Abstract:
Aluminum smelter potrooms are unique in that workplace exposures to hydrogen fluoride (HF), sulfur dioxide (SO2), and particulate matter (PM) occur simultaneously for many tasks. Peak exposure to these contaminants have become of increasing interest in discovering the etiology of respiratory health effects. Potroom work environments involve high temperatures, presence of magnetic fields, multiple contaminants and periodic high exposure concentrations creating a harsh environment for direct-reading instrumentation (DRI) operation. The aim of this investigation is to use DRIs to characterize peak exposures to potroom tasks for each contaminate and develop a family of peak exposure metrics that incorporate the magnitude, duration, and frequency of peak exposures. This first required instrument validation of PM and SO2, while monitoring HF required development of instrumentation that can differentiate between HF and SO2. The latter involved the use of two instruments to indicate the HF concentration: one measuring both HF and SO2 ([HF+SO2]) while the second measuring just SO2. Task average exposures to each contaminant were characterized with DRI providing task average concentrations, 15-minute exposures, and peak exposures based on feasible instrument averaging times (10 seconds for PM and SO2 and 5 minutes for HF). Additionally, peak exposure metrics were investigated to identify metrics not correlated with task averages or that altered the task ranking. Task average concentrations ranged from 0.3 to 3.5 ppm, 0.4 to 1.1 ppm, and 0.3 to 1.9 mg/m3 for SO2, HF and respirable PM, respectively. The highest acid gas 15-minute exposures were for carbon stting and line breaking tasks with exposures of 9.5 (SO2) and 2.3 ppm (HF) with peak exposures of 27.9 ppm for SO2 (over 10 seconds) and 2.5 ppm for HF (over five minutes). Respirable PM exposures were highest for carbon setting with task average of 6.3 mg/m3 and peak exposure of 37.1 mg/m3 (10 second). Peak duration and the GSD of the time series were not correlated with the task average concentrations (TWA) while peak frequency, peak magnitude, and metrics incorporating more that one of these measures were highly correlated with the task average.