Secondary Organic Aerosol and Ozone Formation Potential from Anthropogenic and Biogenic Volatile Organic Compounds in Ulsan, South Korea, in Summer = 울산 여름철 휘발성 유기화합물의 이차유기입자와 오존 생성 잠재력

다국어 초록 (Multilingual Abstract)

A large quantity of anthropogenic and biogenic volatile organic compounds (VOCs) is emitted in Ulsan because Ulsan has huge multi-industrial complexes in its eastern coastal area and high mountainous regions in the western hinterland. Some of the VOCs are oxidized and form tropospheric ozone and secondary organic aerosol (SOA) in the atmosphere. Not only high temperature and radiation but also the transport of VOCs from their sources could aggravate the photochemical oxidation reactions in the atmosphere in Ulsan in summer. Despite the environmental importance of photochemical reactions of VOCs, few studies on photochemical VOCs in Ulsan have been carried out. The objectives of this study are to investigate the spatial concentration levels of photochemical VOCs, estimate the formation potentials of ozone and SOA, and propose further research to figure out how much VOCs have contributed to high tropospheric ozone and particulate matter episodes in Ulsan, South Korea, in summer.

Hybrid VOC monitoring was conducted with diffusive passive samplers (Radiello, Instituti Clinici Scientifici Maugeri, Itay) and active pumped adsorbent tube samplers (Sequential tube sampler-25, PerkinElmer, UK) at 17 sites (5 industrial, 6 rural, 6 urban sites) and three sites (1 control, 1 industrial, 1 rural site), respectively, in Ulsan from May to August 2020. Through the hybrid VOC sampling, the temporal and spatial resolution of Ulsan VOC monitoring was highly improved. The target VOCs were selected as photochemical assessment monitoring stations (PAMS) ozone precursor 53 VOCs (36 aliphatics and 17 aromatics). Both anthropogenic VOCs like benzene, toluene, ethylbenzene, and o,m,p-xylene (BTEX) and a biogenic VOC, isoprene, were included. All 240 VOC adsorbent samples were analyzed with a thermal desorber-coupled gas chromatography/mass spectrometer (TD-GC/MS, UNITY series 2, Markes, UK-7890B/5977A, Agilent, USA). Every sampling trip had field blank samples to track any contaminations from the whole analytical process. A 3:1 signal to noise ratio was applied to the quantification of VOCs. The concentration spatiotemporal distribution of the VOCs was comprehensively interpreted with the data of 16 meteorological observation stations in Ulsan, considering the physicochemical properties of the 53 VOCs.

From May to August, the atmospheric temperature in Ulsan increased except for in July because July is the rainy period. Heavy rain was observed in July, causing relatively low temperatures and radiation. Due to the land-sea breeze in Ulsan, the transport of VOCs from industrial areas to highly urbanized areas occurred in the daytime while transport of VOCs from mountainous areas to the urban region occurred in the nighttime. Criteria air pollutants (CAPs) in rural and industrial sites were compared. Although the concentration of fine particulate matter (PM2.5) did not show statistically significant differences between the sites, NO2 was higher, and O3 was lower in the industrial site than in the rural sites. Also, O3 and the fine particulate matter to coarse particulate matter (PM10) ratio, whose change could indicate the secondary aerosol formation, showed strong diurnal variations in the rural sites but not in the industrial site. These differences in the concentration levels of NO2 and O3 and the diurnal variation between industrial and rural sites need to be contemplated in a further SOA and O3 formation study in Ulsan.

Total VOCs (TVOCs), BTEX, and aliphatics exhibited significantly higher concentrations in industrial sites than in rural and urban sites. However, isoprene, a well known biogenic VOC (BVOC), showed a higher concentration in rural sites than in industrial and urban sites. This was obvious due to the BVOCs being emitted from the vegetation. In addition, isoprene concentrations had strong diurnal cycles depending on temperature and solar radiation. In order to identify the source and aging status of BTEX in each site, diagnostic ratios were applied to the BTEX concentration in each sampling site. Toluene to benzene ratio and m,p-xylene to ethylbenzene ratio were used as indicators for traffic emission and aging, respectively. BTEX in automobile and shipbuilding industrial areas were highly affected by fresh and non-traffic sources while BTEX in the harborside petrochemical industrial area were mostly influenced by fresh and traffic sources. Most rural and urban sites were affected by aged both traffic and non-traffic BTEX sources.

