안산지역 대기 중 다이옥신 및 dl-PCBs의 오염특성 조사

Abstract

After establishment of Banwol industrial complex in 1987, Ansan city becomes the largest industrial sector development in Gyeonggi-do, Korea. As the population and industrial activity grow over this region, toxic air pollutants, particularly POPs (Persistent Organic Pollutants) from various emission sources have been major public concerns. Air samples for POPs monitoring were collected at the industrial sites (<TEX>$A_2$</TEX>), residential sites (<TEX>$B_1$</TEX>, <TEX>$B_2$</TEX>), commercial site (C), and rural/remote site (D) of the area of Ansan during 2008 with a prolonged industrial sampling site <TEX>$A_1$</TEX> from 2001 to 2008. All samples were analysed for 2,3,7,8 substituted-polychlorinated dibenzo-p-dioxin and dibenzofurans (PCDD/Fs) and dioxin like polychlorinatd diphenyls (dl-PCBs). In site <TEX>$A_1$</TEX>, a steady decline of their concentrations from 2003 to 2008 was observed due to the reinforced emission guideline from waste incinerators. The average concentration of the PCDD/Fs and dl-PCBs ranged between 0.118 pg-TEQ/<TEX>$m^3$</TEX> (rural/remote site D) and 0.532 pg-TEQ/<TEX>$m^3$</TEX> (industrial area <TEX>$A_2$</TEX>). These level were generally consistent with previous studies in Gyeonggi-do, while higher than other places. Most of PCDD/Fs congener were partitioned into particle phase, whereas dl-PCBs were partitioned into gas phase. The logarithm of gas-particle partition coefficient <TEX>$K_P$</TEX> of dl-PCBs and PCDD/Fs were well correlated with sub-cooled liquid vapor pressure <TEX>$P_L$</TEX>. The slope <TEX>$m_</TEX><TEX>T$</TEX> of log <TEX>$K_P$</TEX> versus log <TEX>$P_L$</TEX> for PCDD/Fs (-1.22) and dl-PCBs (-1.02) in industrial area (<TEX>$A_2$</TEX>) were high compared to other residential/commercial area. It suggests that this area was likely influenced by the direct emission source of PCDD/Fs and dl-PCBs. To simulate the partition of PCDD/Fs and dl-PCBs between gas and particle phase, Junge-Pankow model (<TEX>$P_L$</TEX>-base) and <TEX>$K_{oa}$</TEX> model were applied. It was found that J-P model was more suitable than the <TEX>$K_{oa}$</TEX> model in this study.