Search

A new methodology for performing long-term source apportionment (SA) using positive matrix factorization (PMF) is presented. The method is implemented within the SoFi Pro software package and uses the multilinear engine (ME-2) as a PMF solver. The technique is applied to a 1-year aerosol chemical speciation monitor (ACSM) dataset from downtown Zurich, Switzerland. The measured organic aerosol mass spectra were analyzed by PMF using a small (14 d) and rolling PMF window to account for the temporal evolution of the sources. The rotational ambiguity is explored and the uncertainties of the PMF solutions were estimated. Factor–tracer correlations for averaged seasonal results from the rolling window analysis are higher than those retrieved from conventional PMF analyses of individual seasons, highlighting the improved performance of the rolling window algorithm for long-term data. In this study four to five factors were tested for every PMF window. Factor profiles for primary organic aerosol from traffic (HOA), cooking (COA) and biomass burning (BBOA) were constrained. Secondary organic aerosol was represented by either the combination of semi-volatile and low-volatility organic aerosol (SV-OOA and LV-OOA, respectively) or by a single OOA when this separation was not robust. This scheme led to roughly 40 000 PMF runs. Full visual inspection of all these PMF runs is unrealistic and is replaced by predefined user-selected criteria, which allow factor sorting and PMF run acceptance/rejection. The selected criteria for traffic (HOA) and BBOA were the correlation with equivalent black carbon from traffic (eBCtr) and the explained variation of 60, respectively. COA was assessed by the prominence of a lunchtime concentration peak within the diurnal cycle. SV-OOA and LV-OOA were evaluated based on the fractions of 43 and 44 in their respective factor profiles. Seasonal pre-tests revealed a non-continuous separation of OOA into SV-OOA and LV-OOA, in particular during the warm seasons. Therefore, a differentiation between four-factor solutions (HOA, COA, BBOA and OOA) and five-factor solutions (HOA, COA, BBOA, SV-OOA and LV-OOA) was also conducted based on the criterion for SV-OOA. HOA and COA contribute between 0.4–0.7 µg m−3 (7.8 %–9.0 %) and 0.7–1.2 µg m−3 (12.2 %–15.7 %) on average throughout the year, respectively. BBOA shows a strong yearly cycle with the lowest mean concentrations in summer (0.6 µg m−3, 12.0 %), slightly higher mean concentrations during spring and fall (1.0 and 1.5 µg m−3, or 15.6 % and 18.6 %, respectively), and the highest mean concentrations during winter (1.9 µg m−3, 25.0 %). In summer, OOA is separated into SV-OOA and LV-OOA, with mean concentrations of 1.4 µg m−3 (26.5 %) and 2.2 µg m−3 (40.3 %), respectively. For the remaining seasons the seasonal concentrations of SV-OOA, LV-OOA and OOA range from 0.3 to 1.1 µg m−3 (3.4 %–15.9 %), from 0.6 to 2.2 µg m−3 (7.7 %–33.7 %) and from 0.9 to 3.1 µg m−3 (13.7 %–39.9 %), respectively. The relative PMF errors modeled for this study for HOA, COA, BBOA, LV-OOA, SV-OOA and OOA are on average ±34 %, ±27 %, ±30 %, ±11 %, ±25 % and ±12 %, respectively., Atmospheric Measurement Techniques, 14 (2), ISSN:1867-1381, ISSN:1867-8548

Source apportionment using the bilinear model through a multilinear engine (ME-2) was successfully applied to non-refractory organic aerosol (OA) mass spectra collected during the winter of 2011 and 2012 in Zurich, Switzerland using the aerosol chemical speciation monitor (ACSM). Five factors were identified: low-volatility oxygenated OA (LV-OOA), semivolatile oxygenated OA (SV-OOA), hydrocarbon-like OA (HOA), cooking OA (COA) and biomass burning OA (BBOA). A graphical user interface SoFi (Source Finder) was developed at PSI in order to facilitate the testing of different rotational techniques available within the ME-2 engine by providing a priori factor profiles for some or all of the expected factors. ME-2 was used to test the positive matrix factorization (PMF) model, the fully constrained chemical mass balance (CMB) model, and partially constrained models utilizing a values and pulling equations. Within the set of model solutions determined to be environmentally reasonable, BBOA and SV-OOA factor mass spectra and time series showed the greatest variability. This variability represents the uncertainty in the model solution and indicates that analysis of model rotations provides a useful approach for assessing the uncertainty of bilinear source apportionment models., Atmospheric Measurement Techniques, 6 (12), ISSN:1867-1381, ISSN:1867-8548

