Minimizing Errors in Measured Yields of Particle-Phase Products Formed in Environmental Chamber Reactions: Revisiting the Yields of β-Hydroxynitrates Formed from 1-Alkene + OH/NO_x Reactions

Abstract

Environmental chamber studies are widely used for measuring the yields of reaction products and secondary organic aerosol formed from the oxidation of volatile organic compounds, information that can be subsequently used to develop chemical reaction mechanisms and models of atmospheric chemistry. Such measurements are challenging, however, and many potential sources of errors are not always appreciated. Here, we describe methods for minimizing and accounting for uncertainties associated with the chamber volume, gas and particle sampling, instrument calibrations, gas–wall partitioning, particle deposition to walls, and secondary reactions with OH radicals and demonstrate these methods by measuring yields and branching ratios for the formation of β-hydroxynitrates from reactions of C14 and C15 1-alkenes with OH radicals in the presence of NO x . Experiments were conducted in a Teflon environmental chamber and included Tenax sampling of alkenes with analysis by gas chromatography with flame ionization detection, filter sampling of particle-phase β-hydroxynitrates and dioctyl phthalate seed particles with analysis by liquid chromatography with UV/vis detection, and measurements of aerosol volume concentration and composition using a scanning mobility particle sizer and thermal desorption particle beam mass spectrometer. The measured yields of β-hydroxynitrates are ∼25% higher than those determined previously using similar methods whose uncertainties were less well minimized and characterized.

Comprehensive Analysis of Products and the Development of a Quantitative Mechanism for the OH Radical-Initiated Oxidation of 1-Alkenes in the Presence of NOx

“…This value is within ∼25% of the value of 0.046 calculated from the fraction of reaction that occurred by H atom abstraction (0.30) and the 1,4-hydroxynitrate yield of 0.13 measured by Yeh and Ziemann 35 for reactions of C 14 −C 16 n-alkanes, corrected again by a factor of 1.19 for losses due to the filter sampling method. 20 The yields measured for each reaction product are reported in Table 1. There are no significant differences between the yields measured for the C 14 and C 15 1-alkenes.…”

“…We note that the yields of alkyl nitrates, 1,4-hydroxynitrates, and 1,4hydroxycarbonyls expected to be formed by H atom abstraction (Scheme S1) are 0.09, 0.05, and 0.16, respectively, based on the corresponding yields of 0.29, 0.16, and 0.55 for the reactions of n-alkanes. 19,35 In these calculations, the measured 1,4-hydroxynitrate yield of 0.13 35 was multiplied by a factor or 1.19 to correct for losses due to the filter sampling method as described above, 20 which also slightly reduced the 1,4-hydroxycarbonyl yield of 0.58 that was calculated by difference. 35 Calculation of Branching Ratios for OH Radical Addition Pathways.…”

“…This result is consistent with a previous experiment in which 1-tetradecene and 1-pentadecene reached gas-wall partitioning equilibrium in ∼15 min, with 12 and 17% partitioned to the chamber walls, respectively. 20 Therefore, the assumption that the 1-alkenes were in equilibrium with the walls during sampling seems appropriate, and the partitioning coefficient that would be used to correct the FID signals for gas-wall partitioning before and after the reaction cancels out in eq 1. 19 Gas-phase aldehydes (formaldehyde, tridecanal, and tetradecanal) were collected using cartridges and denuders.…”

“…Particle-phase product yields were corrected for particle deposition to the walls using the nonvolatile DOP seed particles as an internal standard, 20 where the correction factor was determined from the chamber concentrations of DOP measured on filters collected before and after reaction. The average particle wall loss for the 30-min post-reaction sampling period was ∼5% (corresponding to an average correction factor of 1.05) for the eight experiments, similar to values of ca.…”

Gas- and Particle-Phase Products and Their Mechanisms of Formation from the Reaction of Δ-3-Carene with NO3 Radicals

“…The column temperature was then increased at 10 °C min –1 to 280 °C, and eluted compounds were quantified with the FID. The mass concentration of Δ-3-carene reacted was calculated from the equation Δ­[C] = [C] T,i (1 – FID f /FID i ), where [C] T,i is the mass of Δ-3-carene added/chamber volume (assumed to be 7.4 m 3 based on a previously measured average value) and FID i and FID f are measured FID signals/volume of air sampled before and after reaction.…”