Discovering the True Morphology of Amphibole Minerals: Complementary TEM and FESEM Characterization of Particles in Mixed Mineral Dust

Transcription

1 A. Méndez-Vilas and J. Díaz (Eds.) FORMATEX 2007 Discovering the True Morphology of Amphibole Minerals: Complementary TEM and FESEM Characterization of Particles in Mixed Mineral Dust K. E. Harris, K. L. Bunker *, B. R. Strohmeier, R. Hoch and R. J. Lee RJ Lee Group, Inc., 350 Hochberg Road, Monroeville, PA, 15146, USA This study involved the development and application of a protocol using transmission electron microscopy (TEM), energy dispersive X-ray spectroscopy (EDS), selected area electron diffraction (SAED), and field emission scanning electron microscopy (FESEM) techniques for the particle-byparticle characterization (i.e., chemistry, crystallography, and morphology) of mixed mineral dust. This protocol was developed to characterize amphibole particles that occur as accessory minerals around the former vermiculite mine located near Libby, Montana. Media and public attention has been focused on Libby since the fall of 1999 due to health concerns over potential amphibole asbestos exposure occurring in and around the community. The mineralogy in the Libby area is complex. There is ongoing research and debate on the composition and morphology of the various types of asbestos and non-asbestos amphibole particles present in the mixed mineral dust and their relative quantities and potential health risks. Therefore, the distinction among asbestos and non-asbestos amphibole particles is important from both a scientific and regulatory standpoint. In this study, the Yamate Level III TEM method for asbestos analysis of air samples was supplemented by FESEM imaging. The FESEM procedure involved relocation of all amphibole particles analyzed by TEM/EDS/SAED and the collection of secondary electron images of the full structure, both structure ends, and the structure surface. Direct comparison of the TEM and FESEM images for a specific particle reveals morphological features that cannot be observed in a TEM image. Features key to distinguishing between asbestos and non-asbestos amphibole particles, such as overall particle shape and surface roughness, are readily apparent in the FESEM images. Particle-by-particle examination of multiple air samples generated large volumes of TEM and FESEM images, EDS spectra, and SAED patterns. To facilitate data review, more than 67,000 TEM and FESEM images, spectra, and diffraction patterns were imported into a database. The database allows simultaneous viewing of multiple TEM and/or FESEM images, EDS spectra, and SAED patterns for a given particle. More than 3,700 structures were classified by various categories, such as particle shape, end and side geometries, and surface texture. The organization and viewing capacity of the database are useful for sharing technical results with other experts and training new analysts. This study demonstrated that FESEM imaging and digital database utilization are essential elements for accurate characterization of complex mixed mineral dusts, particularly when dealing with large numbers of samples. Keywords amphibole; asbestos; digital database; FESEM; field emission scanning electron microscopy; mixed mineral dust; SEM; TEM; transmission electron microscopy 1. Introduction "Asbestos" is a commercial term applied to a group of naturally occurring silicate minerals that grow in a specific fibrous form and exhibit unique characteristics such as: flexibility, high tensile strength, and resistance to heat and chemical degradation. Six minerals are specifically regulated as asbestos by the U.S. Federal government [1]: chrysotile (fibrous serpentine) and five varieties of amphibole fibers: crocidolite (riebeckite asbestos), amosite (cummingtonite-grunerite asbestos), anthophyllite asbestos, tremolite asbestos, and actinolite asbestos. In addition to the six regulated asbestos minerals, 388 minerals (including 92 silicate and aluminosilicate species) are known to occur, at least occasionally, in fibrous * Corresponding author: klbunker@rjlg.com 643

