Average Pb concentration.
Average Pb concentration. For the reference method, calculate the average Pb concentration for each filter by averaging the concentrations calculated from the three analyses as described in (g)(1) using equation 1 of this section:
(i) Analytical Bias.
(1) For the audit samples, calculate the average Pb concentration for each strip or filter analyzed by the reference method by averaging the concentrations calculated from the three analyses as described in (g)(1) using equation 2 of this section:
(2) Calculate the percent difference (Dq) between the average Pb concentration for each audit sample and the true Pb concentration (Tq) using equation 3 of this section:
(3) If any difference value (Dqi) exceeds ±5 percent, the bias of the reference method analytical procedure is out-of-control. Corrective action must be taken to determine the source of the error(s) (e.g., calibration standard discrepancies, extraction problems, etc.) and the reference method and audit sample determinations must be repeated according to paragraph (g) of this section, or the entire test procedure (starting with paragraph (e) of this section) must be repeated.
(j) Acceptable filter pairs. Disregard all filter pairs for which the Pb concentration, as determined in paragraph (h) of this section by the average of the three reference method determinations, falls outside the range of 30% to 250% of the Pb NAAQS level in µg/m
3 for Pb in both TSP and PM 10. All remaining filter pairs must be subjected to the tests for precision and comparability in paragraphs (k) and (l) of this section. At least five filter pairs must be within the specified concentration range for the tests to be valid.
(k) Test for precision.
(1) Calculate the precision (P) of the analysis (in percent) for each filter and for each method, as the maximum minus the minimum divided by the average of the three concentration values, using equation 4 or equation 5 of this section:
(2) If a direct reading candidate method is tested, the precision is determined from collocated devices using equation 5 above.
(3) If any reference method precision value (PRi) exceeds 15 percent, the precision of the reference method analytical procedure is out-of-control. Corrective action must be taken to determine the source(s) of imprecision, and the reference method determinations must be repeated according to paragraph (g) of this section, or the entire test procedure (starting with paragraph (e) of this section) must be repeated.
(4) If any candidate method precision value (PCi) exceeds 15 percent, the candidate method fails the precision test.
(5) The candidate method passes this test if all precision values (i.e., all PRi's and all PCi's) are less than 15 percent.
(l) Test for comparability.
(1) For each filter or analytical sample pair, calculate all nine possible percent differences (D) between the reference and candidate methods, using all nine possible combinations of the three determinations (A, B, and C) for each method using equation 6 of this section:
(2) If none of the percent differences (D) exceeds ±20 percent, the candidate method passes the test for comparability.
(3) If one or more of the percent differences (D) exceed ±20 percent, the candidate method fails the test for comparability.
(4) The candidate method must pass both the precision test (paragraph (k) of this section) and the comparability test (paragraph (l) of this section) to qualify for designation as an equivalent method.
(m) Method Detection Limit (MDL). Calculate the estimated MDL using the guidance provided in 40 CFR, Part 136 Appendix B. It is essential that all sample processing steps of the analytical method be included in the determination of the method detection limit. Take a minimum of seven blank filters from each lot to be used and calculate the detection limit by processing each through the entire candidate analytical method. Make all computations according to the defined method with the final results in µg/m
3. The MDL of the candidate method must be equal to, or less than 5% of the level of the Pb NAAQS.