Document Z0rJdE3LxNVOGdXk5E45Z5eO

G. M. Faulring, W. D. Forgeng, E. J. Kleber, and H. B. Rhodes Detection of Chrysotile Asbestos in Airborne Dust from Thermosetting Resin Grinding Authorized Reprint from Journal of Testing and Evaluation. Vol. 3, Mo. 6 Copyright American Society for Testing and Materials 1916 Race StreeT, Philadelphia, Pa. 19103 1975 UCC 025624 G. M. Faulting,1 W. D. Forgeng,1 E. J. Kleber,1 and H. B. Rhodes1 Detection of Chrysotile Asbestos in Airborne Dust from Thermosetting Resin Grinding REFERENCE: Faulrmg, G. M.t Forgeng, W. 1)., Kleber, E. J., .tnrj Rhodes. H. B., "Detection of Chrysorile Asbestos in Air borne Dust from Thermosetting Resin Grinding," Journal of Testing and Ezaiuation, JTEVA. Vol. 3, No. Nov. 1975, pp. 482-490. ABSTRACTT: Airborne dust samples generated during grinding thermosetting resin plagues containing 0.8 to 18% chrysotile and reference samples were examined optically and nn the electron microprobe up to magnifications of X900. Chryaotile in liuBt samples was analyzed by new methods using iodine in solution or ad a vapor for selectively staining chrysotile and inrreaaing its visibility and by the NIOSH phase contrast method. No free fibers of chrysotile were detected m dust collected from 0 8% chrysoule-polyester plaques. Only a low-level concen tration was found in dust from 4-18% chrysolite-polyester plaques. The difference :n texture of chrysotile and cellulose esteT membrane iCEM) filters was emphasized by vacuum deposition of aluminum and carbon. The importance of care in handling rhrypotiie-hearing CEM filters and the improvement in retention with carbon coating are described. Electron microprobe methods are presented that distinguish between chrysotile completely encapsulated in rosin and chrysotile with a free surface. The airborne chrysotile in the grinding dust was usually nonfibrnus, encapsulated in resin, and closely associated with other materials. In the samples analyzed, the number of chrysotile fibers with a free surface varied from zero to two per thousand dust particles. KEY WORDS: thermosetting resins, serpentine, iodine, chrysotile, light (visible radiation), phase contrast, Tesin-encapsu'auon, vacuum deposition, electron microprobe At the present time, the amount of fibrous asbestos in air borne dust rollected on membrane filters* is generally determined by the phase contrast method developed by Edwards and Lynch (7), hereafter designated the NIOSH (National Institute of Occupational Safety and Health) phase contrast method. The most evident defect in this procedure is the failure to discriminate between various kinds of fibers. Crossrnon 0?j and Julian et al [3] identified each of the six asbestos-tvpe minerals with dispersion staining, an optical method of producing color that depends on variations in refractive index of the mounting medium and the constituent being identified with different wavelengths of light. Thus, its application is limited. This paper describes methods that rapidly detect free chrysotile asbestos in dust samples. Reference samples prepared from blends of known materials and from dust samples produced by grinding chrysotile-bearing thermosetting resins, polyester, and epoxy were collected by settling m water and on membrane filters. These samples were - (Senior :wierit.iet. Metals Div.; ';onsuitani. Metals Div,; air monitoring engineer. Caiidna Asbestos; and technology manager. Caiidna Asbestos, respectively, Union Carbide Curp . Niagara Fall.*, N.Y. : Midipore Filter CV-rp . Bedford. Mxs.. eeJiitioee ept-or membrane fillers. examined optically and on the electron microprobe. We used new' procedures for selectively staining chrysotile and for emphasizing the difference in texture of the chrysotile and the filters by vacuum deposition of carbon. The retentive power of membrane filters for chrysotile was investigated. An electron microprobe method for distinguishing between resin-encapsu lated and free chrysotile and for determining the percentage of free chrysotile particles in dust samples was developed. Sample Preparation and Description Resin plaques containing 0.8 to 18% chrysotile were ground with a power-driven hand grinder equipped with a 7-in. (177.8mm) diameter, 16-grit abrasive disk simulating the fabrication operations found in boat yards and the automobile industry. Compositions of the plaques are listed in Table 1. The grinding was carried out in a thoroughly cleaned, closed, 8M by 8)^ by 8-ft (2.6 by 2.6 by 2.4-m) room. The plaque, approximately H-in. (12.7-mm) thick by 1 by 2 ft (0.3 by 0.6 m), was clamped to a bench 30 in. (76.2 mm) from the floor approximately in the center of' the room. The edge was ground for a period of 4-5 min with the grinder rotating in a direction to throw the heavy