Environ. Sci. Technol. 2007, 41, 3959-3965
Rate Coefficients for the OH + Pinonaldehyde (C10H16O2) Reaction between 297 and 374 K M A X I N E E . D A V I S , †,‡ R A N A J I T K . T A L U K D A R , †,‡ GREGORY NOTTE,§ G. BARNEY ELLISON,§ AND J A M E S . B . B U R K H O L D E R * ,† Earth System Research Laboratory, National Oceanic and Atmospheric Administration, 325 Broadway, Boulder, Colorado 80305-3328, Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, Colorado 80309, Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309-0215
of R-pinene leads to ring-opening and a variety of stable end-products that contain carbonyl, alcohol, and carboxylic functional groups (2). Pinonaldehyde (C10H16O2, 3-acetyl2,2-dimethyl-cyclobutyl-ethanal) is an end-product formed in high yield in the OH, O3 and NO3 initiated oxidation of R-pinene. The molecular structure of pinonaldehyde is shown in the Supporting Information, Figure 1S. A pinonaldehyde yield of 60% for the OH initiated R-pinene oxidation has recently been suggested for use in atmospheric model calculations (3). Pinonaldehyde is a possible marker for R-pinene (monoterpene) chemistry in the atmosphere. At present, measurements of pinonaldehyde in the lower troposphere are limited to just a few field campaigns in forested areas (4-7). In addition to direct observations of its atmospheric abundance, a marker requires a detailed understanding of its atmospheric formation and loss processes. Our work focuses on the determination of the rate coefficient, k1, for the reaction
OH + Pinonaldehyde f Products
The rate coefficient for the reaction of OH with pinonaldehyde (C10H16O2, 3-acetyl-2,2-dimethyl-cyclobutyl-ethanal), a product of the atmospheric oxidation of R-pinene, was measured under pseudo-first-order conditions in OH at temperatures between 297 and 374 K at 55 and 96 Torr (He). Laser induced fluorescence (LIF) was used to monitor OH in the presence of pinonaldehyde following its production by 248 nm pulsed laser photolysis of H2O2. The reaction exhibits a negative temperature dependence with an Arrhenius expression of k1(T) ) (4.5 ( 1.3) × 10-12 exp((600 ( 100)/ T) cm3 molecule-1 s-1; k1(297 K) ) (3.46 ( 0.4) × 10-11 cm3 molecule-1 s-1. There was no observed dependence of the rate coefficient on pressure. Our results are compared with previous relative rate determinations of k1 near 297 K and the discrepancies are discussed. The state of knowledge for the atmospheric processing of pinonaldehyde is reviewed, and its role as a marker for R-pinene (monoterpene) chemistry in the atmosphere is discussed.
Introduction Monoterpenes (C10H16) are a class of unsaturated cyclic Biogenic Volatile Organic Compounds (BVOCs) which are emitted from coniferous trees and account for approximately 10% of the total biogenic hydrocarbons emitted globally (1). The oxidation of monoterpenes in the lower troposphere plays an important role in determining the local HOx budget, regional ozone production, and secondary organic aerosol (SOA) formation. The atmospheric oxidation of R-pinene, the most abundant monoterpene emitted in North America, is initiated by its gas-phase reaction with OH, O3, or NO3. The lifetime of R-pinene is on the order of hours for OH and O3 reactive loss and ∼0.15 h for NO3 reactive loss for representative daytime (OH and O3) and nighttime (NO3) concentrations in the lower troposphere (2). The oxidation * Corresponding author phone: 303-497-3252; fax: 303-497-5822; e-mail:
[email protected]. † Earth System Research Laboratory. ‡ Cooperative Institute for Research in Environmental Sciences, University of Colorado. § Department of Chemistry and Biochemistry, University of Colorado. 10.1021/es070048d CCC: $37.00 Published on Web 04/27/2007
