National Technical University of Athens (NTUA)
Characterization of the sources and processes of ambient particulate matter with high-end mass spectrometry
Abstract
dc:descriptionAmbient aerosols are defined as solid or liquid particles suspended in the air. The impact of air pollution on both the human health and the environment has been well established and studies in these fields are increasingly crucial. Epidemiological studies have linked air pollution to cardiovascular and respiratory diseases. These particles can be either primary (directly emitted from a source) or secondary (formed through reactions involving pre-existing particles or by converting gases into particles). Based on their size, aerosols are categorized as PM1, PM2.5, and PM10, corresponding to diameters up to 1, 2.5, and 10 micrometers, respectively. Depending on their diameter, airborne particles can penetrate various depths of the respiratory system; larger particles reach the upper respiratory tract, while smaller particles (up to 0.1 micrometers in diameter) can penetrate the alveoli and enter the bloodstream. Concerning their environmental impact, airborne particles influence climate directly by scattering or absorbing radiation and indirectly through aerosol-cloud interactions, ultimately affecting the Earth’s radiative forcing. This thesis investigated the chemical composition of airborne particulate matter, focusing on organic aerosols (OA) — that remain the most uncertain fraction of ambient aerosol— through three approaches. First, a modelling technique was applied to identify OA sources in an urban background environment, while integration of the inorganic ions improved the model. Next, the study examined OA at a high-altitude background location, where both organic and inorganic aerosol sources were analyzed, while also considering aerosol-cloud interactions. Finally, a novel instrument designed to improve the measurement of volatile organic compounds (VOCs) —precursors to secondary organic aerosols— was evaluated under both laboratory and field conditions. The need for continuous, real-time, high-resolution measurements of aerosol chemical composition has led to the development of instruments like the Time of Flight Aerosol Chemical Speciation Monitor (ToF-ACSM, Aerodyne). The ACSM measures the chemical composition of PM1 particles, which evaporate at around 600°C. These particles include organic matter, as well as inorganic ions such as SO4²⁻, NO3⁻, NH4⁺, and Cl⁻. The instrument operates by sampling particles passing through a Nafion drier, concentrating the particles using an aerodynamic lens, and then introducing them into a vacuum chamber where they impact on a tungsten plate at 600°C. The particles get vaporized, and then get ionized by electron impact ionization from a tungsten filament at 70 eV. Finally, the ions enter the time-of-flight chamber, where an electric pulse periodically accelerates ion packets, which are then detected and classified by their mass according to the time needed to reach the detector, providing mass spectra and concentrations of these chemical components. A common approach for analyzing this kind of data is source apportionment, typically for the organic components, using models such as the Positive Matrix Factorization (PMF). In brief, PMF assumes that the total mass concentration of PM1 equals the sum of the contributions of its various sources. The model is described by the equation: X = F G + E Eq. 1where X represents the input concentration time series, F and G are the profiles and the temporal variability of each factor, respectively, and E is the residual matrix. Within this doctoral research, data from a ToF-ACSM instrument were analyzed from two stations of the Environmental Radioactivity & Aerosol Tech. for Atmospheric & Climate Impacts, (INRaSTES) of the National Center for Scientific Research “Demokritos.” The first station is an urban-background station, located in Agia Paraskevi, 8 km from Athens’ city center and at the foothills of Mount Hymettus. The second station is a high-altitude background station, located at the top of Mount Helmos in Kalavryta. Other instruments used alongside the ACSM included an aethalometer for real-time equivalent Black Carbon (eBC) mass concentration measurements, a Mobility Particle Size Spectrometer (MPSS) for aerosol size distribution, and a PICARRO instrument for greenhouse gases measurements. NOx and O3 analyzers, managed by the Ministry of Environment and Climate Change, provided the concentrations of NOx (NO + NO2) and O3 at the Demokritos Athens station. The first study of this thesis applied a methodology for source apportionment of the total PM1 fraction (organic matter, SO4, NH4, NO3, and Cl) measured by the ToF-ACSM. The Positive Matrix Factorization (PMF) model was applied to the combined organic and inorganic species, including for the inorganic species the m/z variables listed in Table 1.The model was applied to an annual dataset (October 2017 - November 2018) at the suburban DEM station. At first, the conventional methodology was followed, in which only the organic variables were introduced in the model. Subsequently, the new methodology