Rate coefficients for the gas-phase reaction of OH with ( Z )-3-hexen-1-ol , 1-penten-3-ol , ( E )-2-penten-1-ol , and ( E )-2-hexen-1-ol between 243 and 404 K

Rate coefficients,k, for the gas-phase reaction of the OH radical with ( Z)-3-hexen-1-ol (( Z)CH3CH2CH = CHCH2CH2OH) (k1), 1-penten-3-ol (CH3CH2CH(OH)CH = CH2) (k2), (E)-2-penten-1-ol ((E)-CH3CH2CH = CHCH2OH) (k3), and (E)-2-hexen-1-ol ((E)-CH3CH2CH2CH = CHCH2OH) (k4), unsaturated alcohols that are emitted into the atmosphere following vegetation wounding, are reported. Rate coefficients were measured under pseudo-first-order conditions in OH over the temperature range 243–404 K at pressures between 20 and 100 Torr (He) using pulsed laser photolysis (PLP) to produce OH radicals and laser induced fluorescence (LIF) to monitor the OH temporal profile. The obtained rate coefficients were independent of pressure with negative temperature dependences that are well described by the Arrhenius expressions k1(T )= (1.3±0.1)×10 −11exp[(580±10)/T ]; k1(297K)= (1.06±0.12)×10 −10 k2(T )= (6.8±0.7)×10 −12exp[(690±20)/T ]; k2(297K)= (7.12±0.73)×10 −11 k3(T )= (6.8±0.8)×10 −12exp[(680±20)/T ]; k3(297K)= (6.76±0.70)×10 −11 k4(T )= (5.4±0.6)×10 −12exp[(690±20)/T ]; k4(297K)= (6.15±0.75)×10 −11 Correspondence to: J. Burkholder (james.b.burkholder@noaa.gov) (in units of cm3 molecule−1 s−1). The quoted uncertainties are at the 2 σ (95% confidence) level and include estimated systematic errors. The rate coefficients obtained in this study are compared with literature values where possible.


Introduction
Biogenic volatile organic compounds (BVOCs) are emitted into the atmosphere in quantities that exceed the emission of VOCs from anthropogenic sources (Guenther et al., 2000(Guenther et al., , 1995)).BVOCs are chemically active compounds and their gas-phase chemistry has a direct impact on air quality on local to regional scales through their impact on the abundance of HO x (HO x = OH + HO 2 ), ozone production, and contributions to secondary organic aerosol (SOA).The formation of organic nitrates locally also leads to the transport of NO x (NO x = NO + NO 2 ) and subsequent ozone production on the regional and continental scale.It is therefore important to know not only the atmospheric abundance of BVOCs, but also their reaction rates and degradation pathways to enable accurate model calculations used for air quality forecasts as well as regulatory purposes.
A number of unsaturated BVOCs (molecules containing carbon-carbon double bonds) are emitted by a variety of plant species in response to wounding due to their antibacterial properties (Nakamura and Hatanaka, 2002).The unsaturated compounds primarily include C5 and C6 aldehydes, ketones, and alcohols and are collectively referred to as "green leaf volatiles".The atmospheric lifetimes of these oxygenated BVOCs and the formation of atmospheric degradation products from these compounds are greatly influenced by the presence of the >C = C< double bond, to which radicals such as OH, O 3 , and NO 3 can add, and the presence of functional groups.(Z)-3-hexen-1-ol ((Z)-CH 3 CH 2 CH = CHCH 2 CH 2 OH), leaf alcohol, is emitted into the atmosphere following the enzymatic oxidation of αlinolenic acid in response to stress, e.g.drought, (Ebel et al., 1995) and vegetation wounding (Hatanaka and Harada, 1973).(Z)-3-hexen-1-ol is emitted globally from a wide range of plants including grass, clover, alfalfa, grape, lettuce, onion, orange, peach, and oak (Arey et al., 1991;Kirstine et al., 1998).Another significant C6 green leaf volatile associated with the wounding of plants is (E)-2hexen-1-ol ((E)-CH 3 CH 2 CH 2 CH = CHCH 2 OH) (Kirstine et al., 1998).Several C5 oxygenates derived from α-linolenic acid are also emitted by vegetation for the purpose of microbiological protection.These compounds include 1penten-3-ol (CH 3 CH 2 CH(OH)CH = CH 2 ) and 2-penten-1-ol (CH 3 CH 2 CH = CHCH 2 OH) which are emitted by a variety of plants (Fisher et al., 2003;Heiden et al., 2003;Karl et al., 2001;Kirstine et al., 1998).