2007 American Chemical Society
(1)
which is a significant daytime removal process for pinonaldehyde. Several studies of the rate coefficient for reaction 1 around room temperature have been reported in the literature. However, discrepancies exist among the reported rate coefficients with values ranging from 4.0 × 10-11 cm3 molecule-1 s-1 to 9.1 × 10-11 cm3 molecule-1 s-1 (8-11). Each of these studies used the relative rate technique, and there is not a clear explanation for the large range in the measured values. Further rate coefficient measurements are needed to resolve the existing discrepancies and also to determine the temperature dependence of k1. In this paper, measurements of the OH + pinonaldehyde reaction rate coefficient including its temperature dependence are reported. Our results are compared with previous determinations and the discrepancies are discussed. An overview of the current understanding of the atmospheric processing of pinonaldehyde is presented and existing gaps in knowledge are discussed. Experimental Details. Pulsed laser photolysis production of OH coupled with its detection by laser induced fluorescence (PLP-LIF) was used to determine the gas-phase rate coefficient for reaction 1 under pseudo-first-order conditions in OH. Rate coefficient determinations were made at temperatures between 297 and 378 K at 55 and 96 Torr (He). The experimental setup is similar to that used in previous kinetic and photolysis quantum yield studies from this laboratory (12, 13). A schematic of the apparatus is shown in Davis et al. (12) and described briefly below. The LIF reactor consisted of a jacketed Pyrex cell approximately 15 cm in length with an internal volume of ∼150 cm3. The reactor was maintained at a constant temperature by circulating a temperature-regulated fluid through its jacket. The temperature of the gas in the reaction zone was directly measured and was accurate to (1 K. OH radicals were produced by pulsed laser photolysis of H2O2 at 248 nm (KrF excimer laser). The initial OH radical concentration, [OH]0, was calculated using
[OH]0 ) σλ Φλ [H2O2] F
(2)
where σλ and Φλ are the absorption cross section (2.0 × 10-20 cm2 molecule-1) and OH quantum yield, respectively, for H2O2 at 248 nm (Φλ ) 2) (14), and F is the photolysis laser fluence (photon cm-2 pulse-1), measured with a calibrated VOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3959
TABLE 1. Summary of Experimental Conditions and Rate Coefficients, k1(T), Determined in This Work for the OH + Pinonaldehyde Reaction temperature pressure V k1(T) laser fluence [H2O2] [OH]0 [pinonaldehyde] (K) (Torr He) (cm s-1) (mJ cm-2 pulse-1) (1014 molecule cm-3) (1011 molecule cm-3) (1014 molecule cm-3)a (10-11 cm3 molecule-1 s-1) 374 349 325 297 297 297
57 58 58 96 59c 55-57
49 45 42 31 38 20-38
5.9 5.8 6.4 7.0 7.0 4.3-15.6
0.8 0.6 0.8 0.5 1.0 0.5-1.4
0.76 0.6 0.9 0.6 1.1 0.6-3.0
0.25-0.94 (5) 0.19-0.98 (5) 0.10-1.15 (6) 0.81-1.24 (3) 0.54-1.2 (4) 0.23-1.28 (15)
2.35 ( 0.04b 2.56 ( 0.04 2.69 ( 0.06 3.31 ( 0.06 3.30 ( 0.08 3.48 ( 0.02 k1(297) ) 3.46 ( 0.04d
a The value in parenthesis is the number of pinonaldehyde concentrations used in the determination of k at that temperature. b Quoted uncertainties 1 are 2σ from the fits of k′ versus [pinonaldehyde] and do not include estimated systematic errors. c 2 Torr O2 added. d Room-temperature rate coefficient obtained by fitting all the T ) 297 K data simultaneously. This value was used as the room-temperature value in the determination of the Arrhenius parameters.