2 FORMATEX 2007 A. Méndez-Vilas and J. Díaz (Eds.) form, some of which are asbestiform [2]. The term asbestiform is used to describe the unusual crystallization habit (the actual shape assumed by a crystal or aggregate of crystals) of minerals that form as thin, hair-like fibers such as that which occurs with the six asbestos minerals. Asbestiform describes a special type of fibrosity not all fibrous minerals are asbestiform. The vast majority of amphibole minerals occur as ordinary rocks. Amphibole fragments separated or cleaved from these rocks during weathering, crushing, or grinding can occur in a variety of shapes ranging from blocky to prismatic to acicular. Long, thin "cleavage fragments," while rare, may resemble asbestos fibers. In 1992, however, the Occupational Safety and Health Administration (OSHA) made the decision that cleavage fragments are not asbestos and should not be regulated as asbestos based on epidemiology studies that were either inconclusive or showed no adverse health effects from exposure to non-asbestiform mineral particles [3,4]. Various procedures have been developed for the analysis of asbestos minerals over the last 100 years. The historical development of these methods focused on the analysis of commercial-grade chrysotile asbestos found in the workplace and in consumer products, not the occurrence of amphibole asbestos in natural mixed mineral environments. Beginning with polarized light microscopy (PLM) and continuing through transmission electron microscopy (TEM), the method development has progressed from mineral identification in hand samples to mineral identification of respirable airborne fibers. In the last 30 years, TEM has evolved as the primary tool for airborne asbestos fiber analysis related to occupational exposures in the United States. In a typical TEM analysis, the definition of a fiber is often based on the general shape of the particle, including aspect ratio (length:width) and attributes such as parallel sides or substantially parallel sides [5-10]. Recently, there have been studies in which any particle more than three times longer than it is wide is counted as asbestiform [11]. The criteria are very different from the physical characteristics a mineralogist would use to distinguish asbestos, namely that the particle must have a unique fibrous form. A technique that utilizes an asbestos fiber definition specifying a minimum aspect ratio of 3:1 for particles longer than 5 µm is not valid for the analysis of mixed mineral dusts, however, simply because in most typical natural environments there are too many non-asbestos particles that would fit the aspect ratio definition. Although TEM provides a projection of the general overall shape of mineral particles, the true shape and dimensions of a particular particle may be significantly different from what is observed in the projected TEM image. This limitation can result in the incorrect identification of non-regulated, non-asbestiform particles as asbestos fibers. In this study, field emission scanning electron microscopy (FESEM) was utilized for morphological characterization of mixed mineral dusts as an extension of the TEM analysis. Augmenting the TEM method with high resolution FESEM provides the opportunity to examine the overall particle shape, surface topography, and the side and termination (particle end) geometries. Although morphological characterization studies of asbestos using SEM have previously been reported [12-16], this is the first time a detailed and comprehensive FESEM morphological characterization method has been described for supplementing TEM analyses. This combined TEM/FESEM protocol was developed as part of an on-going effort to characterize amphibole particles that occur as accessory minerals in the former vermiculite mine located near Libby, Montana. Media and public attention has been focused on the small town of Libby since the fall of 1999 due to health concerns over potential amphibole asbestos exposure in and around the community. The mineralogy in the Libby area is complex and there is ongoing debate concerning the composition and morphology of the various types of asbestos and non-asbestos amphiboles that are present in this area and their relative quantities and potential health risks [16-21]. Since government agencies currently do not regulate all asbestiform minerals or cleavage fragments, it is crucial to understand the precise mineralogy of any potential asbestos-containing material. 2. Experimental Methods During the period of , the U.S. Environmental Protection Agency (EPA) collected thousands of bulk soil, ambient air, and other related samples in the vicinity of Libby. As part of cost recovery litigation filed by EPA against W.R. Grace Co., the owner of the Libby vermiculite mine from 1960 to its closure in 1990, splits of about two hundred air samples were requested from the EPA by W.R. Grace. 644