In this study, gas-phase rate coefficients for the reaction of four atmospherically relevant unsaturated alcohols with the OH radical OH + (Z)-3-hexen-1-ol → products k 1 (R1) were measured over a range of temperature (243-404 K) and pressure (20-100 Torr, He).This work is a continuation of a previous study from this laboratory in which temperature dependent rate coefficients for the reaction of OH with several green leaf unsaturated aldehydes were reported (Davis et al., 2007).Previous studies have reported room temperature reaction rate coefficients for Reactions (R1) (Atkinson et al., 1995) and (R2) (Orlando et al., 2001).During the course of this work another kinetic study of Reactions (R1) and (R2) was reported (Jiménez et al., 2009).We have compared our results with these previous studies and the rate coefficient for the analogous well studied OH + (CH 3 ) 2 C(OH)CH = CH 2 (2methyl-3-buten-2-ol, MBO) reaction.We also briefly examine the reactivity trends for the (E)-and (Z)-isomers.

Experimental details
Rate coefficients for Reactions (R1)-(R4) were measured under pseudo-first-order conditions in the OH radical, [Alcohol] [OH], using pulsed laser photolysis (PLP) to produce OH radicals and laser induced fluorescence (LIF) to measure OH radical temporal profiles.A schematic of the experimental apparatus is shown in Fig. 1.The PLP-LIF apparatus has been used extensively in our laboratory and is described in detail elsewhere (Davis et al., 2007;Vaghjiani and Ravishankara, 1989).A particular emphasis in this work was placed on the accurate determination of the reactant concentration in the LIF reactor by using on-line spectroscopic measurements and precise gas flow measurements.A description of the experimental methods used and a brief description of the experimental apparatus are given below.

Experimental apparatus
Rate coefficients were measured in a 150 cm 3 jacketed 15 cm long Pyrex LIF reactor.The temperature of the reactor was maintained by circulating fluid from a heating (or cooling) reservoir through its jacket.The temperature of the gas in the reaction zone of the reactor was measured using a calibrated retractable thermocouple to within ±1 K. OH radicals were produced by pulsed laser photolysis of H 2 O 2 or HNO 3 at 248 nm (KrF excimer laser) or HONO (DONO) at 351 nm (XeF excimer laser).DONO photolysis was used to produce OD radicals.The initial OH radical concentration, (OH) 0 (molecule cm −3 ), was estimated from the precursor concentration, its absorption cross section, σ (λ) (cm 2 molecule −1 ), and quantum yield, (λ), at the photolysis wavelength, λ, and the photolysis laser fluence, F (photons mJ −1 pulse −1 ) The photolysis laser fluence was measured at the exit of the LIF reactor using a calibrated power meter and fluences in the range 1-16 mJ cm −2 pulse −1 (1.25 × 10 15 photons mJ −1 at 248 nm and 1.76 × 10 15 photons mJ −1 at 351 nm) were used.The OH precursor concentration was estimated from the measured OH decay in the absence of the alcohol reactant and the rate coefficient for the reaction of the precursor with OH (Sander et al., 2006).The precursor concentration varied in the range (0.4-3.0) × 10 14 molecule cm

Unsaturated
Alcohol Mixture profile measurements.H 2 O 2 was used as the OH radical precursor in the majority of the experiments, while HNO 3 and HONO were used in several experiments to evaluate possible systematic errors in the rate coefficient measurements.The OH radical was excited in the A 2 + (v = 1) ← X 2 (v = 0) band near 282 nm (OD was excited near 287 nm) using the frequency-doubled output of a pulsed Nd:YAG pumped dye laser.The OH fluorescence signal was detected by a photomultiplier tube (PMT) mounted orthogonal to the photolysis and probe laser beams.A band-pass filter (peak transmission at 310 nm with a 20 nm band-pass, FWHM) mounted in front of the PMT was used to isolate the OH fluorescence signal.The PMT signal was averaged for 100 laser shots using a gated charge integrator.OH temporal profiles were obtained by varying the delay between the photolysis and probe lasers, the reaction time, between 10 µs and 50 ms.
