Kinetic Measurements of Cl Atom Reactions with C 5 –C 8 Unsaturated Alcohols

: The reactions of five structurally similar unsaturated alcohols, i.e., ( Z )-2-penten-1-ol, ( E )-2-hexen-1-ol, ( E )-3-hexen-1-ol, ( Z )-3-hexen-1-ol, and 1-octen-3-ol, with Cl atoms in the gas phase, were investigated at 296 ± 2 K and 1 atm by the relative-rate kinetic technique using a 600-L Teflon reaction chamber. Selected ion flow tube mass spectrometry (SIFT-MS) was used simultaneously to monitor the decay of the alcohols of interest and selected reference compounds. Tetrahydrofuran (THF), propan-1-ol, and octane were used as reference compounds. Chlorine atoms were produced by the photolysis of molecular chlorine (Cl 2 ) using broadband actinic lamps near 365 nm. The estimated rate constant values (in 10 − 10 cm 3 ∙ molecule − 1 ∙ s − 1 ) followed the order 2.99 ± 0.53 (( Z )-2-penten-1-ol) < 3.05 ± 0.59 (( E )-3-hexen-1-ol) < 3.15 ± 0.58 (( Z )-3-hexen-1-ol) < 3.41 ± 0.65 (( E )-2-hexen-1-ol) < 4.03 ± 0.77 (1-octen-3-ol). The present work provides the first value of the rate constant for the reaction of 1-octen-3-ol with Cl atoms. The results are discussed and interpreted in relation to other studies where literature data are available. The structure–activity relationship and the atmospheric implications are discussed as well.


Introduction
Volatile organic compounds (VOCs) are trace gases generally present in the troposphere at sub-ppb concentrations. Yet, they play a major role in the chemistry of the troposphere through the formation of free radicals, ozone, and potentially secondary organic aerosols (SOA) [1,2]. It is recognized that the amount of VOCs emitted from biogenic sources (BVOC) largely exceeds that released from anthropogenic sources (AVOC) [3,4]. Green leaf volatiles (GLV), such as unsaturated alcohols [5], are BVOCs which are emitted from vegetation either spontaneously or under stress conditions [6]. Moreover, their emissions may increase due to climate change [7]. Despite this, the atmospheric reactivity of these compounds is less studied with respect to other BVOCs like monoterpenes.
Several biological materials like grasses, grains, fruits, dried fruits, vegetables, trees, and shrubs were shown to be a source of (Z)-3-hexen-1-ol and a series of (E)-3-alken-1-ols in the atmosphere [8][9][10]. Unsaturated alcohols with longer carbon chains such as 1-octen-3-ol were detected in grass emissions such as clover [11,12] and fungi [13]. Some unsaturated alcohols such as (Z)-3-hexen-1-ol and 1-octen-3-ol released by wheat, oat, and barley grains are thought to be insect attractors [14]. Biotic and abiotic stress factors lead to plant damage and stimulate emissions of higher concentrations of unsaturated alcohols in the troposphere [8,15,16]. In the Austrian Alps, (Z)-2-penten-1-ol was detected at a few-ppb level and its presence in such an area was attributed to the peroxidation of fatty acids occurring as a result of freeze damage on leaves [17].
Once in the atmosphere, unsaturated alcohols can be removed via reactions with atmospheric oxidants such as OH and NO3 radicals, O3, and Cl atoms [18]. The average concentration of Cl atoms in the troposphere is on the order of 10 3 -10 4 atoms•cm −3 but concentrations as high as 10 5 atoms•cm −3 were observed at dawn in the marine boundary layer. At such concentrations, the Cl-driven chemistry might play an important role in VOC oxidation in the troposphere [19][20][21][22]. In fact, as the OH concentration is on the order of 10 6 radicals•cm −3 [23], Cl could present a competitive oxidation process for various VOCs especially during morning hours, at the times of maximum rates of Cl-precursor photolysis [24,25]. Recent findings suggest that chlorine atom reactivity in the atmosphere could be important even far inward from the coasts [26,27], where ClNO2 pools are highly available as important Cl atom reservoirs [28,29]. Moreover, for some specific sites, like landfills, the impact of chlorine atoms on the atmospheric oxidation capacity might bring up to 10 times higher contributions to the local processes [25]. The formation of chlorine atoms in the atmosphere proceeds mainly through the oxidation of HCl by OH radicals and the photolysis of photolabile chlorine reservoir compounds (Cl2, ClNO2, and BrCl) [28][29][30]. Heterogeneous reactions involving sea salt and ozone are also a source of Cl atoms in the marine boundary layer [31,32]. Over the last two decades, although several experimental and theoretical works debated the potential impact of Cl oxidation with numerous VOCs in the tropospheric chemistry [33][34][35][36][37], the reactivity of some unsaturated alcohols is still not well established.
The aim of the present study was to determine the rate constants for the reactions of the following series of unsaturated alcohols with Cl atoms: 1-octen-3-ol + Cl → Products.
The reactivity of these unsaturated alcohols with Cl atoms was scarcely investigated [18]. The kinetics of (Z)-2-penten-1-ol with Cl atoms was solely determined in 2010 in a 400-L Teflon bag, using off-line gas chromatography with a flame ionization detector (GC-FID) and following a relative-rate kinetic approach [38]. For unsaturated hexenols, the rate constants for (E)-2-hexen-1-ol, (E)-3-hexen-1-ol, and (Z)-3-hexen-1-ol were determined only once also using the relative-rate kinetic approach in a 1080-L quartz glass simulation chamber with in situ Fourier-transform infrared (FTIR) spectroscopy [39]. No data are available regarding the rate constant of 1-octen-3-ol with chlorine atoms. The present work represents the first study on the relative-rate constant kinetic approach for the 1-octen-3-ol and Cl reaction and the second for (Z)-2-penten-1-ol and the series of hexenols cited above. Moreover, it has to be emphasized that, to our knowledge, the present kinetic measurements were performed for the first time using a selected ion flow tube mass spectrometer (SIFT-MS). The SIFT-MS instrument was used to measure the time decay of the above-listed compounds. A consistent determination of the reaction kinetics will help to better understand the atmospheric impacts of such compounds and develop structure-activity relationships for atmospheric chemistry models.

