Source apportionment of PM10 at residential and industrial sites of an urban region of Kolkata, India

PM10 and its chemical species mass concentrations were measured once in a week at residential (Kasba) and industrial (Cossipore) sites of an urban region of Kolkata for a period of 24 h during November 2003 to November 2004. At each monitoring site, 53 sets of daily average PM10 samples were collected during the study period. Approximately 55% of the monitoring days are weekdays, while 45% are weekends. The PM10 mass concentrations ranged from 68.2 to 280.6 μg m− 3 at the residential site, and 62.4 to 401.2 μg m− 3 at the industrial site. Polycyclic aromatic hydrocarbon compounds (PAH), fluoranthene (Fl), pyrene (Py), benzo(a)anthracene (BaA), benzo(b)fluoranthene (BbF) and benzo(a)pyrene (BaP) have been analyzed using Gas Chromatoghaphy. Metals in PM10 deposited on quartz microfibre filter papers were measured using an Inductively Coupled Plasma-Atomic Emission Spectrometer. Chromium (Cr), zinc (Zn), lead (Pb), cadmium (Cd), nickel (Ni), manganese (Mn) and iron (Fe) are the seven toxic trace metals quantified from the measured PM10 concentrations. Total carbon (TC), inorganic carbon (IC) and organic carbon (OC) were analyzed using a Carbon analyzer. Exposed quartz microfibre filter papers were also analyzed for water-soluble anions of fluoride (F−), chloride (Cl−), nitrate (NO3−), phosphate (PO43−) and sulfate (SO42−) using ion chromatography. In this study, principal component analysis (PCA)/absolute principal component scores (APCS) model was applied to the mass concentrations of PM10 and its chemical species. Principal component analysis with varimax rotation identified five possible sources; solid waste dumping, vehicular emission, coal combustion, cooking and soil dust at residential site. The extracted possible sources at the industrial site were vehicular emissions, coal combustion, electroplating industry, tyre wear and secondary aerosol. A quantitative estimation by principal component analysis–multiple linear regression (PCA–MLR) model indicated that solid waste dumping contributed 36%, vehicular emissions 26%, coal combustion 13%, cooking 8% and soil dust 4% at the residential site. The PCA–MLR apportioned 37% to vehicular emissions, 29% to coal combustion, 18% to electroplating industry, 8% to tyre wear and 1% to secondary aerosol at the industrial site. Due to the limitation in source marker species analyzed, 13% at Kasba and 7% at Cossipore could not be apportioned to any possible sources by this technique.