ARCHIVED REPORTS_1988_2 - Laserfiche WebLink

Home Browse Search

7.3.6 Replicate analyses are performed regularly. 7.3.6.1 A duplicate analysis is performed on at least one sample each day. Method 102 Revision 1 centrations and 4 replicate runs for each concentra- tion. The concentrations must bracket the anticipate) range of sample concentrations. 7.3.3.2 The mean -square error due to lack of fit about the regression line is compared to the total mean -square error of the independent replicates about their individual means. The calibration is accepted if the F -ratio is less than the 95% rejection limit. 7.3.3.3 Any region of concentration that deviates more than b% from the least square line is considered nonlinear. 7.3.3.4 Samples must be analyzed only in the linear range. 7.3.4 Limits of detection must be established. 7.3.4.1 The limit of detection (LOD) is based on three standard deviations (SD) of runs near the LOD (within 10 SD of the LOD, Winefordner and Long, 1983). 7.3.4.2 The LOD should be determined at least on an annual basis. 7.3.4.3 If the benzene calibration factor changes by more than 10%, The instrument must be checked and the LOD redetermined. 7.3.4.4 The presence in a sample of a compound producing in the chart display very large adjacent peaks will often raise the LOD in that sample. 7.3.5 GC column condition parameters should be checked and documented. 7.3.5.1 All GC accessible parameters should be logged when a column is first installed. These parameters should be checked daily and recorded on integrator reports. 7.3.5.2 The efficiency and resolution of th-' column should be checked every thirty days. If the tests show more than a 10% ch&nge the column needs replacement. 7.3.5.3 If the headpressure required to inaintain a specified flow through the column increases by more than 100%, the column needs replacement. 7.3.5.4 If the drift of retention t-imei: of the peaks results in peak misidentification, all instrument parameters need to be checkEi. 7.3.6 Replicate analyses are performed regularly. 7.3.6.1 A duplicate analysis is performed on at least one sample each day.