Experiments
Rate constants were measured at room temperature (296 ± 2 K) in zero air at atmospheric pressure (760 Torr) using the Thalamos simulation chamber which is a 600-L cubic Teflon photoreactor ( Figure S1, Supplementary Materials). More details of the experimental set-up can be found elsewhere [40], and only a brief description is given hereafter. The Teflon bag of the chamber is fixed inside a temperature-controlled box which can be regulated between 231 and 453 K. The chamber is surrounded by a set of 20 ultraviolet (UV) lamps emitting around 365 nm (PL-L 24W/10/4P, Philips, Netherlands). The lamps were used to produce Cl atoms through the photolysis of Cl2. The chamber is equipped with inlet and outlet ports used for reactant injection and reaction mixture sampling, respectively. Two fans are mounted inside the reactor in order to ensure the homogeneity of the gas mixture.
In the present work initial concentrations of reactants were from 0.7 to 1.5 ppm for both alcohols and reference compounds and around 1.7 ppm for molecular chlorine. Such high organic concentrations are not expected to induce secondary chemistry, since reactions of organic peroxy radicals generated as intermediates in atmospheric chemical mechanisms react very slowly with organic compounds [41]. The reaction mixture was continuously pumped through a sample loop using a dry Teflon diaphragm pump and reinjected into the reaction chamber with a flow rate of 4 L•min −1 .
A Voice200 ® SIFT-MS (Syft Technologies, New Zealand) instrument connected to the loop sampled the chamber at 10 mL•min −1 , allowing the decays of the reactants (alcohols and reference compounds) to be recorded. The SIFT-MS used in this study contains technical features similar to those from its first design [42] and all additional descriptions can be found in other recent papers [43][44][45][46]. Details on the SIFT-MS efficiency in measuring unsaturated alcohols (including those investigated in the present study) and about the ion reaction mechanisms were reported by Schoon et al. [47]. The detection limit of the commercially available SIFT-MS is at about pptv levels with a time resolution of 1 s.
The SIFT-MS measurement process is started by the production of reagent ions (R + : H3O + , NO + , and O2 + ) in a microwave discharge of a mixture of zero air and water vapor. The reagent ions are focused on a first quadrupole mass filter. Every few milliseconds, a single ionic species is selected based on its mass-to-charge ratio (m/z). Then, the selected ion passes to the flow tube reactor in a stream of carrier gas (He at 1 Torr). In addition, the Thalamos gas sample enters into this flow tube and undergoes a soft ionization reaction depending on the chemical properties of its components (A) (either alcohol or reference compound in this work) such as proton affinity and ionization energies. The reaction between the reagent ions (R + ) and the components (A) generates ion products (P + ) and neutral products (N). These processes involve (i) ionization by H3O + occurring mainly via proton transfer and water elimination, (ii) NO + reactions proceeding in different ways, such as charge transfer, hydride and/or hydroxide ion transfer, and/or three body association, and (iii) O2 + reactions generally occurring via charge transfer [43,47]. All these chemical species (R + , A, P + , and N) are injected into a second quadrupole mass filter in order to select ions based on their m/z. An electron multiplier detector placed at the exit of the second quadrupole is used to measure the count rates of the ions in the selected m/z range. In the present work, for SIFT-MS measurements, at least 10,000 counts were set as the limit before the instrument switched from one mass to another and the dwell time was on the order of one second. In addition, the mass spectrometer was internally calibrated on a daily basis with the help of a standard mixture of nine compounds provided by the supplier (Scotty, UN1956, Interscience, Belgium). The internal calibration procedure and the known reaction rate constants of the precursor ions with the analytes allow estimating the theoretical concentrations of any organic compound.
