Methods for polynuclear aromatic hydrocarbon determination in air samples using polar-bonded phase HPLC and GC-MS with application to oil refinery samples

Abstract

The first part of this work deals with the development of accurate and reliable procedures for collection and analysis of polynuclear aromatic hydrocarbons (PNAs). The second part shows the application of these procedures to air samples collected in and around oil refineries. Particle samples were collected using high volume air samplers fitted with glass fiber filters or with a cascade impactor containing paper filters. They were then cleaned using either extraction with dimethylsulfoxide and pentane or utilizing a small cartridge containing a diamine polar-bonded phase material, the second method being more effective. Vapor phase PNAs were sampled using an apparatus designed in our laboratory. After collection, the resins were desorbed with solvent and the PNA content was determined. The suitablility of the resins evaluated for PNA recovery from ambient air using these methods decrease in the following order: Amberlite XAD-2, Chromosorb 105, Tenax GC, coconut charcoal, and Ambersorb XE-348. High performance liquid chromatography (HPLC) was used to determine the behavior of PNAs in the normal and reversed phase on polar-bonded phases containing amine, diamine, and pyrrolidone substrates. Results support the proposed mechanism in the normal phase and indicate that both a partitioning and liquid-solid adsorption mechanism takes place in the reversed phase depending upon the mobile phase. Occasionally, these polar-bonded phases can be deactivated by the formation of amine-carbonyl complexes from polar aldehydes or ketones in the solvent or sample. Deactivation can be reversed by flushing with water to hydrolyze the Schiff 's base imine back to the amine. Gas chromatography-mass spectroscopy (GC-MS) was used to analyze air samples from oil refineries in Port Arthur, collected over a period of three years. The analytical procedures are applied to the collected samples to determine if they contain detectable amounts of PNAs. The GC-MS analysis was adequate for this study but the use of SIM detection is preferred because of the greater sensitivity for PNAs.