In situations where large particles can be generated (e.g., abrasive blasting, wood working, and grinding operations) excessive collection of particles up to the millimeter range is likely to occur. There have been some attempts to modify inlets with shields to provide a barrier against the collection of large particles, but these modified inlets have not been demonstrated to provide the same agreement with the inhalable convention as the unmodified ones. Another potential problem with inhalable samplers is the collection of passively sampled particles. Measurements when the sampler airflow is turned off indicate that IOM samplers, which pointed outward from the body and had a large inlet diameter (15-mm), can collect quite significant amounts of dust, with median values of 9 to 32 percent of the mass collected during active sampling [23]. Open-faced cassettes had only 2 to 11 percent of the mass passively collected. These samplers have a larger inlet (37-mm), but point downward, reducing the likelihood of particle settling onto the collection surface. The mechanism of collection is unclear, but the dust may be transported into the inlet by turbulence and deposited by settling or turbulent diffusion. How this passively collected dust modifies the amount collected during active sampling remains under investigation. A comparison of measurements obtained with the 37-mm closed face cassette (4-mm inlet diameter) to the IOM sampler (15-mm inlet diameter) in several workplaces gave similar results when the material on the interior walls of the 37-mm cassette were added to the analyte deposited on the filter [24]. This suggests that the two samplers can have similar inlet efficiencies in spite of differences in inlet size and orientation if the median particle size sampled is not too large. Other studies did not include wall deposits in evaluation of the 37-mm cassette. 3. CLASSIFIER ACCURACY The theory of classifier separation is based on particle aerodynamic diameter, which is defined as the diameter of a 1 g/cm3 density sphere having the same settling velocity as the particle in question. If the particle is markedly non-spherical, the aerodynamic diameter may not be well defined, possibly resulting in sizing errors. For example, fibers and plate-like particles settle slightly differently depending on orientation. Thus, the sampling conventions, based on aerodynamic diameter of particles reaching specified parts of the respiratory system, become somewhat ambiguous for these types of particles. For such non-ideal particles, further testing of classifiers to simulate particle behavior in the respiratory tract may be necessary. For instance, Maynard [25] found that plate-like particles may orient differently in elutriators, impactors, and cyclones. This preferred orientation in a cyclone produced a collection efficiency 15% below that estimated to occur in the respiratory system. Improved understanding of fiber [26] and plate-like particle [25] behavior in the respiratory tract is needed to aid in development of more accurate samplers for these types of particles. Various types of classifiers have been constructed to meet the ACGIH/ISO conventions. For example, respirable samplers have used cyclones [27], impactors [28-30], elutriators [31], and porous foam [32,33] to remove non-respirable particles from the aerosol prior to filter collection. The technology for testing these samplers has improved in recent years through use of a realtime aerodynamic sizing instrument and resulted in quicker and more precise measurements [34,35]; this technique has allowed the accuracy of these samplers to be investigated more carefully [36]. However, a round-robin comparison of 50% cut-point measurements from six
3/15/03
187
NIOSH Manual of Analytical Methods
