Improvements to a Dual Channel Atmospheric Mercury Measurement System and Development of Methods to Characterize Oxidized Mercury Emitting Permeation Tubes Using a GC/MS

Abstract:

KCl-coated denuders have been used to measure atmospheric oxidized mercury in the past, but these systems are biased low, recovering only 30-50% of injected oxidized mercury compounds as oxidized mercury. The amount of bias is also dependent on location and atmospheric conditions. New technology is critically needed to replace KCl denuder-based measurement systems.

We developed a dual channel atmospheric mercury measurement system that fills this gap. The system consists of a Tekran 2537 analyzer that alternately samples air that passes through a pyrolyzer and air that passes through a series of cation-exchange membranes. The pyrolyzer reduces all mercury compounds to elemental mercury, providing a measure of total mercury in the atmosphere. The cation exchange membrane channel measures elemental mercury, and we calculate oxidized mercury as the difference between these two.

We use a permeation tube-based automated calibration system to verify the performance of the dual channel measurement system in ambient air. In order to do this, accurate characterization of the permeation tubes is crucial. We characterize the tubes using two different methods: gravimetrically using a microbalance scale to determine mass loss from the perm tubes, and by routing the output from the perm tubes into a GC/MS to quantitatively determine the composition of emitted mercury compounds.

We deployed the calibration system with the dual channel system to determine the accuracy and stability of both systems. The dual channel system can achieve detection limits for oxidized mercury of 10-20 pg/m³ (hourly averages), though the detection limit is worse if elemental mercury is rapidly changing. Unlike KCl denuder-based instruments, the dual channel system was able to collect oxidized mercury compounds (HgCl₂ and HgBr₂) injected into ambient air by the calibrator without any bias. However, comparison of data obtained from the University of Nevada, Reno’s RMAS system shows that we may be experiencing a loss of oxidized mercury (possibly non-halide compounds or particle-bound mercury) in our sampling line. We have recently made improvements to our system in order to mostly eliminate the sampling line and minimize any loss.