allow better sensitivity in samples taken with them, interlaboratory precision would suffer due to the use of multiple sampling devices. At this time, silica sampling should be done with a 1.7 L/min Dorr-Oliver nylon cyclone to meet the ISO/CEN/ACGIH respirable sampling convention within the United States. A future change to the GK2.69 sampler is a possibility.
Both cyclones and filter cassettes should be leak-tested prior to sampling to avoid gross failure in the field [31,32]. The cyclones may be tested using a simple pressure- (or vacuum-) holding test. The filter cassette should also be checked for leakage while attached to the cyclone. Two approaches to testing the cassettes have been used. One approach used a micromanometer to measure pressure drop across a given cassette and compared it to the average pressure drop across well-sealed cassettes [33]. An alternative approach used a particle counter to measure the penetration of submicrometer ambient aerosol through the cassette, with the percentage of penetration serving as an indicator of leakage [31,34]. Evaluation of cassette leakage by several laboratories indicates that significant leakage can occur when cassettes are hand-assembled, when incorrect closing pressures are used on the cassettes and when cassettes are assembled with non-standard thicknesses of filter/backup pad combinations [34,35,36]. For press-fit cyclones, a press should be used to achieve a good seal, while screw-fit devices should have threads in good condition and be firmly screwed together.
5. ANALYTICAL CONSIDERATIONS
The measurement of airborne crystalline silica can be challenging. Sample preparation techniques may include complex procedures for reducing mineral interferences, redepositing the sample onto an analytical filter or using the collection filter for analysis (direct-on-filter measurement). Appropriate calibration of the analytical technique and the standard reference materials used for calibration are critical for accurate analyses. Identification of the analyte, whether quartz, cristobalite or tridymite, can be complicated by the presence of mineral interferences. There are several analytical methods to choose from, each having associated particle size effects. Thus, a high degree of attention is required throughout the analysis process. Discussion of crystalline silica measurement in bulk samples is beyond the scope of this chapter; however, a method for determining crystalline silica in bulk samples by infrared spectrometry is described elsewhere [37].
Prior to any sample preparation step, it is advisable to check the cassette received from the field for adherence of particles to the top or sides of the cassette. In an evaluation of Min-U-Sil 5 samples received in the American Industrial Hygiene Association (AIHA) Proficiency Analytical Testing (PAT) program for Rounds 146-148, it was observed that up to 20% of the total sample was recovered by rinsing the top of the cassette prior to sample preparation [38].
Sample preparation may include procedures to reduce mineral interferences prior to analysis. A phosphoric acid digestion may be used if the presence of amorphous silica is suspected and alternate analytical peaks which are free of interferences do not provide sufficient
sensitivity [39]. The phosphoric acid dissolves the amorphous silica, although some of the smaller crystalline silica particles may be lost [40]. Calcite, magnetite and hematite may be removed with dilute HCl. Kaolinite can be altered to an amorphous state through heating the sample within a specified temperature range. Some mineral interferences also may be dealt with during sample analysis via a thorough knowledge of the characteristic analytical peaks.
