dispersive x-ray analysis. Although TEM analysis is potentially very powerful and accurate, the process of sample collection and preparation and details involved in sample analysis can degrade the quantitative accuracy of the technique. Several more specialized techniques, such as electron energy loss spectroscopy and secondary ion mass spectrometry, have been used for analyzing particles and can also be applied to fibers [69]. Airborne fiber samples for TEM analysis are typically collected onto a filter, usually a polycarbonate membrane or MCE membrane filter. For the latter filter type, the filter is chemically collapsed to form a smooth upper surface on which collected fibers are trapped. Sometimes the surface is etched using a low temperature asher to expose the fibers collected on or near the surface of the original filter. The filter is coated with a carbon film that entraps fibers exposed on the filter surface and the filter material is then dissolved away. The carbon film is transferred to a TEM grid (usually 3-mm diameter) and the sample can be placed in the TEM for analysis. For Method 7402, the surface is not ashed because some fibers, e.g., cellulose, may be removed and give an inaccurate total fiber count [58]. Ashing can thus affect the measurement of the asbestos fiber fraction. The above approach to preparing MCE filters for TEM analysis is called the direct-transfer approach, since fibers are transferred to the carbon film with minimum disturbance to the way they were collected. An alternative technique is to dissolve the entire filter in liquid, ultrasonicate the suspension to disperse the particles, and deposit an aliquot of the particle suspension onto a polycarbonate filter for final transfer to the carbon film. This is called the indirect transfer technique. With the indirect technique, the optimum particle loading of the TEM sample can be obtained and soluble particles can be removed from the sample. However, the suspension process can change the apparent size distribution of the particles and fibers by breaking apart agglomerates or even breaking apart asbestos fibers into smaller fibers or fibrils [70]. The breakup problem can be especially severe for chrysotile, causing a large increase in fiber count. Quality assurance is especially important with TEM analysis of fibers. The NVLAP program provides quality assurance accreditation for laboratories performing TEM analysis using the Environmental Protection Agency’s Asbestos Hazard Emergency Response Act (AHERA) method. Note that data provided under the AHERA method, because of significant differences in counting rules, the types of structures counted as asbestos, and the size range of fibers, cannot be directly compared with counts by Methods 7400 or 7402. The process of sample collection and preparation is a complex one that can introduce biases into the final measurement. Since only small portions of the filter are measured during TEM analysis, sampled fibers that deposit non-uniformly onto the filter due to inertial, gravitational, and electrostatic effects will be measured inaccurately [71]. Fibers that penetrate the filter surface and are not transferred to the carbon film will be lost. If the filter is incompletely dissolved away from the carbon film, the sample will be difficult to analyze. Many of the sources of bias and variability noted in sampling and counting by PCM also apply to TEM analysis. Fiber counting in a TEM can also introduce biases and variability in the final result. There is a tendency to use the high magnification of the TEM to look for the smallest fibers, while ignoring some of the larger ones. Even so, fibers shorter than 0.5 :m tend to
3/15/03
155
NIOSH Manual of Analytical Methods
