Page:NIOSH Manual of Analytical Methods - 5042.pdf/6

 BENZENE-SOLUBLE & TOTAL PARTICULATE (ASPHALT FUME): METHOD 5042, Issue 1, dated 15 January 1998 - Page 6 of 7

2 = aliquot factor

EVALUATION OF METHOD: Asphalt fume collected during a previous NIOSH investigation [5] was spiked on tared PTFE filters, allowed to dry at least overnight, and extracted using benzene. The results are summarized in the table below. Total Particulates Spiking level (mg)*

Recovery (%)

Benzene-Soluble Fraction

Sr

Recovery (%)

Sr

1.85

102

5.97

97.9

0.738

1.17

103

3.98

98.8

2.02

0.62

94.0

5.85

94.8

1.85

0.23

91.6

3.50

96.9

6.10

0.12

82.1

3.91

80.9

9.54

0.058

110

16.4

92.1

13.5

0.025
 * Six replicates per level

105

11.4

73.1

17.4

The pooled relative standard deviation ( r) for the total particulates was 4.8% at loadings greater than or equal to 0.10 mg per sample. For the benzene-soluble fraction, the pooled relative standard deviation was 6.1% at loadings greater than or equal to 0.21 mg per sample. The accuracy criterion is based on determining the range of analyte loadings and the analyte loading on the sample media that will give at least 95% confidence of obtaining a measurement of the analyte that is within 25% of the true value 95% of the time. Since no independent method for determining the total particulate concentration is available, no estimate of the bias for the total particulate data was made; therefore, the maximum allowable bias was calculated at which the accuracy criterion could still be met. Based on the spiking data, if the total particulate loading was greater than or equal to 0.10 mg per filter, the measurement determination will be within 25% of the true value95% of the time if the true bias is less than 10.0%. The bias for the benzene-soluble fraction was negative (see the data above), and since the bias for the benzenesoluble fraction varied little, the bias was pooled over the spiking range of 1.85 to 0.20 mg per filter. It was determined that the 25% accuracy criterion was met if the benzene-soluble fraction was greater than or equal to 0.20 mg per filter. The limit of detection (LOD) and the limit of quantitation (LOQ) were determined using field blanks [6]. The LOD is equal to three times the standard deviation of the field blank weight differences (post-sampling weight - tare weight), and the LOQ is equal to ten times the standard deviation of the field blank weight differences. Field sample values should be compared to the LOD and LOQ values only after the field samples have been blank corrected. The standard deviations of the field blank weights were 0.013 mg per sample for total particulates and 0.014 mg per sample for the benzene-soluble fraction. Therefore, the LOD and LOQ for total particulates were 0.04 and 0.13 mg per sample, respectively. The LOD and LOQ for the benzene-soluble fraction were 0.04 and 0.14 mg per sample, respectively. These LOD and LOQ values should only be compared to blank corrected field sample data. A user check of the method was performed in which tared PTFE filters were spiked with 1.08, 0.392, or 0.216 mg of pyrene per filter and then analyzed by an independent chemist [6]. A mean total particulate recovery of 103% (Sr = 5.85%) was obtained, and the mean benzene-soluble fraction recovery was 109% r(S= 9.91%). Correlation of benzene-soluble mass with total particulate was linear, with2R= 0.994, and the mean

NIOSH Manual of Analytical Methods (NMAM), Fourth Edition