This study presents the results of regular PM2.5 (aerodynamic diameters equal to or smaller than 2.5 μm) collection at the Banqiao, Zongming, Douliu, Chiayi, Xiaogang, and Hualien sites in Taiwan every six days from January to November in 2018 (with the additional December 2017 data). PM2.5 mass, water-soluble inorganic ions, carbonaceous contents, and metal elements were resolved for further analysis. In addition, the work also assessed the comparability of metal element monitoring from an automated instrument and established a new organic compound analytical method. Utilizing the analyzed water-soluble inorganic ions, carbonaceous contents, and metal elements, this study investigated the characteristics of temporal and spatial distributions of PM2.5 mass and chemical components, apportioned source contributions, and evaluated atmospheric visibility influencing factors. The results showed that PM2.5 mass levels increased in the order from east (10 μg m-3), north, to the south (24 μg m-3) of Taiwan during the sampling period. The season of highest PM2.5 mass levels at the Banqiao, Zongming, and Hualien sites were in spring, while that of the Douliu, Chiayi, and Xiaogan sites were in winter. Summer was the lowest season of PM2.5 mass levels at all sites. In general, SO42- and organic carbon were the two most abundant components at all sites except NO3- at the Douliu, Chiayi, and Xiaogan sites in winter. Among high, medium, and low concentration groups of metal elements, high concentration group was with natural origins, while medium concentration group was with contributions mostly from fuel oil and coal burnings, metal refining, and traffic emissions. In contrast, the low concentration group was with characteristics of coal burning, metal surface coating, and brake and tire wearing. The volatilizations of NO3- and Cl- were the highest in summer and spring, respectively, while NH4+ was relatively stable across seasons. Influencing volatilization factors include ambient temperature, pollution sources, and chemical compound forms of volatilized components. Meanwhile, positive interferences of volatilized organic carbons adsorbed by quartz-fiber filters varied less than negative interferences from volatilization of the collected carbonaceous particles. Volatilization of the collected carbonaceous particles were primarily under the influences of pollution sources and events. For the trace contents of organic compounds, this study successfully utilized gas flow modulator as a core with a full fledge of two-dimension gas chromatography coupled with time-flight mass spectrometry. The Qualitative organic contents of PM2.5 comprised mostly nitrogenous and phthalate compounds at the Banqiao site. During the sampling period, events with PM2.5 greater than 35 μg m-3 tended to increase from north to south. Most events were under the influences of bad environmental ventilation and local source emissions and frequently originated from previous night of the sampling day. The events often got worse accompanied by regional pollution transport on the same day. In addition, high ozone concentration frequently occurred in the pollution events of late winter and early spring. From the resolved component ratios in PM2.5, NO3- was the sole component increased greatly in all high PM2.5 events in contrast to non-event samples, which was consistent with the findings over the past three years. For seasonal variations, the enhancements of SO42-, NO3-, and fuel oil- and coal-burning tracers were predominant in winter. However, in spring, NO3-, Cl , EC, traffic emissions, fuel oil- and coal-burning tracers increased prevalently but with less SO42- increase. Similarly, the enhancements of NO3-, Cl , biomass burning, industrial sources, and traffic emissions were significant but with less SO42- in fall. For source apportionment using positive matrix factorization, “sulfate and combustion emissions” and “nitrate and combustion emissions” were the two most significant factors at all sites. The factor of “sulfate and combustion emissions” was dominant at the Banqiao, Zongming, and Hualien sites in contrast to the dominance of “nitrate and combustion emissions” at the Douliu, Chiayi, and Xiaogang sites. This implies regional transport is prominent in the areas north to Taichung City and Hualien County, while accumulated pollutants derived from local sources are important in the areas south to Yulin County. To estimate atmospheric light extinction coefficient (bext), this study inserted PM2.5 chemical components, gas species, and relative humidity (RH) into the revised Interagency Monitoring of Protected Visual Environments (IMPROVE) equation. Sulfate contributed bext the most of all components at all sites except nitrate at the Douliu, Chiayi, and Xiaogang sites in winter. Similarly, SO42- was the most important factor in influencing atmospheric visibility from statistical regression analysis, which was consistent with the finding in the revised IMPROVE equation. The regression analysis of metal elements between automated and manual collection methods showed that metal elements with good correlation were those from sources of traffic emissions (Zn、Ba、V、Cu), coal burning (Pb), fuel-oil burning (V), biomass burning (K), and metal smelting (Mn、Cu、Pb). As those elements are abundant in the atmosphere, the automated instrument is still valuable in qualitative identification of pollution sources. Looking into the causes of pollution events and control measures, fuel burning emissions and stagnant environment are main causes of pollution. The control of NOx and ozone precursors will help in reducing high PM2.5 events. Stringent control on evening emissions of public and private sectors is necessary. The origins of nitrogenous and phthalate compounds are of concern.

PM2.5 chemical component monitoring, Characteristics of temporal and spatial distributions of PM2.5 chemical components, PM2.5 source apportionment and visibility influencing factors, Assessment of automated monitoring of metal elements