Tobacco Documents Online

CS Electron-Microscope Examination. Specimens are examined in the scanning electron microscope and the million-volt electron microscope, and micrographs are made of random areas until 35 to 50 particles with a 3:1 aspect ratio or larger are located. Visual examination, although obviously much quicker, is liable to large subjective errors in recognizing 3:1 "fibers" and in measurement of particle size. Permanent records are important in case of subsequent need for rechecking or confirmation by others. The total grid• area photographed can be related to the original volume of air samples through the series of dilution and scale factors involved. Typically, one full grid opening cor- responds to about 10 liters of clean ambient air and to about 10 mL of very dusty air or stack samples. In most cases, "average" concentrations of 2 to 10 particles per grid opening are about right. Smaller values lead to prolonged searching to obtain adequate statistics, and higher particle concentrations give difficulties with identification procedures. Statistical Factor and Misidentification Examination of a large number of samples has clearly demonstrated that the concentra- tion (per unit area) of particles follows a Poisson distribution where the variance is about equal to the mean. The corresponding distribution of particles between grid openings is illustrated schematically in figure 11. The mean concentration is 2.5 particles per square. Obviously, differences of nearly an order of magnitude could result by chance if only the left-central or right-central four openings were examined with means of 0.5 and 3.3, respectively. This distribution effect can be minimized by continuing to search and record micrographs until enough data are obtained to be valid at a predetermined confidence level. • • 0 4 3 . • . , 3 4 2 t 3 1 . • • • • • • • ' 4 1 1 , 4 5 2 0 • ~ • ~ . • 4 0 0 i 3 4 ' • 2 r 2 1 4• 5 . ~ , 4 • 2 3 . Figure 11. Schematic illustration of Poisson distribution of particles between grid squares. 397

Random Misidentification. Another source of error that is not so easily circumvented is the misidentification of mineral fragments if only electron diffraction is employed. This can occur by chance even when every care is taken to tilt the particle properly, the pattern is measured accurately, and comparison is made with an appropriate suite of minerals. This error is' particularly serious in view of the growing tendency to equate "amphibole" with "asbestos". The frequency of chance computer matches of diffraction patterns of randomly oriented mineral particles is shown in table 3. The chance of misidentification of a particle as an amphibole when minnesotaite and magnetite are both present ranges from 9 to 18 percent. The misidentification probability is reduced when lattice-symmetry effects are introduced into the calculation. An illustration of the effects of chance misidentification is shown in table 4. It was assumed that 35 particles of a non-amphibole mineral were counted and 7 misidentified as amphibole. If amphiboles could have been present, it is only possible to say that the concentration was less than 103/m3. Because of the larger scale factor for dusty air, or if only a few grid squares were scanned, an even larger apparent concentration would be reported (as has already occurred) [9]. Table 3. Computer matches of randomly generated diffraction patterns.a - - - - Randomly Generated Patterns From - - - - Computer Indexed Solutions Minnesotaite Minnesotaite 100 Actinalite 9 Magnetite 0 Grunerite 11 a Accidental matches are underlined. Actinolite Magnetite 24 12 100 18 0 100 14 13 Table 4. False identification as amphiboles, probability ti0.2. Scale Factora Apparent Concentration Actual Ambient Air l0s ti5 x 10''/m3 0b -<103c Dusty Air 5 x 107 ti2 x 107/m3 0-<10s a(Cubic meter per grid opening)-1. b When none could be present. C Maximum if occurrence possible. More Rapid Electron-Microscope Methods The identification and counting procedures described in previous sections are very time consuming and expensive. Costs per individual air sample are in the range of $300 to $1500, depending an the methods used and the degree of reliability that the analytical laboratory is willing to certify. Deployment of more than 1000 air monitoring stations around the United States might easily be deemed necessary. The financial and manpower resources are just not available for electron-microscope analysis on this scale. 398