Trace elements contribute typically only a few percent to the total mass of air pollutants, however, they canaffect the environment in significant ways, especially those that are toxic. Furthermore, they are advantageouswith respect to a refinement of source apportionment when measured with high time resolution and appropriatesize segregation. This approach is especially advantageous in an urban environment with numerous time-variantemission sources distributed across a relatively narrow space, as is typically the setting of a megacity.Two 1-month long field campaigns took place in the framework of the Megapoli project in Paris, France, in thesummer of 2009 and in the winter of 2010. Rotating drum impactors (RDI) were operated at two sites in eachcampaign, one urban, the other one suburban. The RDI segregated the aerosols into three size ranges (PM10-2.5,PM2.5-1 and PM1-0.1) and sampled with 2-hour time resolution. The samples were analyzed with synchrotronradiation induced X-ray fluorescence spectrometry (SR-XRF) at the synchrotron facilities of Paul ScherrerInstitute (SLS) and Deutsches Elektronen-Synchrotron (HASYLAB), where a broad range of elements (Na, Mg,Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Se, Sr, Zr, Cd, Sn, Sb, Ba, Pb) was analyzed for each sizerange.Time series of the analyzed elements for the different sites and campaigns were prepared to characterize theaerosol trace element composition and temporal behavior for the different weather situations and urban environments.They allow for the distinction of regional vs. local sources and transport, and provide a basis for sourceapportionment calculations. Local and regional contributions of traffic, including re-suspension, break wear andexhaust, wood burning, marine and other sources will be discussed. Indications of long-range transport fromPolish coal emissions in the city center of Paris were also found.

Abundant tiny liquid and solid particles with different diameters and chemical compositions suspended in the air are invisible threat of human lives. Particulate matter (PM) also affects the climate and ecosystem directly and indirectly. PM can not only be emitted directly to the atmosphere as primary aerosols but secondary aerosols can also be formed by various aging processes. With 99% of the global population exposed to high PM2.5 (PM with an aerodynamic diameter smaller than 2.5 µm) levels larger than the new WHO air quality guideline (annual average of 5 µg/m3), it is more important than ever to mitigate the PM level more effectively. Therefore, it is extremely crucial to continuously have a high-quality and timely characterization of its sources to design/validate mitigation strategies. This dissertation focuses on developing and implementing novel source apportionment (SA) analyses of PM across Europe. The datasets were collected using different versions of the aerosol chemical speciation monitor (ACSM) and a compact time-of-flight aerosol mass spectrometer (C-ToF-AMS), which typically provide the aerosol composition in real-time with a time resolution of 30 min. One of the major components of PM is organic aerosol (OA). Positive matrix factorization (PMF) has been widely deployed on the OA mass spectrum to resolve its sources and their respective contributions quantitatively. Typically, the multilinear engine (ME-2) has been proven to be a useful tool to control rotational ambiguity of PMF model to provide environmental-reasonable solutions. In addition, a rolling window PMF approach has been introduced to conduct PMF within a smaller time window of 14/28 days and move with a one day. This approach allows the model to capture temporal variations of factor profiles, which has been proved to provide better separations among different factors. With a well-established air quality monitoring network in Europe, a standardized SA protocol has been developed while practicing advanced SA (rolling window PMF approach with ME-2) using 22 ACSM/AMS datasets since 2013 (13 urban and 9 non-urban sites). Overall, this comprehensive protocol works effectively across all sites governed by different sources and generates robust and consistent source apportionment results. In addition, this thesis has provided a comprehensive picture of OA sources in Europe that covers all seasons with a time resolution of 30-240 min. It enables the quantifications of the most common OA components such as hydrocarbon-like OA (HOA, related to traffic, 10.7 ± 3.8%), biomass burning OA (BBOA, 12.4 ± 6.9%), cooking-like OA (COA, 14.5 ± 2.5%), more oxidised-oxygenated OA (MO-OOA, 44.4 ± 12.6%), and less oxidized oxygenated OA (LO-OOA, 27.9 ± 6.5%). Other components such as coal combustion OA (CCOA), solid fuel OA (SFOA: mainly a mixture of coal and peat combustion), cigarette smoke OA (CSOA), sea salt (mostly inorganic but part of the OA mass spectrum), coffee OA, and ship industry OA could also be identified at certain sites. Oxygenated OA (OOA) components make up most of the submicron OA mass (average = 71.1%, range from 43.7 to 100%). Solid fuel combustion-related OA components (i.e., BBOA, CCOA, and SFOA) are also considerable, totalling 16.0% of yearly contribution to the OA, with peak contributions during winter months (21.4%). The second study presents the chemical composition and source apportionment of the submicron particle measurements in Magadino. This study conducted a novel and advanced SA approach by strictly following the standardized protocol to capture the temporal variations of source factors in a long-term dataset. It is a role model for similar analyses, which provides sophisticated OA information in a southern alpine valley. This study resolves three primary sources, an HOA factor, a BBOA factor, and a 58 related OA (58-OA) factor. Moreover, two secondary factors are identified: a MO-OOA and a LO-OOA factor. This study also demonstrated the robustness of this novel SA technique by comparing it with conventional seasonal PMF results. The third study demonstrates the state-of-the-art SA technique to provide robust OA sources in real-time (a few minutes after every scan) using yearlong ACSM datasets collected in three European cities, Athens, Paris, and Zurich. It shows remarkable robustness with an excellent agreement between this technique and the rolling results generated by following the abovementioned protocol. Although the real-time SA requires some a priori information about the sampling site (i.e., an appropriate number of factors and their corresponding factor profiles), this technique minimizes user interactions (requiring high levels of technical expertise). Overall, this PhD thesis developed and implemented state-of-the-art SA techniques on the OA mass spectrum across Europe. It provided high-quality SA results in high-time resolution across Europe with unprecedentedly extensive spatial and temporal coverage. Similar studies should be conducted in other continents to understand global OA sources better. The broad application of the new real-time SA approach allows for a larger spatial coverage of robust OA source information in real-time, which could be revolutionary to the next-generation climate/air quality model and helps authorities design/validate timely actions to successfully reduce the air pollution levels.