OH temporal profiles were measured under pseudo-firstorder conditions in OH and the OH decay obeyed the relationship

Alcohol concentration determination
The concentration of (Z)-3-hexen-1-ol, 1-penten-3-ol, (E)-2-penten-1-ol, or (E)-2-hexen-1-ol in the LIF reactor was determined by on-line UV and infrared absorption measurements as well as measured flow rates of the sample from dilute gas mixtures.UV (185 nm) and infrared absorption cross sections were measured as part of this study using the Beer-Lambert law where L is the pathlength of the absorption cell and the alcohol concentration was determined by absolute pressure measurements.
UV absorption of the alcohols was determined using a Hg pen-ray lamp light source combined with a solar-blind photodiode detector with a 185 nm band-pass filter.A similar setup was used for the cross section determination and in the kinetic measurements.For the cross section measurements, two absorption cells with quartz windows and optical pathlengths of 1.0 and 1.95 cm were used.The absorption cells used for monitoring the concentration in the LIF reactor during the kinetic measurements had 50 cm pathlengths.I 0 was measured with the absorption cell flushed with bath gas (He) or evacuated.A flow of the pure alcohol vapor was then introduced and the lamp intensity, I, as well as the absolute pressure recorded.The pressure of the alcohol in the absorption cell varied over the range 0.1-0.6Torr (the exact pressure range differed depending on the molecule) to obtain the absorption cross section using Eq.(3).
b Infrared absorption spectra between 800-4000 cm −1 are given in the supporting information.c Integration interval was 2800-3050 cm −1 .
Infrared cross sections were obtained relative to the UV absorption cross section at 185 nm as follows.A slow gas flow of a dilute alcohol/He mixture was passed through a UV absorption cell (50 cm) and then through an infrared absorption cell.Infrared absorption measurements were made using a Fourier transform infrared spectrometer (FTIR) with 100 co-added scans at a resolution of 1 cm −1 .A single pass infrared absorption cell with a pathlength of 10 cm was used for the majority of the measurements.A small volume (∼750 cm 3 ) multi-pass absorption cell with a total optical pathlength of 485 cm was used in a few measurements.Good agreement was obtained for the spectra obtained using the two absorption cells.The sample then passed through a second UV absorption cell after the infrared cell to test for sample loss; no loss was observed.The infrared absorption cross sections were determined using the concentrations determined from the UV absorption measurements and the pressures measured in the absorption cells.The UV and infrared absorption cross sections obtained in this work are given in Table 1.Figures showing the infrared absorption spectra are given in the Supporting Information.
It is important to note that in our methods the UV and infrared absorption cross sections were not measured independently and rely on the absolute pressure measurements used in the UV cross section determination and the preparation of the sample bulb mixtures.In our kinetic measurements, using a combination of UV and infrared measurements provided a self-consistency check of the alcohol concentration determination and also provided a means to evaluate the loss of reactant in the gas flow through the apparatus, which is a particularly important issue at the temperature extremes employed in this study.
For the kinetics measurements, the alcohol samples were introduced into the gas flow from dilute mixtures that were prepared off-line, as shown in Fig. 1.The gas flow of the dilute alcohol/He mixture was measured using a calibrated flow meter.The sample was diluted with a bath gas flow before entering the FTIR.The sample was diluted further prior to entering the first UV absorption cell.A flow of purge gas added in front of the windows of the LIF reactor led to a small, ∼1%, dilution to the alcohol concentration measured after the LIF reactor.The absolute alcohol concentration in the LIF reactor was also measured using calibrated measured flow rates and pressures as well as the on-line spectroscopic measurements described above.The three methods agreed very well, to within <3%, under all experimental conditions.He (UHP, >99.9995%) and O 2 (UHP, >99.99%) were used as supplied.Concentrated H 2 O 2 (>95% mole fraction) was prepared by bubbling N 2 for several days through a sample that was initially ∼60% mole fraction.The H 2 O 2 concentration in solution was determined by titration with a standard KMnO 4 solution.HONO (DONO) was prepared on-line by dropwise addition of a 0.1 M NaNO 2 solution (in The OH precursor was added to the gas flow by bubbling He through the liquid.The OH radical precursor was added to the gas flow just prior to the flow entering the LIF reactor.Gas flows were measured using calibrated flow transducers and pressures were measured using calibrated 1, 100, and 1000 Torr capacitance manometers.