combining all species was implemented. In Figure 1, the profiles of each organic factor retrieved are shown, including their temporal variability and daily fluctuations for each season. Similarly, Figure 2 presents these profiles for the combined solution. Five factors were identified with the first model: three of them represented primary particles, with mass spectra corresponding to Hydrocarbon-related Organic Aerosol (HOA) from traffic emissions, Cooking Organic Aerosol (COA), and Biomass Burning Organic Aerosol (BBOA). The remaining two mass spectra are one More Oxidized Oxygenated Organic Aerosol (MO-OOA) and one Less Oxidized Oxygenated Organic Aerosol (LO-OOA), representing higher and lower oxidation states, respectively, based on the fraction of the variable at m/z 44 of the organic component, which indicates the oxidation level of each aerosol type. The combined solution identified two additional factors consisting mainly of inorganic variables, ammonium sulfate and ammonium nitrate (named as AmSul & AmNi, respectively). Additionally, the secondary organic aerosol factors included both organic and inorganic components, leading to a Less Oxidized Aerosol (LOA) and a More Oxidized Aerosol (MOA) factor. LOA exhibited significant inorganic contributions and a seasonal variation in the organic-to-inorganic ratio. More specifically, Figure 3 displays the contribution of each species in the two oxidized aerosols through the seasons. As shown, MOA has a stable composition dominated by organic compounds, while LOA displays seasonal variability. Another interesting observation from these results involves the formation of ammonium nitrate and its daily variation. According to the reaction of ammonium nitrate formation (Equation 2):NH3 (g) + HNO3 (g) -> NH4NO3 (s) Eq. 2 when the relative humidity of the atmospheric air exceeds the deliquescence relative humidity (DRH) of ammonium nitrate (62%), ammonium nitrate formation is favored. However, when the relative humidity falls below this value, the reaction in Eq. 2 shifts to the left, making ammonium nitrate inversely proportional to temperature. The two solutions were compared based on the uncertainty of each factor, which was defined here as the ratio of the interquartile range of each factor (75th-25th) to each median concentration. As shown in Table 2, the solution of combined PMF demonstrated overall lower uncertainty, suggesting that this combined approach provides a more comprehensive picture of the sources contributing to the PM1 fraction components, with slight improvement of the overall model. The second study of this thesis focused on measurements from a high-altitude background location at the Eastern Mediterranean region, where aerosol-cloud interactions can be studied. These interactions contribute the highest uncertainties in climate models. The measurement station for this study was the Helmos Hellenic Atmospheric Aerosol and Climate Change (HAC)² station of Demokritos, located at the top of Mount Helmos at 2314 m above sea level. This site is representative of background pollution in the Eastern Mediterranean. The intensive CALISHTO (Cloud-Aerosol InteractionS in the Helmos Background Troposphere) campaign took place in the fall of 2021, aiming to investigate aerosol-cloud interactions using a large suite of instrumentation. The ToF-ACSM instrument was deployed there for yearlong real-time measurements of the chemical composition of atmospheric aerosols (September 2021 - September 2022). This enabled a detailed analysis of aerosol chemical composition at a high altitude, where measurements are often conducted in the Free Troposphere (FT), above the Planetary Boundary Layer (PBL). A combined PMF model was applied again, yielding five factors for the PM1 fraction at the (HAC)² station. To determine the PBL height and distinguish air masses originating from the PBL from free tropospheric air masses, a synergy of remote sensing techniques, and in-situ observations were sued. More specifically, a HALO Wind Doppler lidar, which provides the vertical distribution of wind speed, was installed at the temporary site “Vathia Lakka”, at 500 m below (HAC)², enabling PBL height retrieval. A set of metrics was established after comparison with HALO and to compensate for periods when remote sensing data were not available. These metrics included the eBC/CO ratio, with a threshold of 0.5 for FT periods, the total number concentration of particles in the accumulation mode, with a threshold of 100 #/cm³, and the water vapor mixing ratio, with a threshold of 3.2 g/kg (Figure 4).Finally, another condition studied was the presence of clouds at the station during sampling, verified by Liquid Water Content (LWC) measurements. For periods when measurements were conducted within clouds, aerosol particles were classified into "activated droplets" (particles converted into cloud droplets) and "residual aerosol" (non-activated particles) using the "virtual filtering" technique (Foskinis et al., 2024).Aerosol source identification was performed using PMF analysis, dividing them into five factors: primary organic aerosols (POA), ammonium nitrate (AmNi), ammonium sulfate (AmSul), and two secondary