PM2.5 samples were collected from a semi-rural site in the industrial city of Ulsan, South Korea during 2018–2019 to characterize the chemical composition and determine source contributions on polluted and normal days. The daily PM2.5 concentration exceeded the United States National Ambient Air Quality Standards (35 µg/m3) on 11 of the 56 days. The mean concentration of Σ9 water-soluble inorganic ions (WSIIs) was 17.1 and 6.07 µg/m3 on polluted and normal days, respectively, with SO42–, NO3–, and NH4+ the dominant species. The elements were dominated by Mg and Fe, with a mean concentration of Σ11 elements of 1973 and 855 ng/m3 on polluted and normal days, respectively. The mean concentration of Σ21 polycyclic aromatic hydrocarbons (PAHs) was 4.44 and 3.26 ng/m3 on polluted and normal days, respectively, indicating a higher human health risk on polluted days. High PM2.5 concentrations were due to an increase in the concentration of almost all WSIIs, elements, and PAHs on polluted days. Potential sources of PM2.5 were identified using positive matrix factorization (PMF), the conditional bivariate probability function (CBPF), and the potential source contribution function (PSCF). According to PMF results, secondary inorganic aerosols (39.8%), primary sulfate (18.2%), and coal/biomass burning (16.3%) were dominant on polluted days, whereas marine sources (23.1%), secondary inorganic aerosols (19.3%), and primary sulfate (16.2%) were the largest contributors to PM2.5 on normal days. Results from the CBPF and PSCF indicate that polluted days may be influenced by both long-range atmospheric transport (LRAT) and local emission sources, while LRAT became more important on normal days. Based on these findings, other studies are necessary to further investigate and distinguish the PM2.5 source contributions in industrial and urban areas. [ABSTRACT FROM AUTHOR]