Results and discussion
Tables 2-5 summarize the experimental conditions used in the determination of k 1 -k 4 and the measured rate coefficients.The kinetic measurements for (Z)-3-hexen-1-ol, 1penten-3-ol, (E)-2-hexen-1-ol, and (E)-2-penten-1-ol exhibit very similar behavior and were found to be independent of pressure over the range 20-100 Torr (He). Figure 2 shows a set of OH temporal profiles that illustrate the precision of the measured pseudo-first-order decays and the dependence on the alcohol concentration.The OH decays shown in Fig. 2 are for the OH + 1-penten-3-ol reaction at 297 K but similar quality data was obtained for the other compounds as well as under other experimental conditions.
Figure 3 shows the pseudo-first-order rate coefficient data for the OH + (Z)-3-hexen-1-ol, OH + 1-penten-3-ol, OH + (E)-2-hexen-1-ol, and OH + (E)-2-penten-1-ol reactions obtained at 244, 297, and 374 K.The k determinations were very precise at all temperatures and under all experimental conditions as outlined in Tables 2-5.The obtained rate coefficients were independent of the variations in experimental parameters such as laser fluence, radical source (H 2 O 2 , HNO 3 , HONO), initial OH concentration, precursor concentration, flow velocity, and the presence of O 2 (discussed further below).The final rate coefficient at each temperature was obtained by fitting all available pseudo-firstorder rate data together, Eq. ( 2).
As expected, Reactions (R1)-(R4) are all efficient, although differences in reactivity were observed within the high precision of our measurements.The greatest difference in the measured rate coefficients was between the (Z)and (E)-isomers, which is discussed further below.The final room temperature rate coefficients are k 2 (297K) = (7.12± 0.14) × 10 −11 cm 3 molecule −1 s −1 where the quoted uncertainties are 2σ from the precision of the least-squares analysis.Uncertainties quoted throughout this paper are at the 2σ (95% confidence interval) level unless stated otherwise.
The negative activation energies (E/R) for Reactions (R1)-(R4) are similar with values between −580 and −690 K.The E/R value for the OH + (Z)-3-hexen-1-ol reaction, k 1 , of −580 ± 10 K, is ∼15% lower than for the other C5 and C6 unsaturated alcohols included in this study.The negative activation energies observed for these reactions is consistent with a reaction mechanism dominated by the addition of the OH radical to the >C = C< double bond.[(E)-2-penten-1-ol] Laser Fluence k 3 (T ) c (K) (Torr, He) (10 14 molecule cm −3 ) (10 11 molecule cm −3 ) (10 14 molecule cm −3 ) (mJ cm −2 pulse −1 ) (10 −11 cm 3 molecule −1 s −1 )  [(E)-2-hexen-1-ol] Laser Fluence k 4 (T ) c (K) (Torr, He) (10 14 molecule cm −3 ) (10 11 molecule cm −3 ) (10 14 molecule cm −3 ) (mJ cm −2 pulse −1 ) (10 −11 cm 3 molecule −1 s −1 ) For the compounds studied here, experiments were performed with and without the addition of O 2 to the reaction mixture.Experiments performed with the addition of O 2 were intended to scavenge the OH-alcohol adduct as a more stable peroxy, RO 2 , radical thus minimizing the possible regeneration of OH from the OH-Alcohol adduct formed as a product in Reactions (R1)-(R4).Previous studies of the analogous OH + 2-methyl-3-buten-2-ol (MBO) reaction have shown that OH regeneration from the OH-MBO adduct dissociation with the release of OH from the alcohol group yields an apparent decrease in the measured rate coefficient in the absence of O 2 (Baasandorj and Stevens, 2007;Rudich et al., 1995).The rate coefficients measured in this study with and without added O 2 were found to be statistically identical, Table 6.Summary of rate coefficients for the OH radical reaction with (Z)-3-hexen-1-ol, 1-penten-3-ol, (E)-2-penten-1-ol and (E)-2-hexen-1-ol obtained in this work and taken from the literature.±20% estimated uncertainty in the rate coefficient of the reference compound, where k(296 K) = 6.48 × 10 −11 cm 3 molecule −1 s −1 for the OH + (E)-2-butene reference reaction was used; c The error is 2σ of the least-squares fit combined with the estimated overall uncertainties; d The error is 2σ precision with an estimated 10% uncertainty in the reference rate coefficient where k(298 K) = 2.63 × 10 −11 cm 3 molecule −1 s −1 for the OH + Propene reference reaction was used.PLP ≡ pulsed laser photolysis -laser induced fluorescence.RR ≡ relative rate.