es A more simplified method, currently under development, is to group particles observed in the microscope into classes based on their appearance and their x-ray emission spectra, that is, their approximate composition, using automatic image-analysis techniques for measurements and recording. Complete electron-diffraction identification is then carried out on randomly selected particles from each class. This cuts the time required by a factor of five or more, admittedly at some loss in certainty. However, the improved statistics appear to overcome this deficiency. In some cases of particular interest, simpler methods may be developed to achieve reasonably reliable identification. For example, the development of different crystal faces by amosite and grunerite affect their orientation on the electron-microscope grid and the diffraction patterns most commonly observed. This could be used to screen samples for amosite and grunerite with a reasonable degree of accuracy on a quality-control basis. Other specific analysis problems might be solved in similar fashion. However, this method will not work on ambient-air samples, which could contain a large number of minerals or chemical compounds. Finally, automatic image-analysis facilities to process scanning electron-microscope images are coming into use. Effective methods of utilizing these devices are being developed to expedite analysis of air and water samples by eliminating manual data logging and particle counting and measurement. Physical Characteristics of Mineral Fragments Aside from questions of errors in identifying and counting amphibole and other mineral fragments, the very pronounced distinctions between their physical characteristics and those of asbestos must be recognized. In addition to obvious differences in particle size and shape, more subtle features such as crystal structure, face orientation, surface chemistry, associated impurities, lattice imperfections, and concentration could well be important in the biological response to small particles. These factors cannot be over- looked when attempting to generalize from health-effects studies with a particular mineral variety. Scanning-electron-microscope micrographs in figure 12 demonstrate the difference in appearance of crocidolite fibers and several cleavage fragments of hornblende with a 5:1 aspect ratio. The difference in character between the two types of particles is quite apparent, yet there has been a growing tendency to equate them. caocznoaixa aoMeLace Figure 12. Scanning electron micrographs of crocidolite fiber and fragments of hornblende. N ~ 399 W ~` 0 un .+ ~

The cumulative distribution of particle aspect ratios of amosite fibers is shown in figure 13 with a similar plot for comminuted actinolite particles. Differences in the distribution of the aspect ratio of particles are marked. Very few actinolite cleavage fragments have an aspect ratio greater than 10:1, whereas very few amosite particles are less than 10:1. - CUMULATIVE FRACTION, % 100 Figure 13. Plot of measured aspect ratio of amosite fiber and ground actinolite. Studies of more basic structural differences are just beginning in a number of laboratories. Differences in cleavage properties and growth characteristics between amosite and grunerite that are reflected in particle morphology and surfaces were discussed in the Section, "Characteristic Diffraction Patterns." Planar lattice defects, responsible for extensive streaking in certain electron-diffraction patterns, can be seen by high-resolution electron microscopy [11]. When present, these grown-in faults may promote a tendency to split into long narrow fibers. Several years ago, electron-microscope studies revealed that chrysotile fibers are actually hollow tubes [12]. Recently, high-resolution studies of cross-sections of crocidolite have shown the presence of somewhat similar micropores between the fibers [13]. Summary Identification of mineral fragments and determination of their concentration in air and water samples with a satisfactory degree of accuracy can be achieved using electron optical methods if the procedures reviewed in this report are followed. The most important precautions are: 1. Loss of particles or changes in their dimensions during specimen preparation must be avoided. 2. Crystal fragments must be tilted into a zone-axis orientation before selected- area electron-diffraction patterns are recorded. 3. Spot spacings and angles must be measured accurately and compared with computed patterns for any minerals that could occur in the sample. 400