Table S1 (Supplementary Materials) summarizes all the product ions (P + ) observed for alcohols and references. To determine the relative rate constants, pairs of representative product ions from the used alcohol and reference compounds were selected and used to plot the reactant decays according to the relative-rate kinetic approach. The ions that could be used for the kinetic analysis were selected based on the absence of any possible interference arising from other chemical processes happening simultaneously with the studied reaction. Therefore, a series of tests were carried out before any experiment, where only the compound of interest (either the alcohol or the reference) and Cl2 were present and mixed with purified air in the Thalamos chamber. Gas samples were analyzed by SIFT-MS, selecting all the product ions arising from alcohols and references. The alcohol and reference product ions displaying interferences with reaction products ions were excluded from the analysis.
Each experiment started by cleaning the reaction chamber at least for 2 h at 50 °C using a flow rate of about 20 L•min −1 of zero air. Once the temperature of the reactor dropped to room temperature, the cleanliness of the reactor was evaluated by measuring the background levels of the selected ions for both alcohol and reference compound using the SIFT-MS system over 10 min. The background levels for all the selected ions never exceeded 5% of their initial experimental concentrations. Wall losses of the alcohols and the reference compounds in the dark were investigated in preliminary experiments, and negligible contributions were determined for these processes. The coefficient of variation (CV, expressed as the ratio of one standard deviation to the mean value of 40-min measurements) for the recorded signals showed <2.7% values. The reported values actually reflect small fluctuatios in the performances of the instrument such as small variations in the precursor ion concentrations, temperature of the instrument core, etc. The stability of the alcohols and reference compounds under UV-A irradiation was also tested, and all compounds were shown to be stable (CV < 2.8% over 30 min). The studied alcohol was injected in the chamber. The ion signal delivered by the SIFT-MS apparatus was recorded for 30 to 45 min until signal stabilization was achieved. The reference compound was then added to the reaction mixture and left to stabilize in the dark for 15 min. The precursor of chlorine atoms (Cl2) was finally introduced into the reactor, and the mixture was left again for 20 min in the dark. The UV lamps were turned on for 30 min in order to monitor the reaction initiated by Cl atoms.
At the moment of Cl2 injection in the dark, the alcohol concentration showed a very fast decay in its concentration in the first 2-3 min, which was immediately followed by stabilization. Alcohol decays were of the order of 5-10% for 1-octen-3-ol, 15-20% for (Z)-2-penten-1-ol, and about 30% for the C6 alcohols. In the experimental set-up, the inlet and outlet ports are relatively close to each other (20 cm) and, therefore, it is believed that Cl2 concentrations inside the injection port could be for a while up to 10 4 times higher than their final values in the chamber. These very high local concentrations of Cl2 may accelerate the reaction with the studied alcohol in the dark within the first few minutes after Cl2 injection. Details concerning such dark reactions were already reported in previous studies for similar structurally unsaturated compounds [38,39,[48][49][50]. Nevertheless, these suspected side reactions are not expected to modify the kinetic measurements since (1) alcohol concentrations were stabilized a few minutes after the Cl2 injection, and (2) studied kinetics between unsaturated alcohols and Cl atoms are very fast compared to the kinetics of these side reactions.
The relative-rate kinetic method was used to determine the alcohol + Cl rate constants. The alcohols of interest and reference compounds simultaneously react with Cl atoms with rate constants kAlc+Cl and kRef+Cl, respectively.

