 12 cm 3 molecule 1 s 1 . The results obtained in this work together with those found in the literatures for ethanol + OH radicals are summarized in Table 1.
As seen in Table 1, our result agrees well with the previous values obtained in absolute methods [18][19][20][21][22] and the result obtained by Wu et al. [23] by similar technique. We could believe that our smog chamber system is suitable for the simulation of gas phase reactions of OH radicals with VOC.

Kinetics of OH reaction with EMS
(1) Influences of adsorption and particles. Using the apparatus proved suitable above, experiments were conducted to investigate OH + EMS rate constant at 298 K and 1.01×10 5 Pa pressure. Initially, experiments performed under different concentrations of EMS, cyclohexane and H 2 O 2 (80 wt%) were repeated for more than 20 times, and repeated puzzling curving plots derived from eq. (II) were obtained. The concentration conditions of a part of these experiments are listed in Table 2 and the combined plots are shown in Figure 2.
A careful comparison was made to analyse the reasons for the puzzling curving plots. Comparison between the results of experiments 1-3 shows that H 2 O 2 concentration influences the extent to which the curving plots differ from the normal linear plots. However, if the conversion ratio of experiments 1-3 was considered, the agreement between these experiments at the same conversion ratio is good.
As shown in Figure 2, the initial part of each plot exhibits  good linearity and the agreement between these initial parts is very good. For each experiment, the plot derived always exhibits good linearity in the initial part and becomes more and more curving with the proceeding of the reaction. Further analysis reveals that if conversion ratio of the experiments was fixed, the higher the EMS concentration is, the larger the plots will differ from normal linear plot. It has been reported that aerosols have catalytic effects on the reaction kinetics of the VOCs [24][25][26]. In our experiments, because of the relatively high EMS concentration and large conversion rate, we could believe the accumulation of a nice bit of small particles in the reactor. Thus, the catalytic effect of particulate matter on the oxidation rate of EMS in the presence of OH radicals may lead to the puzzling error of the plots.
To minimize the catalytic effect of particles, the conversion ratio was controlled to less than 30%. At the same time, the steps mentioned above to control the adsorption were also taken into account. Experiments under these conditions were conducted at 297 K and 1.01×10 5 Pa, and the corresponding plot for a typical experiment is listed in Figure 3. The good linearity of the plot in Figure 3 proves the necessity and validity of the steps we adopted to solve the puzzling problem of the curving plots.
(2) Temperature dependence of OH radical reaction with EMS. The rate constants for the OH radical reaction with EMS were determined under the improved experiment conditions over the temperature range of 297-346 K. The results are listed in Table 3. The data given in Table 3 are plotted in Arrhenius form in Figure 4. Unweighted linear least-squares analysis of these data leads to the following Arrhenius expressions, k EMS = 7.25×10 11 (cm 3 s 1 molecule 1 )e (3.9±0.37)(kJ/mol)/RT . At room temperature (297 K), the rate constants we obtained are in good agreement with that reported by Wang  To the best of our knowledge, the activation energy of OH reaction with DMS is still not well characterized. Wallington et al. [29] reported a negative Arrhenius activation energy, while the studies of Wine et al. [11], Hynes et al. [28], Hsu et al. [30], and Abbatt et al. [31] report a positive Arrhenius activation energy. The latest IUPAC data evaluation recommends a positive Arrhenius activation energy. For OH radical reactions with both DMS and EMS, the observation of a positive activation energy can be taken as an indication of the dominance of the hydrogen abstraction reaction in the systems used to study the reaction, while measurement of a negative Arrhenius activation energy can be taken as an indication that the association reaction of adduct formation probably also contributes to the rate coefficient measurement under the conditions employed.
In this work, we observed a positive activation energy (3.9±0.37) kJ for OH radical reaction with EMS in the presence of O 2 , which indicates the dominance of the hydrogen abstraction reaction under our experimental conditions. Although the association reaction of adduct formation should not be excluded since the air was used as the buffer gas in our system, the relatively high value of the positive activation energy seems to imply that the adduct formation reaction contributes very little to the total reaction. At temperatures below 270 K, the results of Albu et al. [32] and Williams et al. [33] show that the rate constants for OH reaction with DMS in the presence of O 2 increase with the decrease of the temperature and the contribution of adduct formation reaction becomes more and more important with the decrease of the temperature. These results are generally in fair agreement with our results.

Atmospheric implication
In this work, an improved smog chamber system was tested and used to investigate the temperature dependence of the gas-phase reaction of the OH radicals with EMS. At 298 K, the constant of OH radicals with ethanol was determined and the results proved our smog chamber is suitable for the simulation of gas phase reactions of OH radicals with VOC.
Using the tested smog chamber system, the OH radical reaction with EMS was investigated over the temperature range of 297-346 K under 1.01×10 5 Pa pressure and the obtained Arrhenius expression is k EMS =7.25×10 11 (cm 3 s 1 molecule 1 )e (3.9±0.37)(kJ/mol)/RT . Investigation of the mechanism of the reaction reveals that at temperatures over 298 K, the H abstraction reaction is the major channel of the reaction.
During the initial experiments, it is found that when the conversion ratio of the EMS+ OH reaction is large enough, the rate constant increases with the proceeding of the reaction. It seems that the sulfate particles formed during the reaction has a catalytic effect on the reaction. In China, because of the wide use of coal as the energy source, the average content of sulfate aerosol is much higher than that in the Northern Hemisphere [34]. Reduced sulfur compounds and a significant fraction of the total mass of semi-volatile compounds may be absorbed on these particles and the potential catalytic effect of the particles may influence the atmospheric oxidation of these compounds.