OCR Text |
Show used; however, trains such as the carbonyl train (Method 0011) should not use a filter and collection of large amounts of particulate in the impingers requires modification of analytical procedures. Another complication in many cases when sampling at the inlet to control devices is the availability of sampling ports. At the Cardinal plant, the ports available at the inlet to one of the control device provided for vertical probe sampling only. This presented a significant problem for the EPA Method 29 sampling train because the amount of particulate collected over the 4-hour ampling period would be large. EER overcame this by developing an in-stack filter arrangement (Figure 3) that prevented loss of collected particulate matter. The assembly worked well and total metals concentrations compared well to concurrent measurements made using a conventional Method 29 train at a parallel sampling location; however, less quantity of volatile metals (mercury, selenium, boron) was collected on the filter fraction of the in-stack probe system. This is believed to be due higher filter temperature using the in-stack assembly, which increases the amount of these metals in the vapor phase. High S02 concentrations create problems with nearly every measurement method. S02 can consume reagents such as hydrogen peroxide and potassium hydroxide prematurely. The amount of reagent required to prevent depletion should be calculated based on expected S02 concentrations. This usually indicates an increase in the amount of reagent normally added to the trains. In general, EPA has indicated they would prefer to see the quantity of a prepared solution increased to compensate, rather than increasing the strength of solutions. For example, at the Cardinal plant EER increased the amount of HN031H202 solution in the second impinger of the EPA Method 29 train by a factor of five to prevent depletion. This required use of a "jumbo" twoliter impinger in this position. High S02 also can interfere with the sampling and analysis of organics. High S02 levels can interfere with purge-and-trap GC analysis used for VOCs, and it has been reported that recovery of surrogate standards from the XAD-2 resin used in the modified method 5 trains for SVOCs can be poor. DISCUSSION Quality Assurance/Quality Control (QA/QC) Overall, the QNQC results obtained using these methods were excellent, with only isolated instances of questionable results. Depending on the laboratory performing sample analysis, the 8 |