Products Cl Alcohol
Cl Alc Assuming that the alcohol and reference compounds only disappear through reaction with Cl atoms, their concentration changes between t = 0 and time t can be related through the following equation: For each reference, using the determined slope (b) and the known kRef+Cl, the kAlc+Cl values could be calculated. The uncertainty of the determined rate constant (ΔkAlc+Cl) was estimated with the following expression: where σ and ΔkRef+Cl are the uncertainties of the slope (b) and kRef+Cl, respectively. The final rate constant kAverage and its uncertainty ΔkAverage (standard error of the mean) were calculated using Equations (10) and (11), respectively.

Kinetic Measurements for (Z)-2-Penten-1-ol
Three experiments were performed for the (Z)-2-penten-1-ol reaction with Cl atoms in zero air. Each experiment was done with one of the three reference compounds (THF, propan-1-ol, and octane), with this strategy leading finally to three distinct determinations. The kinetic plots for (Z)-2-penten-1-ol using THF, propan-1-ol, and octane as reference compounds are displayed in Figure 1a-c, respectively, for all the selected combinations of fragment ions.   The relative rate plots show very good linearity with zero intercepts, indicating absence of secondary reactions. Use of the known reference rate constants kRef+Cl allows kAlc+Cl to be determined for each investigated situation. All the kAlc+Cl/kRef+Cl ratios and the corresponding rate constant values for (Z)-2-penten-1-ol reactions with Cl atoms are reported in Table 1. For the (Z)-2-penten-1-ol with Cl atom reaction, an average rate constant value of (2.99 ± 0.53) × 10 −10 cm 3 •molecule −1 •s −1 was determined.

Kinetic Measurements for 1-Octen-3-ol
Six experiments were carried out for 1-octen-3-ol (two experiments for each reference compound used in the present study). Figure 5 displays the kinetic plots corresponding to the performed experiments.   All the kAlc+Cl/kRef+Cl ratios and the corresponding rate constants for each reference compound are reported in Table 1. The average rate constant for the Cl atom reaction with 1-octen-3-ol at 296 K was (4.03 ± 0.77) × 10 −10 cm 3 •molecule −1 •s −1 .