3 A. Méndez-Vilas and J. Díaz (Eds.) FORMATEX 2007 These air samples were collected from residential, outdoor, and former vermiculite screening plant locations in the Libby area. The EPA provided the samples and related analysis results to RJ Lee Group, Inc. as part of their litigation data production. A description of the EPA sample collection procedure is available online at: Twenty air samples (11 residential, one outdoors, and 8 screening plant), which the EPA contract laboratories had identified as having the presence of Libby Amphiboles, were selected for analysis in this study. In general, the air samples were collected on 0.4 µm mixed cellulose ester (MCE) filters at flow rates ranging from 2-10 litres per minute at stationary sampling locations selected to represent typical ambient air. The submitted filters were exposed to acetone vapour for 20 seconds and placed in a plasma asher for approximately one minute. Samples were carbon coated and the MCE filter media were dissolved in acetone following the procedure of Burdette and Rood [22], leaving the airborne particulate trapped in a carbon film supported by a 400 mesh locater TEM grid. Samples were examined with either JEOL Model JEM 1200EX or JEM 2000FX TEMs operating at an accelerating voltage of 120 kv. All particles in 40 grid openings having a 3:1 aspect ratio or larger were analyzed following the Yamate Level III procedure [5]. The method requirement that particles have parallel or substantially parallel sides was not enforced for the purpose of this analysis. The Yamate method requires the collection of energy dispersive X-ray spectroscopy (EDS) spectra and selected area electron diffraction (SAED) patterns to examine the particle chemical composition and crystal structure, respectively. A detailed discussion of the EDS/SAED results is beyond the scope of this article, which is focussed on the TEM/FESEM imaging comparison, and will be presented in another publication. All particle mineral assignments mentioned in this article were determined from the EDS/SAED results using the nomenclature developed by Leake, et al. [23]. During the TEM analyses, a sketch was made showing the locations of the particles of interest on the TEM grid, as per the National Voluntary Laboratory Accreditation Program (NVLAP) 1876b method for verified counting [24]. Prior to FESEM analysis, the TEM grids were platinum coated for 30 seconds at a deposition current of 20 ma using a K950X Turbo Evaporator with K350 Sputter Coating/Glow Discharge Attachment Magnetron target assembly. TEM grids were mounted in a JEOL EM-Q2 Quick Change Retainer, which was attached to a modified aluminium SEM pin mount stub with aluminium screws. Secondary electron (SE) images were obtained with an FEI Sirion 400 FESEM operated in the ultrahigh resolution mode with a Through-the-Lens Detector (TLD) and an accelerating voltage of 3 kv, a beam current of 200 pa, and a 4.0 mm working distance. Particles previously analyzed by TEM were relocated in the FESEM with the aid of TEM field images and TEM analyst sketches of the appropriate grid openings. 3. Results and Discussion In this study, more than 3,700 individual amphibole particles were examined by TEM followed by the complementary relocation analysis by FESEM. The FESEM protocol required the collection of SE images of the full structure, both structure ends, and the structure surface, as well as three-dimensional FESEM stereo images of the full structure and structure surface. The colorized stereo images cannot be properly viewed in black and white print such as this book, but can be viewed on the RJ Lee Group, Inc. website at The overall morphologic shape of the mineral particles relocated by FESEM can be described using six primary classifications: fiber, acicular, prismatic, bladed, columnar, and bundle. These classification terms were originally developed to differentiate between common mineral rock fragments and their asbestiform varieties using optical microscopy [25]. FESEM, however, provides an opportunity to observe the morphological and surface characteristics of fine dust particles that cannot be readily observed by optical microscopy or TEM. Figure 1 shows examples of TEM images and corresponding FESEM images for five of the particle classifications, which will be described in more detail below. a. Fibers (see Fig. 1a): Fibers are thin particles (width 0.5 µm) with parallel sides, smooth surfaces, and no discernable crystal faces (i.e., an apparently round cross-section). Fibers have very high (20:1 to 100:1, or higher) aspect ratios and often display curvature (i.e., bending and 645

