Abstract Sampling and analytical procedures used to determine total mercury content in crude oils were examined. Three analytical methods were compared with respect to accuracy, precision, and detection limit. The combustion method (U.S. EPA 7473 hybrid) and a commercial extraction method (non-standard) were found adequate to provide a good combination of sensitivity and accuracy, while instrumental neutron activation analysis was found to suffer from interferences from elements other than mercury (Ni, Se) that are typically found in crude oil. Aliquot removal and processing steps were found to be important to accuracy and precision of the combustion method. Potential errors caused by incomplete homogenization and loss of volatile mercury were investigated in detail. Homogenization of oils that contain suspended mercury compounds was found important to allow retrieval of representative aliquots for analysis. Ultrasonication at slightly elevated temperature yielded the more reproducible and higher magnitude results. It was discovered that volatile mercury can be lost from sample containers when they are opened to remove aliquots. Alternative sample containers and aliquot removal techniques were developed that minimized loss of volatile mercury. Speciation experiments were utilized to determine the identity of volatile species and the rates at which they are lost from containers. Both elemental mercury and dimethylmercury were identified as volatile mercury components of one freshly sampled crude oil.
Read full abstract