Atmospheric gas and particle phase concentrations of polychlorinated biphenyls (PCBs) in a suburban site of Bursa, Turkey


CİNDORUK S. S. , Tasdemir Y.

ENVIRONMENTAL FORENSICS, vol.9, pp.153-165, 2008 (Peer-Reviewed Journal) identifier identifier

  • Publication Type: Article / Article
  • Volume: 9
  • Publication Date: 2008
  • Doi Number: 10.1080/15275920801888442
  • Journal Name: ENVIRONMENTAL FORENSICS
  • Journal Indexes: Science Citation Index Expanded, Scopus
  • Page Numbers: pp.153-165
  • Keywords: back trajectory, Bursa, dry deposition, Junge-Pankow model, partitioning, PCBs, POLYCYCLIC AROMATIC-HYDROCARBONS, AIR PARTITION-COEFFICIENT, WATER-SURFACE SAMPLER, ORGANIC-COMPOUNDS, VAPOR-PRESSURE, DRY DEPOSITION, LAKE-MICHIGAN, GREAT-LAKES, ORGANOHALOGEN PESTICIDES, TEMPERATURE-DEPENDENCE

Abstract

Ambient air samples were collected between July 2004 and May 2005 with a high-volume air sampler (HVAS) in order to investigate the gas and particle concentration levels of polychlorinated biphenyls (PCBs) in a suburban site (the Uludag University Campus) in Bursa, Turkey. Average particle and gas-phase PCB concentrations (average +/- standard deviation) were 85.8 +/- 127.8 pg m(-3) and 328.1 +/- 284.2 pg m(-3), respectively. Total (gas+particle) PCB concentrations were higher in the warmer periods, while an opposite case was observed for the colder periods. The application of the Clausius-Clapeyron equation representing the relationship between gas-phase PCB concentrations and air temperature yielded a significant correlation. Moreover, the relationships between PCB congeners and meteorological variables were investigated. Based on meteorological variables and back trajectory results, evaporation and atmospheric transport were assumed to be the major sources of PCBs in the sampling site. PCBs are partitioned between gas and particle phases, and this fate was examined using different available approaches such as logK(P)-logP(Lo), logK(P)-logK(OA), and Junge-Pankow model. Shallow slopes between -0.18 and -0.48 were obtained from the logK(P)-logP(Lo) plot, while the slope of logK(P)-logK(OA) ranged from 0.2 to 0.42. P-Lo and K-OA values obtained from the literature were both considered good descriptors for determination of K-P experimental partitioning coefficient in the site. Correlations were significant at p<0.01 for all approaches.