see Tables 2-5.Therefore, it was concluded that OH regeneration was insignificant, estimated to be <5% for Reactions (R1)-(R4), under the conditions of our measurements.
In addition, rate coefficients for the reaction of OD with (Z)-3-hexen-1-ol and (E)-2-hexen-1-ol were also measured.The rate coefficient for the OD + (Z)-3-hexen-1-ol reaction was measured to be (10.3 ± 1.1) × 10 −11 cm 3 molecule −1 s −1 at 50 Torr and 297 K in the absence of O 2 , nearly identical to the OH rate coefficient for Reaction (R1).The direct monitoring of OH during the OD reaction also indicated no detectable formation of the OH radical in this system.These results combined with the O 2 independence of k 1 indicate that OH regeneration was negligible in Reaction (R1).Similar behavior was observed for the OD+ (E)-2-hexen-1-ol reaction where the measured rate coefficient at 297 K was (5.91 ± 0.57) × 10 −11 cm 3 molecule −1 s −1 .

Error analysis
The absolute accuracy of the rate coefficients measured for Reactions (R1)-(R4) was dependent on the precision of the measurements, the uncertainties in the determination of the alcohol concentration in the LIF reactor, and the possible contributions of systematic errors to the measurement.The precision of the kinetic measurements was very good and contributed <3% to the overall uncertainty in the measured rate coefficients.The alcohol concentration in the LIF reactor was determined from the measured gas flows as well as the on-line infrared and UV absorption measurements.
The agreement between the infrared and UV absorption concentration determinations was better than 5% over the entire range of experimental conditions employed in this study.
The UV absorption measurements performed before and after the LIF reactor also agreed to better than 5% under all experimental conditions.This is particularly noteworthy for the measurements at the temperature extremes and indicates that the reactant alcohol was not lost in the gas flow through the apparatus.The UV and infrared absorption cross sections used in the alcohol concentration determination were measured as part of this work and are estimated to have an uncertainty of <5%.Systematic uncertainties were primarily evaluated experimentally through the use of a range of experimental conditions during the kinetic measurements including variations in total pressure, buffer gas, photolysis laser fluence, gas flow velocity, initial OH radical concentration, OH precursor, and precursor concentration over the course of the rate coefficient determinations.Measurements were also performed with O 2 added to the reaction mixture to evaluate the influence of possible secondary chemistry of the radicals formed in Reactions (R1)-(R4) on the measured rate coefficients.The measured rate coefficients were found to be independent of the variations in experimental conditions, within the precision of the measurements, as summarized in Tables 2-5.
The overall 2σ uncertainty in the measured rate coefficients for (Z)-3-hexen-1-ol, 1-penten-3-ol, (E)-2-penten-1ol, and (E)-2-hexen-1-ol are estimated to be 10, 7, 7 and 8%, respectively.The estimated systematic uncertainties are included in the reported Arrhenius A factors while the uncertainty in the E/R values was taken from the precision of the least-squares fits.The recommended rate coefficients, including total absolute uncertainties, are given in Table 6.