power meter. The H2O2 concentration was estimated from the measured first-order decay of OH in the absence of pinonaldehyde using the OH + H2O2 rate coefficient (14). [OH]0 was varied during the course of the study but was 15%, of the total He gas flow over the semiliquid pinonaldehyde sample held at 297 K. The flow containing the pinonaldehyde sample was diluted at the exit of the sample reservoir. The largest [Pin]LIF was ∼1.5 × 1014 molecule cm-3 which is roughly a factor of 10 lower than its room-temperature vapor pressure. Infrared absorption spectra were recorded using Fourier transform infrared spectroscopy (FTIR) and quantified using literature absorption cross section data, σ(1725.4 cm-1) ) 8.08 × 10-19 cm2 molecule-1 (base e) (10). Infrared absorption spectra were recorded in 50 coadded scans between 500 and 4000 cm-1 at 1 cm-1 resolution. The infrared absorption cell was a small volume, ∼750 cm3, multipass cell mounted in the FTIR sample compartment. The path length of the cell was calibrated in separate measurements to be 485 ( 8 cm and the detection sensitivity for pinonaldehyde was ∼2 × 1012 molecule cm-3. It should be noted that if the pinonaldehyde infrared absorption cross section is revised, the rate coefficient data presented in this study can be scaled accordingly. UV absorption measurements were made continuously during the kinetic measurements. The UV measurements were made at 185 nm using Hg pen-ray lamps and solar blind photodiodes with 185 nm narrow bandpass filters. The Pyrex absorption cells used for the kinetic measurements had 50 cm pathlengths and quartz windows. The pinonaldehyde absorption cross section at 185 nm was determined relative to the infrared cross section reported by Hallquist et al. (10) as part of this study. In our cross section determinations, the infrared measurements were performed using the FTIR spectrometer setup described above. The 185 nm cross section measurements were made using an 84 cm path length absorption cell. UV absorption measurements were also made using a diode array spectrometer (210-400 nm) with a 30W D2 light source at a resolution of ∼1 nm. A He flow passed through the pinonaldehyde sample reservoir, the FTIR, and then into the UV absorption cell. The total pressure in the absorption cells was in the range 150-230 Torr and the residence times of the gas mixture in the absorption cells were ∼5 s (FTIR) and ∼15 s (UV). The pinonaldehyde concentration was varied between 0.7 × 1014 molecule cm-3 and 3.0 × 1014 molecule cm-3 in the diode array/FTIR measurements and between 0.3 × 1013 molecule cm-3 and 1 × 1013 molecule cm-3 for the 185 nm/FTIR measurements. The absorption signals varied linearly with pinonaldehyde concentration over this range obeying Beer-
Lambert’s law. The shape of the pinonaldehyde absorption spectrum recorded between 210 and 400 nm was in good agreement with that reported by Hallquist et al. (10). We obtained an absorption cross section at 285 nm, the peak of the spectrum, of (1.35 ( 0.08) × 10-19 cm2 molecule-1 in excellent agreement, within 2%, with the value reported by Hallquist et al. (10). Measurements made with the gas flow reversed (UV measurements first) yielded identical infrared and UV absorption cross sections, i.e., loss of pinonaldehyde was 99.9995%) was used as the buffer gas. O2 (99.99%) and N2 (99.99%) were used as supplied. Concentrated hydrogen peroxide (>95%) was prepared by bubbling N2 through a H2O2 sample initially at 60% concentration for several days prior to use. H2O2 was introduced into the main gas flow just prior to entering the LIF reactor by passing a small flow of He through a bubbler containing the concentrated sample. Gas flow rates were measured using calibrated electronic mass flow transducers. Pressures were measured using 100 and 1000 Torr capacitance manometers.