Summary of All the Determined Rate Constants and Comparison with Literature Data
The obtained rate constants (in 10 −10 cm 3 •molecule −1 •s −1 ) are summarized in Table 1 and compared with the available literature data. All the details for the calculations were explained above in the experimental description. For the (Z)-2-penten-1-ol and Cl atom reaction, Rodríguez et al. [38] reported a rate constant value determined by the relative-rate kinetic method with CCl3COCl as the Cl atom precursor. In this study octane, propene, and cyclohexane were used as reference compounds. The average value between the three determinations of Rodríguez et al. [38] was 3.00 × 10 −10 cm 3 •molecule −1 •s −1 . This value is in excellent agreement with the value determined in the present study (2.99 × 10 −10 cm 3 •molecule −1 •s −1 ).
A series of reactions between Cl atoms and several unsaturated alcohols was investigated by Gibilisco et al. [39] using Cl2 as a source of Cl atoms. Three hexenols among this series were also studied in the present work. The obtained rate constants are in good agreement with Gibilisco et al. [39], with the relative differences between our determinations and Gibilisco et al.'s determinations [39] being of the order of 10% or less for all hexenols. For 1-octen-3-ol, no other measurements are available in the literature.
Although, for 1-octen-3-ol, the rate constant values determined with THF as reference compound appear to be higher than those determined using propan-1-ol and octane, the values are reported as obtained since (i) these systematic higher values were not observed for C5 alkenols, and (ii) if necessary, the rate coefficient values for the alkenol reaction with Cl atom can be easily calculated by averaging the values obtained with the other two reference compounds. Moreover, one should also note that similar differences were observed for the hexenols. This may indicate a systematic error on k(THF + Cl) and, therefore, further studies on k(THF + Cl) are needed.

Structure-Activity Relationships
The reactions between unsaturated VOCs and Cl atoms proceed mainly via the addition to the double bond to produce a chlorine-containing alkyl radical. Addition occurs preferably on the least substituted carbon, thus forming the most stable radical [60]. This chlorinated alkyl radical adds O2 to form a chlorinated peroxy radical whose further reactions lead to oxygenated chlorine-containing compounds [48]. A second minor pathway is the abstraction of the allylic hydrogen from the C-H bond. For example, a very recent work showed that the reaction of (Z)-3-hexene proceeds via both Cl atom addition and hydrogen abstraction with branching ratios of around 70% and 30%, respectively [61]. Table S2 (Supplementary Materials) presents a synthesis of all the known rate constants of unsaturated alcohol + Cl from C3 to C8. The comparison of these data could help to better understand the relationships between molecular structure and rate constants. For a group of alcohols with the alcohol function located on the first carbon atom (alken-1-ol) and a C=C bond placed the second or third carbon atom, data are presented in Figure 6 for comparison purposes. Considering the series of 2-alken-1-ols (red) with 2-propen-1-ol [49], 2-buten-1-ol [50], (Z)-2-penten-1-ol ( [38] and this work), and (E)-2-hexen-1-ol ( [39] and this work), the increase in chain length seems to lead to higher rate constants. A similar trend is also observable for the reaction of Cl atoms with the series of 3-alken-1-ols (blue) with 3-buten-1-ol [62], (E)-3-hexen-1-ol, (Z)-3-hexen-1-ol ( [39] and this work), (Z)-3-hepten-1-ol, and (Z)-3-octen-1-ol [39]. The activation effect of the chain length on the rate constants was also observed in previous works for the reaction of Cl atoms with linear alcohols [63] and alkenes [22,[64][65][66].