They reported k 1 (298 K) = (10.8± 2.2) × 10 −11 cm 3 molecule −1 s −1 , which is in excellent agreement with the value obtained in the present study.Jiménez et al. (2009) reported k 1 (298 K) = (9.57± 2.42) × 10 −11 cm 3 molecule −1 s −1 obtained using a PLP-LIF technique, which was very similar to the method used in the present study.Although the reported rate coefficient is slightly less than obtained in our work the agreement is within the combined uncertainties of the measurements.Orlando et al. (2001) used a relative rate method to determine k 2 (298 K) = (6.7 ± 0.9) × 10 −11 cm 3 molecule −1 s −1 in   6.
good agreement with the value determined in our work.As shown in Fig. 5, the rate coefficient data reported by Jiménez et al. (2009) for Reaction (R2) over the temperature range 263-353 K is systematically less, by ∼25%, than measured in our work.The temperature dependence of k 2 obtained in the two studies is in reasonable agreement, E/R = -690 ± 20 K in our work compared with E/R= −606 ± 30 K reported by Jiménez et al.The difference between the rate coefficients is reduced to ∼15% when the rate coefficient data are normalized to the same 185 nm absorption cross section for 1penten-3-ol, see Table 1; absorption measurements at 185 nm were used in both studies to measure the 1-penten-3-ol concentration.
The room temperature rate coefficients for the reaction of OH with (E)-2-penten-1-ol and (E)-2-hexen-1ol can be compared with the values predicted using the structure activity relationships (SAR) of Kwok and Atkinson (1995) combined with updated enhancement factors for the −CH 2 OH and −CH 2 CH 2 OH groups from the work of Papagni et al. (2001).The SAR estimated rate coefficients for Reactions (R3) and (R4) are identical since the reactivity factors for C 2 H 5 and C 3 H 7 are the same.The SAR estimated rate coefficient is 1.33 × 10 −10 cm 3 molecule −1 s −1 , which is approximately twice the experimentally measured values of (6.76 ± 0.70) × 10 −11 cm 3 molecule −1 s −1 and (6.15 ± 0.75) × 10 −11 cm 3 molecule −1 s −1 .Note that the rate coefficient measured in our work for the shorter chain length molecule is actually greater.The level of agreement between the SAR estimate and the experimental values is probably within the acceptable range for the SAR estimation method.The SAR rate coefficient for the OH + (Z)-3-hexen-1-ol reaction is in good agreement with the experimental value, to better than 10%, while the SAR rate coefficient for the 1-penten-3-ol reaction is only ∼15% less than the experimental value.The difference between the estimated and measured rate coefficients for the (E)-2-penten-1-ol and (E)-2-hexen-1-ol reactions may in part be due to the steric hindrance in the OH addition to the trans isomers, which may not be adequately accounted for in the SAR reactivity coefficients.We mention this because the measured rate coefficient for the OH + (Z)-2-penten-1-ol (Orlando et al., 2001) reaction, which should have less steric hindrance, is (1.06 ± 0.15) × 10 −10 cm 3 molecule −1 s −1 in agreement with the SAR value of 9.0 × 10 −11 cm 3 molecule −1 s −1 .The steric effect seems to offset the enhancement associated with the presence of the −CH 2 OH group adjacent to the double bond.
It is also worthwhile to compare the rate coefficients for the C5 compounds included in this study with that for the OH reaction with (CH 3 ) 2 C(OH)CH = CH 2 (2-methyl-3-buten-2ol, MBO) which has a similar molecular structure.MBO is a C5 unsaturated alcohol commonly found in remote forests and has been studied in the laboratory by several groups (Baasandorj and Stevens, 2007;Imamura et al., 2004;Rudich et al., 1995).Room temperature rate coefficients for the OH + MBO reaction have been reported in the range (5.8-6.6)× 10 −11 cm 3 molecule −1 s −1 in the presence of O 2 .The absolute rate coefficient measurements for the OH + MBO reaction are complicated somewhat by the regeneration of OH radicals from the elimination of the alcohol OH group from the OH-MBO adduct, which was found (as described earlier) not to be a problem for the C5 compounds in the present study.Overall the rate coefficients for the C5 compounds studied in this work are very similar to that for the MBO reaction.The temperature dependence of the rate coefficients studied here are also similar to that of the OH + MBO reaction; Rudich et al. (1995) reported E/R = (−610 ± 50) K.
