Ein massenspektrometrischer Ansatz zur Auswertung von Atmosphären-Simulationskammer Experimenten

Monoterpenes are the most important precursors for atmospheric secondary organic aerosols (SOA). The oxidation of monoterpenes by ozone, OH- and NO3-radicals yields a variety of products, which are partitioned between the aerosol and the gas phase. The SOA yields from monoterpene oxidations and the partitioning of the oxidation products is not well characterized, especially the influence of relative humidity and temperature is rather unclear. The characterization of the partitioning of semivolatile oxidation products is important to extrapolate simulation experiments at higher concentrations to atmospheric conditions. The understanding of the secondary organic aerosol component is an important prerequisite to estimate climatic effects of atmospheric aerosols. In the present work simulation experiments on the ozonolysis of monoterpenes were carried out at the large aerosol-chamber of the research center Jülich and at the temperature controlled AIDA-chamber of the research center Karlsruhe. The investigated monoterpenes were alpha-pinene, limonene, sabinene and myrcene. Mass spectrometric methods were used to determine the partitioning of organic compounds between the aerosol and the gas phase. The monoterpenes and their gas phase oxidation products were determined by proton transfer reaction mass spectrometry (PTR-MS). The aerosol phase was characterized by an aerosol mass spectrometer (AMS) to determine the total aerosol mass and a scanning mobility particle sizer (SMPS) to determine the size distribution and the total volume of the aerosol. Mass spectrometry allows for the direct measurement of a wealth of compounds with high time resolution and low detection limits. In this work a data evaluation routine based on fragmentation patterns of organic compounds has been developed to evaluate PTR mass spectra. It allows for the extraction of mass spectra of individual compounds from the mass spectra of the mixture. It provides the basis for an interference free quantification of fragmented products and precursors in the investigated reaction systems. Additionally, statistical methods were used to evaluate the PTR mass spectra based on time resolved, simultaneous measurements of the precursors and the oxidation products in the gas-phase. A correlation analysis for the identification of precursor and product signals has been developed and validated. Regression analysis was used to determine molar reaction yields of the identified gas phase products. Experiments carried out at the aerosol chamber in Jülich focused on the molar reaction yields of gas phase oxidation products and on mass and carbon balances of the multiphase systems. The molar gas phase reaction yields of the monoterpene oxidation products pinonaldehyde (0.21 +/- 0.04), sabinaketone (0.54 +/- 0.13) and endolim (0.04 +/- 0.01) were determined as the major first generation oxidation products of the ozonolysis of alpha-pinene, sabinene and limonene, respectively. Acetone has been detected as a product of the ozonolysis of all investigated monoterpenes with molar yields of 0.05 - 0.40. Based on the determined yield the global emission rate of acetone from the ozonolysis of alpha-pinene has been estimated to be 4 - 10 % of the total global annual source of acetone. In a temperature range of 292 - 298 K SOA-yields were determined for the ozonolysis in the absence of an OH-scavenger for sabinene (0.10 - 0.15), limonene (0.65), alpha-pinene (0.18 - 0.25) and myrcene (0.16). The monoterpenes with exocyclic double bonds (myrcene and sabinene) have lower SOA-yields than such with endocyclic double bonds alpha-pinene and limonene). The corresponding SOA-yield of a monoterpene mixture (sabinene, myrcene and limonene) is the weighted average of the SOA-yields of the single compounds. The mass spectrometric approach allows for the determination of a time resolved gas-particle mass balance. Product partitioning and the influence of secondary processes were deduced from the mass balance. Secondary oxidation did neither change the aerosol nor the gas phase fractional yields for sabinene and limonene. Aerosol yields for the alpha-pinene oxidation showed a time dependent increase in the SOA-yield caused by secondary oxidation products. Myrcene showed a time dependent increase in the gas phase yield after complete consumption of the monoterpene, i.e. the secondary oxidation produced volatile compounds. Fractional carbon yields of gas-phase and particulate-phase products were determined to describe the partitioning of the products independent of the degree of oxidation. The sum of these fractional yields provides the carbon balance of the closed system "reaction chamber". The employed analysis methods recovered within experimental errors all the reactive carbon in the reaction systems with the one exception of the alpha-pinene ozonolysis. Experiments in the AIDA chamber in Karlsruhe aimed at the influence of relative humidity and temperature on SOA-yields and the partitioning of monoterpene oxidation products. In contrast to the Jülich experiments an OH-scavenger was used during the ozonolysis. In these experiments the SOA-yields of limonene and alpha-pinene were found to be dependent on relative humidity. alpha-Pinene and limonene produced SOA-yields reduced by 50 % under dry conditions and temperatures of 253 K. At 303 K the influence of relative humidity on the SOA-yield was negligible. The impact of the relative humidity on the monoterpene degradation mechanism was derived from PTR-MS-measurements. The reaction of stabilized criegee-biradicals with water vapor is an important reaction path to yield non volatile compounds which form SOA. In these experiments particulate water did not contribute significantly to the aerosol mass under humid conditions. SOA-yields of the ozonolysis of alpha-pinene and limonene were found to be temperature dependant. At 303 K the SOA-yields of alpha-pinene and limonene were 0.18 and 0.43, respectively. Reducing the temperature from 303 K to 253 K under humid conditions increased the SOA-yield for alpha-pinene by a factor of 4.7. At the same conditions the SOA-yield of limonene increased by a factor of 3.3. Using a two-product model to calculate temperature dependent SOA-yields provided good agreement with the observations within a temperature range of 243 - 273 K. Temperature effects on partitioning dominated over gas phase mechanistic effects. The determined model parameters are not applicable for temperatures > 273 K. At these temperatures the temperature dependence of the reaction kinetics gained influence on the SOA-yields. The contributions of semivolatile products to the SOA-yield at decreasing temperature were determined from the AIDA-experiments. At temperatures A temperature dependent rate coefficient for the ozonolysis of alpha-pinene has been determined for a temperature range of 243 - 303 K. Differential equations provided by the master chemical mechanism were fitted to the experimental data of ozone and alpha-pinene. Fit parameters were the rate coefficient of the alpha-pinene ozonolysis and the wall loss of ozone. The functional dependance of the rate coefficient has been expressed by the following Arrhenius parameters: A preexpontial factor of A = (1.5 +/- 0.4)E-15 cm³/s and a temperature dependent factor of E(A)/R = (860 +/- 70) K. The rate coefficient has been determined for the first time for atmospherically relevant temperatures below 276 K