Document pmoVLXeL8yzOpbovvG05amVLB

A. P. GREEN REFRACTORIES CO. RESEARCH CENTER, MEXICO, MISSOURI Laboratory Report I PLAINTIFF'S < EXHIBIT JWRG-738 Raw Materials - Others Loborotory Number: RM-1973 Vermiculite W.R. Grace Co. Dote: April 30, 1S79 Copies: Environmental Control Laboratory WRG MICROSCOPIC EXAMINATION OF FOUR SAMPLES OF W.R. GRACE CO. EXPANDED VERMICULITE Environmental Control-Product Safety (EC-PS) submitted four samples of W.R, Grace Co. expanded vermiculite for the purpose of identifying fibers which had been observed by the U.S. Gypsum laboratory. The EC-PS group was particularly interested whether or not the fibers might contain any asbestifcrm serpentine or amphibole minerals. The four samples were designated as 1) Sulphur Springs-fine, 2} Sulphur Springs-coarse, 3) TroyNo. 2, and 4) Troy-No. 3. These four materials were first examined by binocular stereomicrcscope, then representative examples of the detected fibers were selected for examination under the petrographic microscope, and in the scanning elec tron microscope (SEM). Binocular Microscope Examination Samples of each of the vermiculites were removed from their sample bags by means of a spatula. The materials were passed through a 70 mesh sieve, and the -70 M fraction was examined under the microscope. The fiber search was largely confined to this fraction because the U.S. Gypsum SEM study had shown that the fibers were quite small in cross-section. A trace amount of white fiber bundles was detected in each of the four samples. The appearance and distribution of the fiber bundles is shown in Figure 1. The length cf the bundles ranges from approximately 0.3 to 2.0 mm, but the most common bundle length is 1.0-1.5 mm. An out standing physical characteristic of the bundles is that they may be easily broken into individual acicular fibers of greater than 5:1 (length: width] aspect ratio by applying pressure with a steel needle probe. The amount of such fiber bundles in each sample of vermiculute is certainly less than 12 by volume, and is probably less .than 0.1-0.22. /.r LAB REPORT RM-1973 April 30, 1979 Page 2 Other particles whose aspect ratio is greater than 5:1 are present in each sample, but these cannot be broken into asbestiform fibers. These are probably vermiculite, biotite, non-fibrous amphibole, and other mineral phases. Petrographic Microscope Examination Fiber bundles were removed from the binocular microscope samples with steel needle and/or forceps, and placed upon glass slides for examina tion under the petrographic microscope. All the fiber bundles appear to be the same mineral phase. One of the prismatic refractive indices is approximately 1.63, and the birefringence is in the range of 0.01 0.02. Most important is the non-parallel extinction of the individual fibers. The Zac angle ranges from 15-25. This data indicates that the fibers are not the most common aiestiform mineral, chrysotile. The fiber bundles have the optical properties of the asbestiform amphiboles (i.e. tremolite-actimolite). Other elongate, but not fibrous grains were also examined under the petrographic microscope, and these appear to be similar to the hornblende (i.e. amphibole) group. Scanninc Electron Microscope and Enerov Dispersive Spectrometry (SEM/EDS) Since the petrographic microscope examination showed that the fiber bundles from all four samples were the same, detail SEM/EDS work was only done upon material from the Sulphur Sorincs-fine sample. Figure 2 is a typical fiber bundle, at moderate magnification. At higher magnification, the ends of the bundles have the "splintered wood" form shown in Figures 3 and 4. Then, when one separates the bundle with a steel needle probe and forceps, the individual fibers have the appearance shown in Figure 5. For comparison purposes, a known sample of asbestiform tremolite was examined. The microstructure of this material is shown in Figures 6 and 7. The optical properties of tne known tremolite fibers are similar to those of fibers from the white fiber bundles. The final point of investigation leading to the identification of the fiber bundles was EDS analysis of individual fibers. The first step in this process was to make EDS spectra of known asbestiform tremolite and known massive (.i.e. non-fibrous) tremolite. These spectra are Figures 8 and S. Figures 10 and 11 are EDS spectra of individual fibers in the bundles displayed in Figures 2-5. The similarity in the chemical com positions of the fiber bundles in the vermiculite and the known tremolites adds to the weight of evidence that the fiber bundles are asbestiform tremolite. *./,it >?r:on:) LAB REPORT RM-1973 April 30, 1979 Page 3 Conclusions and Discussion The results of the stereomicroscopic, petrographic