In general, the rate coefficients for the OH + unsaturated alcohol reactions are only weakly dependent on the chain length of the alcohol.The room temperature rate coefficients for 1-penten-3-ol, (E)-2-penten-1-ol, and (E)-2-hexen-1-ol range between 6.15 × 10 −11 cm 3 molecule −1 s −1 and 7.12 × 10 −11 cm 3 molecule −1 s −1 .Papagni et al. (2001) measured reaction rate coefficients for a series of C3 to C5 unsaturated alcohols.Their measured rate coefficient for MBO was similar to the rate coefficients for allyl alcohol (CH 2 = CHCH 2 OH) (5.46 ± 0.35) × 10 −11 cm 3 molecule −1 s −1 , 3-buten-1ol (CH 2 = CHCH 2 CH 2 OH) (5.50 ± 0.20) × 10 −11 cm 3 molecule −1 s −1 , and 3-buten-2-ol (CH 2 = CHCH(OH)CH 2 ) (5.93 ± 0.23) × 10 −11 cm 3 molecule −1 s −1 .This data set suggests that the rate coefficient is also only weakly dependent on the position of the carbon-carbon double bond with respect to the position of the alcohol group.The (E)-/(Z)-geometry of the compound seems to play a more important role in determining the compounds reactivity.For the compounds with (E)-or trans functional groups around the carbon-carbon double bond a rate coefficient in the range (6-7) × 10 −11 cm 3 molecule −1 s −1 with E/R≈ −680 K was observed.When the functional groups are (Z)-or cis there appears to be less steric hindrance leading to higher reactivity and a less negative activation energy, E/R ≈ −580 K.The rate coefficients for the (Z)-or cis isomer were found to be a factor of ∼1.6 greater than for the (E)-or trans isomer.

Conclusions
Gas-phase rate coefficients for the reaction of several atmospherically relevant unsaturated alcohols with the OH radical were measured as a function of temperature (243-404 K).The rate coefficients show a negative temperature dependence that is consistent with a reaction mechanism involving the addition of OH to the carbon-carbon double bond.No pressure dependence was observed over the pressure range 20-100 Torr (He) indicating that these reactions are in the high-pressure limit for the temperatures and pressures included in this study.Unsaturated compounds are expected a priori to have short atmospheric lifetimes due to their high gas-phase reactivity with the OH radical.The BVOCs included in this study have estimated lifetimes due to loss by reaction with the OH radical of several hours (about 2.5 to 5 h for an OH concentration of 1 × 10 6 molecule cm −3 ).In this study, differences in reactivity between (E)-and (Z)geometrical isomers were observed with the (Z)-isomers exhibiting greater reactivity, which is not accounted for quantitatively in the Kwok and Atkinson (1995) structure activity relationship.The rapid oxidation and degradation product formation of the unsaturated compounds included in this study need to be accounted for in air quality models on local and regional scales.The rate coefficients measured in this work are appropriate for use in atmospheric models.

Fig. 4 .
Fig. 4. Temperature dependent rate coefficient data for Reactions (R1)-(R4), Arrhenius plot, from this work (•).The error bars are the 2σ uncertainty including estimated systematic errors.The lines are least-squares fits of the data to the Arrhenius expression, ln(k) = ln(A)-E/RT where the results are given in Table 6.For comparison, results from previous studies, as described in the text, are included: OH + (Z)-3-hexen-1-ol at room temperature by Atkinson et al. (1995) (♦) and Jiménez et al. (2009) ( ); OH + 1-penten-3ol at room temperature by Orlando et al. (2001) ( ) and between 263 and 353 K by Jiménez et al. (2009) ( ).The results from the previous studies are also included in Table6.
Torr of O 2 added; b OH generated by photolysis of HNO 3 at 248 nm.Value given is [HNO 3 ]; c Quoted uncertainties are 2σ from the precision of the least-squares analysis of k versus [1-penten-3-ol], Eq. (2).
Uncertainties quoted for this work are 2σ including estimated systematic errors; b The quoted error is 2σ of the measured rate coefficient ratio (1.67 ± 0.06) combined with a a