Impactors have been used as samplers and, especially with cascade impactors, the deposits on the impaction stages are measured. Particle bounce from the collection substrates on the impaction plates can be severe, especially for solid particles impacting onto a smooth metal plate. Several modifications to the collection substrate are available to improve collection efficiency of each stage. These modifications should be compatible with the analytical method. Oil can be placed on the collection substrate that wicks up over collected particles and continually provides an oiled surface. A filter or sintered metal can be used to provide a reservoir for this oil. For gravimetric analysis, this oil must have a low vapor pressure and not migrate off the collection substrate. Alternatively, grease can be used, but after the surface is coated with collected particles, additional particles are more likely to bounce. Filters have also been used as substrates and provide a convenient substrate that is somewhat better than a smooth metal surface. Selection and use of an impactor is a complex issue and has been described in reviews [88, 89]. Accurate analysis of cascade impactor data can also be difficult and simple regression analysis of the data may not provide the best answer [89-91]. 9. SAMPLER FIELD COMPARISONS Direct field comparisons of various samplers are frequently reported in the literature. Because of the typical high variability of aerosol concentrations and size distributions in workplaces, it is difficult to use these situations for accurate assessment of sampler performance. However, field studies are important to verify the overall performance of a sampler and to indicate specific sampler issues. The problems with samplers as discussed above can be highlighted with some examples observed in field studies. a. Sampler Bias Affected By Internal Deposits A study of wood dust sampling comparing collocated free-standing samplers indicated that an MSA cassette (having an aluminum cartridge crimped onto the filter) used as a sampler gave two times better precision and collected 2.6 to 3.5 times more dust than the standard 37-mm closed-face cassette [92]. Both these samplers have the same size and shape of inlet. The same study showed that the IOM sampler collected 1.3 times more dust than the MSA cassette, indicating that the particle size, inlet shape and inlet orientation are important factors in inhalable sampling. Among a number of inhalable samplers in current use around the world, the IOM sampler appears to agree the best with the inhalable dust sampling convention [9, 93]. Several studies have shown that the IOM sampler collects anywhere from slightly more to 3.5 times more dust than the closed-face cassette [94-98]. However, Demange et al. showed that for several work sites with relatively small MMAD particles (about 15 :m), the 37-mm cassette results agreed well with the IOM results if the deposits on the internal surfaces of the cassette were added to the filter analyte [23]. Measurements with the 37-mm cassette are not expected to agree as well with the IOM when the particle sizes are much larger because of differences in aspiration efficiency. However, by including all aspirated material, i.e., all material entering the 37-mm cassette inlet, in the analysis, it appears that agreement with the inhalable convention is improved. b. Sampler Precision Affected By Internal Deposits Internal wall deposition in a sampler can also contribute to variability of results. In a field study of lead dust, it was found that the results from a closed-face 37-mm cassette gave a coefficient
3/15/03
194
NIOSH Manual of Analytical Methods