microscopic, and the SEM/EDS examination of white fiber bundles (0.3-2.00 mm long), which are present in trace quantities in the four samples of expanded vermiculite, suggests hat the fiber bundles are asbestiform tremolite- actinolite, Ca2(Mg, Fe )g Sig022(* FJg* . The chemical composition of the fiber bundles (by EDS) is in good agree ment with data for tremolite fibers published by Dement (1). The micro scopic fiber bundles are similar in appearance to a macroscopic photograph of fibrous tremolite from Tuscany, Italy presented by Haartz and Lange (2). The occurence of tremolite-actinolite with vermiculite has been reported by Buie and Stewart (3) in the Enoree, So. Carolina vermiculite deposits. These workers suggest that the chemical alteration (hydrothermal ar:' weathering) of ampnibole and biolite produces ve-miculite. The point is that the presence of asbestiform ampnibole probably indicates an area of the vermiculite deposit where the vermiculite-forming reac tions are not complete. One could then expect vermiculite materials mined from different areas of such deposits to contain more or less of the asbestiform ampnibole. This may be the reason why W.R. Grace vermi culite customers have noticed variations in the amounts of asbestiform material in shipments received at different times. In regard to the potential health hazard of the tremolite-actinolite fiber bundles, it must be remembered that the bundles must be physical crushed, abraded, or pulled apart in order to produce respirable fibers, i.e. the asbestiform fibers are not loosely bound as in the reference tremolite material (compare Figures 6 and 7 to Figure 2). Perhaps the mechanical handling of the vermiculite at Sulphur Springs or Troy could destroy some of these fiber bundles, but it would certainly be necessary to take air samples of the work area to establish whether or not the limits of "safe" concentration had been exceeded. R.P. Stevens:jes :ti02 1 1 9 b Figure 1: The distribution of asbestiform tremol ite-actinolite bundles (Sulphur Springs-fines sample) in a field of vermiculite flakes is shown. The particles indicated by (X)s are identifiable asbestiform bundles. 18X. J` 0 2 6 1 1 9 6 Figure 2: The SEM photograph shows the end of a tremolite-actinolite fiber bundle. The broom-like bundle can be mechanical broken into individual asbestiform fibers. 500X. . Figure 3: The tip of a fiber bundle is shown. Some of the individual fibers are less than 1 micron across. 1000X. Fiaure 4: Some of the bundle tips have a definite "splintered wood" structure Aaain, individual fibers of less than 1 micron diameter can be seen 1000X. Figure 5: The separated asbestiform fibers range from 0.3 to 2 microns in diameter, and bear a strong resemblance to known asbestiform tremolite (see Figure 7). 2000X. S01'6 1 2 0 0 Figure 6: This moderate power SEM photograph of known asbestiform tremolite shows much looser "packing" than the fiber bundles in the vermiculite (see Figure 2). 40CX. :S026 1201 Figure 7: At higher power the known tremolite fibers appear to be morphologically similar to the fibers separated from one of the bundles (see Figure 5). 130GX. r. ri cr porjj .etf.-cVMVT&-Afc'-' t.jj t.?1*-*. . , IIH )(if -!>'' V,U- \'i.*-v'VSvk ^i5E^B3l#4^5^6?V7^8 r,:qure 8: Tne EDS spectrum of a known asbestiform tremolite single fiber is shown. Tne major amounts of SiO~, MgO, and CaO and the trace amount of FeO are typical for tremolite. Figure 9: The EDS spectrum of known massive (i.e. non-fibrous) tremolite grain shows the same major elements as the asbestiform variety, but in this case a trace of MnO is present. rj tu Fipure 10: ' The EOS spectram cT a fibers in a fiber bundle (Figures 2-5) are similar to th'^e of known tremolite, but the smaller amount of CaO, and the prese-'.ce of f!a,,C, M,,0,, and more FeO suggests that the fibers are asbes tiform trmelite-actinolite solid solution modified by the presence of some alkali-bearing amphibole. The Ag peaks are from material used to cement the fibers to the SEM stub. Fi gure 11: -------------------------------------------- ------------ ------ ----- This EDS spectrum of a fiber similar to those shown in Figure 5 shows that the unknown xibers are comDositionally quite similar to known asLestiform tremolite (see Figure 8). Ctf.026 1204 REFERENCES 1) John M. Dement, "Asbestiform Minerals in Industrial Talc", NBS Special Pub. 506, Workshop on Asbestos: Definitions and Measurement Methods, p. 318. 2) J.C. Haartz and B.A. Lange, "Selection and Characterization of Fibrous and Non-fibrous Amphiboies for Analytical Methods Development", Op. Cit., p. 298. 3) B.F. Buie and O.F. Stewart, "The Origin of the Vermiculite at Tiaerville, So. Car., (abstract) G.S.A. Bulletin, Vol. 65, No. 12 (1356-57;.