4 FORMATEX 2007 A. Méndez-Vilas and J. Díaz (Eds.) flexability, see Fig. 1a where the fiber end wraps around another particle). The fiber shown in Figure 1a does have an asbestiform habit, but the EDS and SAED identified the particle as richterite. Therefore, although the fiber is asbestiform, it is not a regulated asbestos fiber. b. Acicular (see Fig. 1b): Acicular particles are thin particles (width 0.5 µm) with generally moderate (10:1 to 20:1) to high (> 20:1) aspect ratios and developing crystal faces. The sides of acicular particles are generally parallel, although non-parallel sides can be observed. The ends of acicular particles are often needle-like or tapered, but perpendicular ends can also be observed. Acicular can be thought of as an intermediate state between a fiber (i.e., no crystal faces) and prismatic particles (i.e., well developed faces, see below). The surface characteristics of acicular particles are not a critical or defining property and therefore can include smooth, textured, mottled, or rough features. c. Prismatic (see Fig. 1c) and Bladed (see Fig. 1d): Prismatic and bladed particles are similar classifications with both types having well-developed crystal faces and generally low (< 10:1) to moderate (10:1 to 20:1) aspect ratios. Prismatic particles have one elongated dimension and two shorter, approximately equal dimensions (i.e., similar particle width and thickness), while bladed particles have one longer dimension and two unequal, shorter dimensions. In contrast to prismatic particles, the widths of bladed structures are typically 4 to 5 times larger than the thickness. The sides of prismatic and bladed particles are typically parallel, but non-parallel sides can be observed. Prismatic particles often have a well-defined corner or edge and a crystalline termination. Bladed particles typically have a wedge-shaped cross-section. The surface texture of prismatic and bladed perticles can between smooth and rough. d. Bundles (see Fig. 1e): Bundles are clusters of individual readily separable particles (typically fibers, but also acicular or prismatic particles). Bundles often have splayed ends and a lack of cohesion between particles in the group, as indicated by the relative displacement of particles with respect to one another along the length of the bundle (i.e., twisting and spreading of substructures). Long, thin, individual asbestos fibers can readily be seen in the FESEM image of the end of the chrysotile bundle in Fig. 1e. e. Columnar: Similar to bundles, columnar particles consist of an approximately parallel, column-like arrangement of substructures (fiber, acicular, prismatic, or bladed) running the length of the main particle. The individual substructures in columnar particles are obvious but do not appear to be readily separable. f. Irregular: Large ( 2 µm) "blocky" particles with non-parallel, irregular sides and irregualr ends that do not fit any of the six primary particle catagories discussed above are separately classified as "irregular." The visibility (i.e., field-of-view) and image resolution of FESEM have been shown to be more than adequate for characterizing long, thin (i.e., 10 nm width) mineral particles. Direct comparison of the TEM and FESEM images shown in Figure 1 demonstrate the unique morphological features that can be observed with FESEM, but cannot be observed in a standard TEM image. All of the amphibole particles shown in Fig. 1 meet the counting criteria of a 3:1 aspect ratio; however, FESEM indicated that only the particles seen in Figs. 1a and 1e are actually asbestiform structures. Features historically used to differentiate between asbestos and non-asbestos amphibole particles, such as overall particle shape, end and side geometries, and surface roughness, are readily apparent in the FESEM images. In addition, FESEM has the added advantage that the entire structure can be imaged, unlike TEM, which can only image those portions of particles that lie within a given TEM sample grid square (see Figs. 1b and 1d). Particles that lie completely on a sample grid bar are "invisible" to TEM, but can readily be imaged with FESEM. 646

5 A. Méndez-Vilas and J. Díaz (Eds.) FORMATEX 2007 Figure 1: TEM image of overall particle (left), FESEM SE image of overall particle (center), and FESEM SE image of particle end (right) of a) Richterite fiber with length=31.0 µm, width=0.3 µm (103:1 aspect ratio), b) Acicular richterite particle with length=6.0 µm, width=0.25 µm (24:1 aspect ratio), c) Pristmatic richterite particle with length=16.2 µm, width=1.0 µm (16:1 aspect ratio), d) Bladed richterite particle with length=6.25 µm, width=0.2 µm (31:1 aspect ratio), e) Chrysotile asbestos bundle with length=13.0 µm, width=1.0 µm (13:1 aspect ratio). The white box in the SE image of the overall particle shows the location of the higher magnification image of the particle end. 647

