Tobacco Documents Online

National Bureau of Standards Special Publication 506. Proceedings of the Wohkshop on Asbestos: Definitions and Measurement Methods held at NBS, Gaithersburg, MD, July 18-20, 1977. (Issued November 1978) INTER-LABORATORY MEASUREMENTS OF AMPHIBOLE AND CHRYSOTILE FIBER CONCENTRATION IN WATER + K. S. Chopra Union Carbide Corporation - Metals Division Niagara Falls, New York 14302 Abstract A3TM Committee E-4 has been experimentally evaluating high magnification microscopic techniques being used for the analysis of fiber contamination in water. This paper will describe the procedures and present status of this technique evaluation. Key Words: Amphibole; ASTM; asbestos; amphibole; chrysotile; fiber; transmission electron microscope; water. Introduction Great interest in the identification, characterization, and concentration determina- tion of mineral fibers in environmental samples has been generated in recent years due to the fibers' potential carcinogenic effect for humans. The variety of sample preparation techniques, instrumentation, identification methods, technical definitions, and levels of analyst experience have often, produced scattered and inconsistent results for related or shared samples. A Task Group was formed under the ASTH E-4 Committee to study the reasons for this inter-laboratory divergence and to establish a recommended standard method for determining fiber concentrations in water. The Task Group' is composed of 17 experts in fine particle analysis from government, industry, and commercial service laboratories in the United States and Canada. Members of the Task Group agreed on the necessity of using a transmission electron microscope (TEM) for the determination of concentratjons of very small fibers, such as asbestos fibers, which have diameters as small as 200 A The (TEM) technique will serve as a reliable method of calibration for other more rapid and less expensive techniques which, hopefully, can be developed. The scanning electron microscope (SEN) was not selected for use because: 1. The SEM lacks the selected area electron diffraction capability for identification of fiber mineral type (e.g., amphibole or chrysotile). 2. The SEN has inferior imaging capability because the image is distorted by sample movement, and the brightness and contrast are less than in the TEM at 20,000X. 3. Some distinctive fiber morphglogies, such as the hollow core of single chrysotile fibrils (200-400 A), cannot be observed by SEN. 0 4. Searching sample areas at magnifications of 10,000 to 25,000 A for fibers is more fatiguing with the SEN. Analyst fatigue contributes significantly to a loss of precision. 5. These observations are meant to define the current limitations of the instruments. N O 377 ~ 0 ~ r+ a on

S61 The Task Group analyzed four Duluth, Minnesota, tap water samples containing amphibole fibers and two samples of filtered water with a chrysotile standard added. The laboratories were supplied with filtered samples on Nuclepore and/or Millipore filters. - Analytical Methods Techniques for the preparation of samples and TEN counting of fibers have been published by Task Group members [1,2,3,4,5,6,7]I. In almost every case, water is filtered onto Millipore or Nuclepore filters. Sections are cut from the filters and placed on TEN grids. The process, whereby the filter is dissolved in a solvent to leave the sample on or in a carbon film on the grid, is a direct transfer method. The filter dissolution step can be done in several different ways and is a key difference between many methods of sample preparation. Most Nuclepore filter preparations are carbon-coated prior to the filter piece dis- solution step so that the resulting grid has the fibers held in the carbon film on the grid. The inclusion of the fibers in the carbon film is made possible by the very flat surface of the Nuclepore filter and is intended to prevent loss of fibers during filter extraction in a Jaffe washer. Millipore filter preparations usually involve the acetone dissolution of filter pieces on a carbon-coated grid in a condensation washer or a Jaffe washer. The condensation washer employs the careful regulation of the level of acetone condensation near a point in a condenser at or just below the position of the grid, so that only acetone vapor is present to dissolve the filter. Fiber identification is often based on the morphology and selected area electron diffraction (SAED) characteristics of the fiber. Many laboratories also rely on energy- dispersive spectrometry (EDS) to classify fibers. by elemental intensity ratio. The observation of morphology at high magnification in the TEN is particularly useful for identifying chrysotile fibrils because of the hollow core or tubular appearance frequently observable. SAED patterns are used to distinguish amphibole and chrysotile fibers from each other and other fibers which have different crystal structures or are amorphous. High-voltage TEN allows the analysis of SAED patterns from fibers too thick for SAED at the normal TEN operating voltages of 60-125 kV. The voltages available on most TEN's do not allow the identification of all mineral fibers, particularly if they are very thin or thick. Considerable controversy exists as to the ~dequacy of SAED for the positive identification of single chrysotile fibrils (200-400 A diameter). Some analysts rely on the observation of the chrysotile magnesium/silicon intensity ratio in the energy-dispersive spectrum instead of a positive SAEg pattern. There are some cases when EDS spectra from different minerals are similar. Conse- quently, an identification based on a combination of morphology, SAEO pattern, and EDS spectrum is considered most reliable, particularly for samples which are collected from previously uncharacterized systems. The members of the Task Group used the combinations TEM-SAEO or TEM-SAEO-ED5 for characterization and identification. Figure 1 shows the inter-laboratory reproducibility for the group analyses and is plotted chronologically. It must be stressed that the inter-laboratory reproducibility is a measure of precision and not accuracy. The Task Group is presently characterizing a sample containing a known chrysotile mass. It is apparent that improvement has occurred in a year and that reproducibility of ±50 percent is possible for fiber concentrations above 70 NFL. The reproducibility at lower concentrations was not this good. The data imply that when all aspects of the analysis are under rigid control, the inter-laboratory reproducibility achievable with the existing TEN technique could be about ±25 percent for relatively clean samples of the type studied herein. Considering the fact that these analyses correspond to the measurement of 50 ppb of amphibole fibers in environmental samples, reproducibility in the range of 25-50 percent is respectable. iFigures in brackets indicate the literature references at the end of this paper. 378