(C 3 4. X-ray emission spectral data from overlapping or closely adjacent particles must be discarded, effects of particle size on line intensities must be recognized, and data must be compared with well characterized reference standards. 5~ Appropriate statistical criteria must be used to interpret the significance of apparent particle concentrations. In our view, a definition of a fiber, incorporating the following points, would resolve most analytical difficulties and remain compatible with all known facts concerning health effects. 1. Aspect ratio greater than 10:1. 2. Parallel edges. It is also necessary to establish a lower limit on particle size in keeping with capabilities of analytical procedures to ensure consistency in count between various laboratories. In the absence of any conclusive evidence for adverse biological effects due to very small particles, the current >5 pm length standard should be maintained. A recent Bureau of Mines Information Circular [14] contains discussion and documenta- tion of many of the foregoing points including: (1) Detailed discussion of definitions relating to asbestos and the importance of eliminating loose terminology from the scientific and popular literature. (2) Discussion of the differing growth and cleavage origins of fibers and fragments as , emphasized in this report. (3) Criticism of the 3 to 1 aspect-ratio criterion for mineral fibers on the basis that fibers from asbestiform minerals always average more than 10 to 1 and often go up to 200 to 1 or eore, whereas the usual ratio for cleavage fragments is less than 10 to 1. References [1] Zoltai, T. and Stout, J. H., "Comments on Asbestiform and Fibrous Mineral Fragments, Relative to Reserve Mining Company Taconite Deposits," Contract report to Minnesota Pollution Control Agency, March 1976. [2] Zoltai, T., "History of Asbestos-Related Mineralogical Terminology," Proceedings, Workshop on Asbestos: Definitions and Measurement Methods, National Bureau of Standards, Gaithersburg, Md., July 18 to 20, 1977. [3] Ross, M., "Minerals Commonly Occurring in the Earth's Crust That Might be Defined as Asbestos by Various Regulatory Agencies," Proceedings, Workshop on Asbestos: Defini- tions and Measurement Methods, National Bureau of Standards, Gaithersburg, Md., July 18 to 20, 1977. [4] Lee, R. J., "Electron Optical Identification of Particulates," Proceedings of Symposium on Electron Microscopy of Microfibers, Pennsylvania State University, August 23 to 25, 1976, in press (1977). [5] Lally, J. S. and Lee, R. J., "Computer Indexing of Electron Diffraction Patterns Including the Effects of Lattice Symmetry," Proceedings of Electron Microscopy Society of America Annual Meeting, Boston, August 22 to 26, 1977. [6] Champness, P. E. , Cliff, G. , and Lorimer, G. W., "The Identification of Asbestos," Journal of Microscopy, 108, Pt. 3, 231-249 (1976). [7] Nord, G. L., Jr., "State-of-the-Art of the Analytical Transmission Electron Micro- scope," Symposium on Electron Microscopy and X-ray Diffraction in Environmental and Occupational Health, in press (1977). _ 401 N ~ 0 N f+ N

[8] Brown, A. L., Jr., Taylor, W. F., and Carter, R. E. , "The Reliability of Measures of Amphibole Fiber Concentration in Water," Environmental Research, 12, 000-000 (1976). [9] Nichelson, W. J., "Analysis of Mesabi Iron Range Samples," Environmental Sciences Laboratory, Mount.Sinai School of Medicine, Contract report to Minnesota Pollution Control Agency, July 1975. [10] Chadfield, E. , Ontario Research Foundation, private communication (1976). [11] Irusteta, M. C. and Whittaker, E. J. W., High-Resolution Electron Microscopy and Diffraction Studies of Fibrous Amphiboles, Acta Cryst. A31, 794 (1975). [12] Yada, K., High Resolution Electron Microscopy of Chrysotile, Acta Cryst. 23, 704 (1967). [13] Franco, N. A., Hutchison, J. L., Jefferson, D. A., and Thomas, J. M., "Structural Imperfection and Morphology of Crocidolite (blue asbestos)," Nature, 266, 7 (1977). [14] Bureau of Mines Information Circular (IC8751), "Selected Silicate Minerals and Their Asbestiform Varities: Mineralogical Definitions and Identification-Characterization," U. S. Dept. of the Interior, 1977. Discussion NOTE: Discussion of this paper was included in the General Discussion at the end of this session. 402