Performance testing of NIOSH Method 5524/ASTM Method D-7049-04, for determination of metalworking fluids.

A performance test of NIOSH Method 5524/ASTM Method D-7049-04 for analysis of metalworking fluids (MWF) was conducted. These methods involve determination of the total and extractable weights of MWF samples; extractions are performed using a ternary blend of toluene:dichloromethane:methanol and a binary blend of methanol:water. Six laboratories participated in this study. A preliminary analysis of 20 blank samples was made to familiarize the laboratories with the procedure(s) and to estimate the methods' limits of detection/quantitation (LODs/LOQs). Synthetically generated samples of a semisynthetic MWF aerosol were then collected on tared polytetrafluoroethylene (PTFE) filters and analyzed according to the methods by all participants. Sample masses deposited (approximately 400-500 micro g) corresponded to amounts expected in an 8-hr shift at the NIOSH recommended exposure levels (REL) of 0.4 mg/m(3) (thoracic) and 0.5 mg/m(3) (total particulate). The generator output was monitored with a calibrated laser particle counter. One laboratory significantly underreported the sampled masses relative to the other five labs. A follow-up study compared only gravimetric results of this laboratory with those of two other labs. In the preliminary analysis of blanks; the average LOQs were 0.094 mg for the total weight analysis and 0.136 mg for the extracted weight analyses. For the six-lab study, the average LOQs were 0.064 mg for the total weight analyses and 0.067 mg for the extracted weight analyses. Using ASTM conventions, h and k statistics were computed to determine the degree of consistency of each laboratory with the others. One laboratory experienced problems with precision but not bias. The precision estimates for the remaining five labs were not different statistically (alpha = 0.005) for either the total or extractable weights. For all six labs, the average fraction extracted was > or =0.94 (CV = 0.025). Pooled estimates of the total coefficients of variation of analysis were 0.13 for the total weight samples and 0.13 for the extracted weight samples. An overall method bias of -5% was determined by comparing the overall mean concentration reported by the participants to that determined by the particle counter. In the three-lab follow-up study, the nonconsistent lab reported results that were unbiased but statistically less precise than the others; the average LOQ was 0.133 mg for the total weight analyses. It is concluded that aerosolized MWF sampled at concentrations corresponding to either of the NIOSH RELs can generally be shipped unrefrigerated, stored refrigerated up to 7 days, and then analyzed quantitatively and precisely for MWF using the NIOSH/ASTM procedures.