Intercomparison of concentration results from fine particle sulfur monitors

A one week intercomparison study was carried out to evaluate the ability of selected analytical instruments to measure fine paniculate sulfur concentrations. The instruments compared included five modified flame photometric detection systems and an automated dichotomous sampler that was coupled to an on-line, wavelength dispersive X-ray fluorescence analyzer tuned to measure sulfur on fine particle filters. All instruments were connected to a common duct. Concentrations were obtained in periods ranging from 1 to 30 min, but data were reported and intercompared on an hourly basis. Paniculate sulfur concentrations varied from 1 to 9 μg m ^-3. Results at all concentration levels from four of the six systems agreed to better than ± 5 % throughout the study; while all six agreed within ± 25 % of their composite-mean concentration values at the higher concentrations measured. Linear regression between the composite means and each sampler system's data set showed that the standard estimate of errors ranged from 0.2 to 0.6μgm^-3, correlation coefficients from 0.979 to 0.994, and slopes ranged from 0.88 to 1.22. Sulfate concentrations were also determined every 6 h from the fine particle stage of another dichotomous sampler. They were 26 % lower than expected relative to the aerosol sulfur concentrations; thus it appears that either the aerosol sulfur was not entirely in sulfate form, or sulfate was not completely extractable by conventional procedures.