6 FORMATEX 2007 A. Méndez-Vilas and J. Díaz (Eds.) Figure 2: Screen image of one page of the Filemaker TM Pro 8 database for a richterite particle showing the FESEM secondary electron image (left) and TEM image (right). The screen image shows the overall structure of the database including the classification system. 648

7 A. Méndez-Vilas and J. Díaz (Eds.) FORMATEX Utilization of the Database The particle-by-particle TEM/FESEM analysis of multiple samples generated large volumes of images, EDS spectra, and SAED patterns. To facilitate data review, all TEM/FESEM images, spectra, and SAED patterns were imported into a FileMaker TM Pro 8 database. For this study, more than 67,000 TEM and FESEM images, spectra, and SAED patterns were imported into the database. TEM images were rotated using Adobe Photoshop 9.0 software prior to importation into the database so that the TEM and FE- SEM images for a given particle had the same relative orientation for a direct side-by-side comparison. The database allows simultaneous viewing of multiple TEM/FESEM images, spectra, and SAED patterns for a given particle. The database format also allows querying and summarizing classifications from a single sample or a group of samples. The organization and viewing capacity of the database are useful for sharing technical results with other experts and for training new analysts. Particle morphologies were first classified, using the terminology described above, by the overall shape observed in the FESEM images. Additional morphological classifications were made for features such as the particle surface texture, as well as the side and end geometries. The "particle association," which refers to the relationship of the particle of interest with respect to other components of the mineral dust, was documented. The orientation of the particle with respect to the sample support grid was also noted because morphological descriptions of the particle features are often orientation dependent. Classifications were assigned to the particles identified in the air samples as well as to reference minerals. This allowed a comparison of the physical characteristics of asbestiform versus non-asbestiform particles. Figure 2 shows an example of a typical database page comparing the TEM and FESEM full structure images obtained for a prismatic, euhedral particle with a summary of detailed morphological classification information (EDS/SAED results indicated this particle was a richterite mineral). Most of Particle Classifications Percentage Fiber 6.0% Acicular 16.0% Prismatic 61.5% Bladed 8.0% Bundles 1.0% Columnar 6.0% Irregular 1.5% Table 1: Classification summary of a subset of 1205 particles from Libby, MT. the detailed morphological information, which indicated that the particle shown in Fig. 2 was not asbestiform, could not be obtained from the TEM image alone. In addition, the particle shown in Fig. 2 was supported by other adjacent mineral debris and is projecting outward from the plane of the page, which was not evident in the TEM image. A subset of 1205 particles collected in the vicinity of the mine, in and out of the town of Libby, and from a processing plan in Libby were analyzed and classified according to the procedure and definitions described above (see Table 1). The majority of the particles were determined to be non-asbestiform particles classified as prismatic, bladed, and acicular particles. In general, these particles have developed crystal faces, rough surfaces, and low aspect ratios. The population of particles did contain a small number of fibers and bundles that were determined to be asbestiform particles. The asbestiform particles have high aspect ratios with smooth surfaces, parallel sides, and no discernable crystal faces. The attributes of asbestiform and non-asbestiform identified using the FESEM classification method described herein are in agreement with the findings of several published research papers [14,17,25]. 4. Summary This study has demonstrated that FESEM is a valuable complementary tool to TEM for characterizing the morphology and surface characteristics of amphibole particles suspected to be asbestos. The accurate analytical distinction between asbestos and non-asbestos minerals is critical in regulatory settings. FESEM results indicate that amphibole cleavage fragments have dimensions and morphological features very different than true asbestos. These results indicate that the concept suggested by some researchers that a significant population of fiber-like cleavage fragments with the same dimensions as asbestos fibers exists in typical mineral dusts, is a myth. In this study, the added benefits of FESEM helped the 649