180 160 ~ 0 J y 140 ~s :E W m 120 x_,i y O x ~+ o&6 100 P P 80 ~o z - • W r V J 60 o a ca F MPHIBOLE 40 Z H 4 0 W L a 20 LEGEND 0 -t- ALL LABS 120 AMPHIBOLE CHRYSOTILE m \ CHRYSOTILE ~ 0 80 0° 0 m d ¢ b ~ s 0 ~ ~ 40 0 m a J ~ 20 ~ L 0 x 11/18/75 5/3/76 8/23/76 12/6/76 6/23/76 12/6/76 DATE Figure 1. Plot of chronological inter-laboratory reproducibility for the group analysis. 379

C~% Summary These methods offer a feasible means of measuring relatively low levels of fiber contamination in environmental water samples. Other bulk-type methods lack the needed sensitivity and selectiYity. The transmission electron microscope is the best basic instrument for performing analysis, particularly when equipped with selected area electron diffraction and energy-dispersive spectroscopy capabilities. The mean fiber concentration by different groups agree within a factor of two. The inter-laboratory reproducibility of 50 percent can be expected in relatively clean water samples unless the concentration is low. In samples with high concentration of interfering solids, the precision will not be as good. Inter-laboratory reproducibility of 25 percent is as good as the method can provide. When applied on a broad scale, there are variable (0-84%) and significant (mean = 30%) losses associated with the condensation washing of samples containing amphibole. The losses are low (mean = 14%) and less variable when using condensation washing to prepare samples containing chrysotile. References [1] Cook, P. M., Rubin, I. B, Maggiore, C. J., and Nicholson, W. J., X-ray diffraction and electron beam analysis of asbestisform minerals in Lake Superior waters. Proc. Int. Conf. on Environmental Sensing and Assessment, 34(2), 1-9 (1976). [2] Millette, J. R. and McFarren, E. F., EDS of waterborne asbestos fibers in TEM, SEM, STEM scanning electron microscopy/1976 (Part III) 451-460 (1976). [3] Beaman, D. R. and File, 0. M., Quantitative determination of asbestos fiber concentra- tions, Anal. Chem. 48 1, 101-110 (1976). [4] Anderson, C. H. and Long, J. M., Preliminary interim procedure for fibrous asbestos, U. S. Environmental Protection Agency, Athens, Ga. (1977). [5] Chatf.ield, E. J. and Glass, R., Improved methodology for determination of asbestos as a water pollutant, Report ORF, Mississauga, Ontario, Canada, Center for Inland Waters (1976). [6] Dement, J., Zumwalde, R., and Wallingford, K. M., Asbestos fiber exposure in Hard Rock Gold Mine, Annals of New York Academy of Sciences, 271, 341-352 (1976). [7] FDA Symposium on Electron Microscopy of Microfibers, August 19, 1976, Penn State University, State College, Pa. (in press). [8] Lee, R. J., Debray, D. R., and Szirmae, A., Sample preparation and measurement of losses in electron optical analyses of particulates in air and water samples, Thirty- fourth Annual Emsa Proceedings, Miami Beach, F1., Aug. 1976, pp. 556-557. This paper will be published in its entirety in the July 1978 issue of Testing and Evalua- tion published by ASTM. Discussion NOTE: Discussion of this paper was included in the General Discussion at the end of this session. 380