organic aerosol (SOA) factors, less oxidized aerosol (LOA) and more oxidized aerosol (MOA). During the intensive CALISHTO fall campaign, results indicated that organics were the dominant species, particularly in September, due to the station's greater exposure to the ABL. However, sulfate levels increased in October and November. Furthermore, during cloudy periods, sulfate ions exhibited the highest activation rates in cloud droplets, reaching 84% (Figure 5).Throughout the annual cycle, results showed significant seasonal variation, with organic concentrations peaking during summer. Overall, PM1 levels were ten times higher in summer compared to winter, attributed to the increased height of the ABL and fewer cloud occurrences. In contrast, the AmNi factor displayed an opposite trend, increasing in winter due to lower temperatures. A notable finding was the increase in aerosol originating from black carbon during fall when aerosols were representative of the FT. This is potentially significant for climate impact, as black carbon is a strongly light-absorbing component. During cloudy periods in fall, spring, and summer, particle concentrations generally decreased compared to clear conditions. However, in winter, this pattern was more complex due to the distinct cloud formation processes, resulting in similar or even higher concentrations during cloudy periods (Figure 6). Statistically significant increases in specific aerosol factors were observed within clouds during winter: POA and AmSul within the ABL, and AmNi and MOA within the FT. Finally, the origin of the air mass was found to significantly affect the concentrations and the absolute contribution of PM1. Air masses from continental Europe (mainly Central and Eastern Europe) were associated with higher pollutant levels, while air masses from North Africa showed higher levels of sulfate ions and black carbon. The last study of the doctoral research focused on the evaluation of a new mass spectrometer (ToF-MS) under laboratory conditions, as well as through field measurements. The new instrument uses low-temperature plasma for ionizing VOCs. Ionization occurs simultaneously from ions present in the system, such as NO+ and O2+. The instrument has high resolution, equal to 20,000 fwhm at m/z 250. A significant linear correlation in the concentration range of ppb (R² > 0.99) was found during the calibration of the instrument for 12 VOCs (Acrolein, α-pinene, benzene, chlorobenzene, dichlorobenzene, ethylmethylketone, isoprene, methacrolein, o-xylene, and toluene), while low detection limits reaching 1.4 ppt for compounds like benzene were also detected (Figure 7).To study the fragmentation pattern of the instrument (the breakdown of the compound being measured during ionization), experiments were conducted in a simulation chamber where the ozonolysis of three sesquiterpenes (β-caryophyllene, α-humulene, and δ-cadinene) was investigated. These compounds have the same mass and ionization method but different fragmentation patterns. The results were encouraging for the new mass spectrometer, which detected the precursor compounds, their expected fragments verified by online libraries, as well as their expected oxidation products from ozone. Finally, field measurements with the new mass spectrometer were carried out in real-time. These measurements were taken at the Demokritos station in Agia Paraskevi (DEM Athens), in parallel with other instruments (ACSM-ToF, Aethalometer AE33, and NOx and O3 analyzers) to evaluate its performance. The analysis focused on the detection of BTX (benzene, toluene, and xylene).Figure 8 shows the daily variations of the compounds detected by the new mass spectrometer, with external data (PMF factors for organic particles, black carbon fractions related to biomass burning (eBCbb) and fossil fuel combustion (eBCff), and NOx). As seen, toluene corresponds with indicators related to vehicle movement (HOA, eBCff, and NOx), while benzene only follows the morning rush hour peak and then does not increase again in the afternoon. This agrees with previous publications suggesting that benzene undergoes more significant atmospheric dilution compared to toluene, thus showing different daily variation. These results demonstrate that the new ToF-MS is reliable for real-time VOC measurements in different environments. Future steps include further development of the instrument's software for more automated measurements, as well as the study of additional VOCs in field measurements and the source apportionment of these compounds.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Zografou, Olga
- Ζωγράφου, Όλγα
Subjects
dc:subject × 17- Ατμοσφαιρικά αερολύματα
- Ατμοσφαιρική χημεία
- Φασματοσκοπία μάζας χρόνου πτήσης
- Φασματοσκοπία μάζας αντίδρασης μεταφοράς πρωτονίων
- Σταθμός υποβάθρου
- Θετική παραγοντοποίηση πινάκων
- Atmospheric aerosols
- Atmospheric chemistry
- Proton transfer reaction mass spectrometry
- High Altitude Platform Stations (HAPS)
- Positive matrix factorization
- Χημεία
- Φυσικές Επιστήμες
- Χημεία, άλλοι τομείς
- Chemical Sciences
- Natural Sciences
- Chemistry, miscellaneous
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/58255