8 FORMATEX 2007 A. Méndez-Vilas and J. Díaz (Eds.) geological and microscopy communities better understand the morphology of the asbestiform and nonasbestiform amphibole minerals found around the former vermiculite mine in Libby, MT. Asbestos is a unique carcinogen in that dose is determined by the number of fibers, not the mass. The FESEM microscope is an essential element in evaluating potential hazards of mixed mineral dusts. Advances in SEM technology have made FESEM a common and reliable tool in analytical laboratories and has allowed significant advances in the characterization of mineral fibers. This study demonstrates the value of morphological characterization in differentiating between asbestos and non-asbestos particles. Acknowledgements. Financial support by the W. R. Grace Co. is gratefully acknowledged. References [1] "40 CFR Part 763," U.S. Code of Federal Regulations, 28, 2003, p [2] H. C. W. Skinner, M. Ross, and C. Frondel, "Asbestos and Other Fibrous Minerals: Mineralogy, Crystal Chemistry, and Health Effects," Oxford University Press, New York, NY, p. 94, [3] U.S. Code of Federal Regulations, 57, June 8, 1992, pp [4] E. B. Ilgren, Indoor Built Environ., 13, 343 (2004). [5] G. Yamate, S. C. Agarwal, and R. D. Gibbons, "Methodology for the Measurement of Airborne Asbestos by Electron Microscopy," EPA Contract No , IIT Research Institute, Chicago, IL, [6] "Interim Transmission Electron Microscopy Analytical Methods Mandatory and Nonmandatory And Mandatory Section to Determine Completion of Response," Federal Register, 52, pp , [7] A. V. Samudra, C. F. Harwood, and J. D. Stockham, "Electron Microscope Measurement of Airborne Asbestos Concentrations: A Provisional Methodology Manual," EPA Report 600/ , U.S. Environmental Protection Agency, Washington, D.C., [8] "Test Method for Microvacuum Sampling and Indirect Analysis of Dust by Transmission Electron Microscopy for Asbestos Structure Number Concentration," ASTM D5755, American Society for Testing and Materials, West Conshohoken, PA, [9] "Test Method for Microvacuum Sampling and Indirect Analysis of Dust by Transmission Electron Microscopy for Asbestos Mass Concentration," ASTM D5756, American Society for Testing and Materials, West Conshohoken, PA, [10] "Asbestos by TEM," NIOSH Method 7402, NIOSH Manual of Analytical Methods (NMAM), Fourth Edition, edited by P. M. Eller, National Institute for Occupational Safety and Health, Cincinnati, OH, [11] El Dorado Hills, Naturally Occuring Asbetos Multimedia Exposure Assessment: Preliminary Assessment and Site Inspection Report, Interim, Final, Ecology and Environment, Inc., U.S. Environmental Protection Agency, Region 9, Contract No. 68-W , San Francisco, CA, [12] R. J. Lee, J. S. Lally, and R. M. Fisher, "Identification and Counting of Mineral Fragments," National Bureau of Standards (NBS) Special Publication 506, Proceedings of the Workshop on Asbestos: Definitions and Measurement Methods, held at NBS, Gaithersburg, MD, July 18-20, [13] J. E. Chisholm, "Discrimination Between Amphibole Asbestos Fibres and Non-asbestos Mineral Fragments," Project Report Ir/L/MF/95/16, Health and Safety Laboratory, Broad Lane, Sheffield S3 7HQ, [14] T. Hartikainen and A. Tossavainen, AIHA Journal, 58, 264 (1997). [15] M. Dorling and J. Zussman, Lithos, 20, 469 (1987). [16] G. P. Meeker, A. M. Bern, I. K. Brownfield, H. A. Lowers, S. J. Sutley, T. M. Hoefen, and J. S. Vance, Am. Mineral., 88, 1955 (2003). [17] A. G. Wylie and J. R. Verkouteren, Am. Mineral., 85, 1540 (2000). [18] M. E. Gunter, M. D. Dyar, B. Twamley, F. F. Foit, Jr., and S. Cornelius, Am. Mineral., 88, 1970 (2003). [19] B. R. Bandli, M. E. Gunter, B. Twamley, F. F. Foit, and S. B. Cornelius, Can. Mineral., 41, 1241 (2003). [20] B. M. Brown and M. E. Gunter, Microscope, 51, 121 (2003). [21] M. S. Sanchez and M. E. Gunter, Am. Mineral., 91, 1448 (2006). [22] G. J. Burdett and A. P. Rood, Environ. Sci. Technol., 17, 643 (1983). [23] B. E. Leake et al., Am. Mineral., 82, 1019 (1997). [24] "A Chrysotile Asbestos Standard Reference Material for Transmission Electron Microscopy," Standard Reference Material (SRM) 1876b, National Institute of Standards and Technology, Gaitersburg, MD, [25] W. J. Campbell et al., "Selected Silicate Minerals and Their Asbestiform Varieties Mineralogical Definitions and Identification-Characterization," Information Circular 8751, Bureau of Mines, United States Department of Interior, Washington, D.C. pp. 1-55,