Volatile organic compounds (VOCs) contribute significantly to the formation of ozone (O3) and secondary organic aerosol (SOA) in the troposphere. Besides their influence on tropospheric chemistry, VOCs also affect air quality and climate both directly and indirectly. Biogenic VOC (BVOCs) emissions are the predominant source on a global scale, while in the urban areas anthropogenic VOCs are more important. Among them, traffic and solid fuel combustion are important emission sources. After emission, they react with various oxidants, yielding oxygenated VOCs (OVOCs). Due to the high variabilities of VOC sources and reactivity, the determination of VOC sources remains challenging. Due to limitations in the measurement techniques, a large fraction of OVOCs in the urban areas is still poorly understood. In this dissertation, we used a proton-transfer-reaction time-of-flight mass spectrometer (PTR-ToF-MS) in different urban environments to investigate the VOC emission sources and to obtain a deeper understanding of their chemical evolution with support from the positive matrix factorization (PMF) receptor model. Intensive campaigns were carried out at two sites in Delhi, India, as well as an urban site in Beijing, China, and an urban site in Zurich, Switzerland. The measurement campaigns in Delhi aimed at characterizing for the first time a wide range of VOCs in this severely polluted environment. At both the urban and suburban site, high nocturnal concentrations of VOCs were observed with strong day-night variations. Primary emissions were predominant at both sites, mostly from traffic and solid fuel combustion. Large amount of aromatics, furans and phenols were emitted and were significantly elevated at night. The secondary factors, which were dominated by oxygenated VOCs were substantially higher during daytime. A comparison between the two sites showed that primary sources were significant at both sites, however, OVOC concentrations were higher at the suburban site. While the urban OVOCs were tightly associated with aging of local primary emissions, the suburban OVOCs represented both local oxidation and regional aging. The measurement campaign in Beijing was performed to investigate VOC evolution during winter and the potential roles of VOCs to SOA formation under haze conditions. Ten VOC families were classified, which showed clear variation of the relative composition as a function of the VOC concentration. The haze periods were driven by very high concentrations of aromatics originating from traffic and solid fuel combustion. Trajectory analysis indicated that both accumulation and regional influences were important for high VOC concentrations during haze periods. The OVOC factor was predominant during the clean periods, but decreased significantly during the haze periods. Meanwhile, oxygenated organic aerosol (OOA) concentration were enhanced during high VOC concentration periods, indicating the importance of OVOCs to secondary organic aerosol formation. The Zurich campaign focused on the chemical evolution of VOCs and their oxidation products in a relatively clean urban environment in summer. Traffic was the major anthropogenic source, with an additional local source related to activities such as cigarette smoke. Three OVOC factors were identified representing oxidation processes of different oxidants and precursors. Two OVOC factors featured strong contributions from aromatic oxidation and were separated by day-night chemistry. The third OVOC factor was characterized by strong biogenic influences and exhibited similar temporal variation as biogenic SOA. A comparison with the biogenic SOA composition showed similarities in important organic family time series in both the gas and aerosol phase.