Positive bias in particulate matter emissions data due to sulfur dioxide adsorption and oxidation on glass fiber filters

Recent particulate matter emisssion tests at an ArcelorMittal sintering line in Vitoria, Brazil have indicated a change in particulate matter emissions that was not consistent with the steady and proper operation of the combined electrostatic precipitator and pulse jet fabric filter control system. ArcelorMittal Brazil, ERM Brasil, and Air Control Techniques, P.C. have evaulated the use of borosilicate glass fiber filters in the U.S. EPA Method 5 sampling train as a possible reason for the anomolous change. The scope of this study included (1) possible adsorption and oxidation of SO2 on borosilicate glass and quartz filter media, and (2) the possible loss of fibers from borosilicaate glass and quartz filter media at normal sampling train vacuums and temperatures. Air Control Techniques, P.C. conducted a series of laboratory tests replicating gas stream conditions using both borosilicate and quartz circular and thimble filters. The laboratory tests confirmed the formation of sulfate artifacts due to sulfur dioxide adsorption on the borosilicate filters. These tests also confirmed that neither the borosilicate glass-fiber filters or the quartz filter lost significant filter weight during sampling at normal Method 5 sampling train vacuum levels and filter temperatures. Tests conducted by ERM Brasil, SA on the ArcelorMittal sintering process stack using both quartz and borosilicate glass fibers also indicated significant adsorption and oxidation of sulfur dioxide on the borosilicate filters. These laboratory and field tests confirmed the positive bias associated with glass fiber filters in SO2 environments reported as early as 1970. The authors recommend that only quartz fiber filter be used for particulate matter testing in SO2-containing gas streams. Furthermore, EPA Method 5 Section 7.1.1 should be revised to explicitly approve the use of quartz fiber filters. Implications: Filterable particulate matter emission tests using EPA Method 5 in sulfur dioxide laden gas streams have had a large positive bias due to outdated sampling practices. Our manuscript provides bias quantification data from both laboratory tests and tests on a sintering process stack. We recommend that (1) particulate matter emission tests conducted in SO2-containing gas streams be conducted using only quartz filter and (2) The U.S. EPA change the definition of filter media to be used in EPA Method 5 and other particulate matter test methods based on Method 5 specifically allow the use of quartz filters.