The top 5 VOC contributors of ozone and SOA formation potentials (OFP and SOAFP) in different sites were compared in this study. In urban and rural sites, toluene, ethylbenzene, and xylenes (TEX) were dominant in the top 5 OFP contributors. In addition to TEX, n-octane and 3-methylpentane significantly contributed to OFP in industrial sites. Regardless of the sites, TEX made the biggest contribution to SOAFP. OFP and SOAFP from VOCs were highest in petrochemical and automobile industrial areas, respectively, in this study. However, the estimated formation potential of O3 and SOA could not explain the spatiotemporal variations of O3 and SOA based on the observed data. To improve the accuracy of the estimations, more VOCs, especially BVOCs, should be included in VOC monitoring and a better methodology to calculate formation potential with meteorological conditions needs to be developed.

In conclusion, TEX largely influenced OFP and SOAFP in Ulsan in the summer. While controls for TEX over the Ulsan need to be enhanced to reduce photochemical oxidation reactions forming O3 and SOA, the study on BVOCs such as isoprenes and terpenes, is also needed due to the lack of understanding BVOCs in Ulsan. In order to improve the estimation of O3 and SOA formation, the key factors, such as meteorological conditions and atmospheric composition, should be investigated and considered in a non-linear way like through a machine learning approach.

목차 (Table of Contents)

• Abstract I
• Contents IV
• List of Figures VI
• List of Tables VIII
• 1. Introduction 1
• 1.1. Environmental importance of volatile organic compounds in Ulsan 1
• 1.2. VOC Monitoring 3
• 1.3. Sources of VOCs 5
• 1.4. Previous studies on the VOCs in Korea 6
• 1.5. Formation of secondary organic aerosol and ozone from VOCs 7
• 1.6. Objectives of this study 9
• 2. Materials and Methods 10
• 2.1. VOC sampling 10
• 2.1.1. Sampling sites 10
• 2.1.2. Active and passive air samplers 11
• 2.1.3. Sampling methods 15
• 2.1.4. Meteorological and criteria air pollution monitoring stations 18
• 2.2. Analysis and Quality Assurance/Quality Control (QA/QC) 19
• 2.2.1. Target VOCs 19
• 2.2.2. TD-GC/MS 25
• 2.2.3. QA/QC 28
• 2.3. Estimation of SOA and ozone formation 30
• 2.3.1. Ozone formation potential (OFP) 30
• 2.3.2. SOA formation potential (SOAFP) 30
• 2.4. Approaches to data analysis 33
• 2.4.1. Comparison of the data distributions 33
• 2.4.2. Source and aging identification 33
• 3. Results and Discussions 34
• 3.1. Criteria air pollutants (CAPs) with meteorological data 34
• 3.1.1. Meteorological conditions 34
• 3.1.2. Spatial distribution of CAP concentrations 37
• 3.1.3. Diurnal cycle in CAP concentrations 42
• 3.2. Spatial and temporal distributions of VOCs 44
• 3.2.1. Spatial distribution of VOC concentrations 44
• 3.2.2. Source and aging identification 50
• 3.2.3. Temporal distribution of VOC concentrations 52
• 3.3. Formation potential of SOA and ozone from VOCs 57
• 3.3.1. OFP 57
• 3.3.2. SOAFP 61
• 3.3.3. Comparison between formation potential and concentration 64
• 4. Conclusions 66
• 4.1. Implications 66
• 4.2. Limitations 68
• 4.3. Proposal for a further study 69
• References 71
• Supplementary Materials 77
• Acknowledgements 88