Effect of the OH Group
In order to better understand the role of the OH group in the unsaturated alcohol chains, the rate constants for the reaction of five unsaturated alcohols with Cl atoms were compared to the rate constants of their corresponding alkenes reaction with Cl atoms (Figure 7). Table S3 (Supplementary Materials) lists the rate constants for the reactions of Cl atoms with a series of alkenes, i.e., propene, 1-butene, 1-pentene, cis-3-hexene, and 1-octene. Comparing the data presented in Tables S2 and S3 (Supplementary Materials) clearly shows that the rate constants for the reactions of Cl atoms with propene [67], 1-butene [65], 1-pentene [22], and 1-octene [66] are higher than the rate constants for the reactions of Cl atoms with 2-propen-1-ol [49], 3-buten-1-ol [62], 1-penten-3-ol [38], and 1-octen-3-ol (this work), respectively. A similar trend is observed when comparing (Z)-3-hexene + Cl [61] and (Z)-3-hexen-1-ol + Cl ( [39] and this work). However, the data from the present work prove also that the presence of the hydroxyl group in the molecular structure clearly has a deactivating effect on the kinetics of unsaturated alcohols with Cl atoms through a negative inductive influence.
Moreover, Gibilisco et al. [39] suggested a stronger deactivating effect of the OH group for α-alkenols compared to β-alkenols based on literature data. Yet, the present results do not confirm this assumption, probably because the measured rate constants are close to the collision rate. Finally, it can be underlined that an activating effect of the OH group was observed for OH radical addition to alkenols [68], suggesting different reaction mechanisms.
The estimated atmospheric lifetimes of the studied alcohols for their reactions with the atmospheric oxidants range from 1.24 to 4.7 h, except for the reactions with Cl atoms for which much higher lifetimes were found in a continental background atmosphere. Thus, C5-C8 unsaturated alcohols are rapidly removed in the gas phase by O3, NO3, or OH and disappear closely to their emission sources. Yet, it is worth stressing that, in the marine boundary layer and in coastal areas, the Cl-initiated oxidation of alkenols can become competitive compared to other oxidants with lifetimes on the order of a few hours. In addition to the marine boundary layer and coastal areas, specific continental areas were recently shown to display even higher Cl atom concentrations [27,81,82]. The present results will help in better understanding the atmospheric chemistry in these regions.

Conclusions
In the present work, five structurally similar unsaturated alcohols including (Z)-2-penten-1-ol, (E)-2-hexen-1-ol, (E)-3-hexen-1-ol, (Z)-3-hexen-1-ol, and 1-octen-3-ol were investigated with regard to their reactivity toward Cl atoms. Investigations were performed in a 600-L Teflon reaction chamber at 296 ± 2 K and 1 atm zero air. The relative-rate kinetic approach was used to determine rate constant values of the target unsaturated alcohols reaction with Cl atoms produced by the photolysis of molecular chlorine. Tetrahydrofuran, propan-1-ol, and octane were used as reference compounds. For the first time, the decays of alkenols and reference compounds were monitored using selected ion flow tube mass spectrometry (SIFT-MS).
The rate constant values (in 10 −10 cm 3 •molecule −1 •s −1 ) determined in the present work, 2.99 ± 0.53 for (Z)-2-penten-1-ol, 3.05 ± 0.59 for (E)-3-hexen-1-ol, 3.15 ± 0.58 for (Z)-3-hexen-1-ol, 3.41 ± 0.65 for (E)-2-hexen-1-ol, and 4.03 ± 0.77 for 1-octen-3-ol, within the experimental error limits, are in very good agreement with literature data. The present work reports, for the first time, the rate constant value for the reaction between 1-octen-3-ol and Cl atoms. Data obtained in the present work proves the activation effect of the chain length on the rate constants values specific for the kinetics between alkenols and Cl atoms. Moreover, evidence was obtained of the effect of the hydroxyl group presence in the molecular structure, with a deactivating effect on the kinetics of unsaturated alcohols with Cl atoms through negative inductive influence. The atmospheric lifetimes estimated in the present work with regard to atmospheric oxidant abundances suggest that alkenol reactivity toward Cl atoms becomes an important competitive process in alkenol sinks, especially in marine boundary layer and coastal areas.
Supplementary Materials: The following are available online at www.mdpi.com/2073-4433/11/3/256/s1 Figure  S1: General description of Thalamos facility, Table S1: Product ion distributions of the H3O + , NO + , and O2 + reactions with the studied unsaturated alcohols and reference compounds (BR-branching ratios of product ions; data come from the SIFT-MS library; BR could vary depending on the experimental conditions and working status of the instrument), Table S2: Rate constants for the reactions of Cl atoms with unsaturated alcohols at 298 K and atmospheric pressure (all rate constants are expressed in 10 −10 cm 3 •molecule −1 •s −1 ), Table S3: Rate constants for the reactions of Cl atoms with alkenes at 298 K and atmospheric pressure (all rate constants are expressed in 10 −10 cm 3 •molecule −1 •s −1 ).

Conflicts of Interest:
The authors declare no conflicts of interest.