Results Table 1 summarizes the experimental conditions used in the kinetic measurements and the rate coefficients obtained for reaction 1. The OH temporal profiles measured with different pinonaldehyde concentrations were of high quality and obeyed eq 3 under all variations in the experimental conditions. Figure 2S in the Supporting Information shows a representative set of OH temporal profiles measured at 297 K. The first-order rate coefficient, k′, data obtained at 297 K, including all variations in experimental conditions, are shown in Figure 1. The error bars on the data points are the 2σ values from the precision of the weighted linear leastsquares fits to the OH temporal profiles using eq 3. The precision of the kinetic data shown in Figure 1 is representative of that obtained at 325, 349, and 374 K, although fewer variations in the experimental conditions were made in those measurements. Analyzing all room-temperature data together yields k1(297 K) ) (3.46 ( 0.04) × 10-11 cm3 molecule-1 s-1 where the quoted uncertainty is at the 2σ level (fit precision). This value is in good agreement with the values
FIGURE 1. Summary of all the OH + pinonaldehyde rate coefficient data measured at 297 K. The symbols highlight sets of data measured under different experimental conditions. The error bars shown are the 2σ (95% confidence limit) uncertainties obtained from the weighted linear least-squares fit of the OH decay profile to eq 3.
FIGURE 2. Arrhenius plot for the OH + pinonaldehyde reaction. Results from this work are shown as the solid circles. The solid line is a weighted linear least-squares fit to our data yielding k1(T) ) (4.5 ( 1.3) × 10-12 exp((600 ( 100)/T) cm3 molecule-1 s-1 (see text for error analysis). The dashed lines represent the uncertainty limits in the rate coefficient in the format used in data evaluations (14) where the error, f(T), is given by f(T) ) f(298 K) exp|g(1/T 1/298)| and f(298 K) ) 1.15 and g ) 50 (1σ level). The values reported in the previous relative rate studies are included for comparison; Glasius et al. (9) (]), Hallquist et al. (10) (4), Alvarado et al. (8) (0), Nozie` re et al. (16) (O).
obtained from the individual determinations at room temperature, as summarized in Table 1, and was used for the determination of the temperature dependence of k1. An Arrhenius plot of the rate coefficient data is shown in Figure 2. The results of a weighted least-squares fit of the data given in Table 1 (weighted by the number of measurements made at each temperature) to a linearized Arrhenius expression are given in Table 2. Error Analysis. A source of uncertainty in the determination of k1(T) using the PLP-LIF method is the systematic VOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3961
TABLE 2. Summary of OH + Pinonaldehyde Rate Coefficient Data k1(297 K) ( 2σa 3.46 ( 0.4 4.0 (1.0d 4.8 ( 0.8e 9.1 ( 1.8d 8.72 ( 1.14d 2.4 3.5
Ab (4.5 (
1.3)c
×
10-12
E/R (K)
T (K)
techniquej
600 ( 100
297-374 293 296 300 298 298 298
PLP-LIF RRf RRg RRh RRi SAR estimation DFT calculation
pressure (Torr) 55-96 760 740 740 760
reference this work Nozie` re et al., 1999 (16) Alvarado et al., 1998 (8) Glasius et al., 1997 (9) Hallquist et al., 1997 (10) Kwok and Atkinson, 1995 (17) Vereecken and Peeters, 2003 (18)
a In units of 10-11 cm3 molecule-1 s-1, literature values are for temperatures near 297 K as listed in the table. b In units of cm3 molecule-1 s-1. The quoted errors from this work are at the 2σ level (95% confidence) and include estimated systematic errors. d The errors quoted are twice the standard deviation of the weighted fit. e The error given is twice the standard deviation of the weighted fit combined with the estimated overall uncertainty in the rate coefficient of the reference compound. fCyclohexane, 1,2-butadiene and isoprene reference compounds. g Propene, 1-butene, and m-xylene reference compounds. h Isoprene and 1,3-butadiene reference compounds. i Propene reference compound. j PLP-LIF: pulsed laser photolysis-laser induced fluorescence, RR: relative rate, SAR: structure activity relationship, DFT: density functional theory. c
errors associated with the determination of [Pin]LIF. The determination of the pseudo-first-order OH decays was very precise and makes a small contribution to the overall uncertainty, typically