particles toward the floor. Four petri dishes filled with distilled water with a total surface area of 425 cm1 were used to collect samples of dust as it settled from the air during each grinding operation. Another set was collected over a 2-h period after each grind. Three of the sample dishes in each set were placed in the comers of the room, two at the 5-ft (1.5-m) level and one at the 2-ft (0.6-m) level and the fourth was located on top of the bench, 3 ft (0.9 m) from the grinding area. Each set wins combined into one sample and the residues concentrated by evaporation at room temperature. fn addition, individual samples were collected before, during, TABLE l--Compositions of chrynotde-bearing plaques. J~amp e No. Resin Weight, Type % Chrysotile Weight, Type- % Fiber Glass Weight, % Sisal Weight. or tC Lime stone Weight. % polvester 79.2 RG-244 OR 20.0 o polyester 99.2 RG-244 0.8 o polyestej 98.0 RG-244 2.0 4 polyester yfi.o RG-244 4.0 5 epoxy 70.0 KC-144 4.0 20.0 <; polyester :15 0 RG-110 5.0 10.0 10.0 41). 0 7 polyester 03.0 RG-110 10.0 10.0 10.0 35.0 8 polyester :). 0 SG-200 18.0 2 0 50,0 'Types RG-110 and RG-144 are opened fibers: SG-2(Xh pelletized; IKj-214. silica nKiU'd (Union Carbide Corporation designations of re fined Californii! <;hry>nie ro/mneroal producis'-. 482 UCC 025650 G. M. Faulring/ \V. D. Forgeng,1 E. J. Kleberg and II. B. Rhodes1 Detection of Chrysotile Asbestos in Airborne Dust from Thermosetting Resin Grinding REFERENCE: Faulring, 0. M.. Forgeng, VV. D.. Kleber, E. J., and Rhodes, H. 13., "Detection of Chrysotile Asbestos in Air borne Dust from Thermosetting, Resin Grinding," Journal of Testing and Evaluation, JTEVA, Vol. 3, No. 6, Nov. 1375, pp. 482-490. ABSTRACT: Airborne dust samples generated during grinding thermosetting resin plaques containing 0.8 to 18% chrysotile and reference samples were examined optically and on the electron naicroprobe up to magnifications of X900. Chrysotile in dust samples *as analyzed by new methods using iodine in solution or as a vapor for selectively staining chrysotile and increasing its visibility and by the N'lOSH phase contrast method. No free fibers of chrysotile were detected in dust collected from t>.$% chryao tile-poiyester plaques. Only a low-level concen tration was found in dust from 4-18% ebrvsotile-polyester plaques. The difference in texture of chrysotile and cellulose ester membrane iCEM) filters was emphasized by vacuum deposition of aluminum arid carbon. The importance of ear in handling chrysotile-bearmg CEM filters and the improvement in retention with carbon coating are described. Electron microprobe methods are presented that distinguish between chrysotile completely encapsulated in resin and chrysotile with a free surface. The airborne chrysotile in the grinding dust was usually nonfibrous. encapsulated in resin, and closely associated with other materials. In the samples analyzed, the number of chrysotile fibers with a free surface varied from zero to two per thousand dust particles. KEY WORDS: thermosetting resins, serpentine, iodine, chrysotile. light -.visible radiation), phase contrast, reem-encapsu'ation, vacuum deposition, electron microprohe At the present time, the amount of fibrous asbestos in air borne dust collected on membrane filters' is generally determined by the phase contrast method developed by Edwards and Lynch [1], hereafter designated the ATOSH (National Institute ol Occupational Safety and Health) phase contrast method. The most evident defect in this procedure is the failure to discriminate between various kinds of fibers. Crossmon [2] and Juliuu et al (3] identified each of the six asbestos-tvpe minerals with dispersion staining, an optical method of producing color that depends on variations m refractive index of the mounting medium and the constituent being identified with different wavelengths of light. Thus, its application is limited. This paper describes methods that rapidly detect free chrysotile asliesto.-, in dust, samples. Reference samples prepared from blends of known materials and from dust .samples produced by grinding chrysotile-bcaring thermosetting resins, polyester, and epoxy were collected by settling in water and on membrane filters. These samples were 1 Senior scientist. Metals Div., rrinsuUam., Metals L>iv.: air monitoring engineer. Caiiciria Asbestos: an-i tenhnn.ogy manager. Calidna Asneslos. restiortiveiy. (.ns.n Carbide C:,r|>., Niagara Fails, N.V. : Mhlipore Filler Carp., Bedford. Mass., .Uuiose ester membrane filters. examined optically and on the electron microprobe. H'e used new procedures for selectively staining chrysotile and for emphasizing the difference in texture of the chrysotile and the filters by vacuum deposition of carbon. The retentive power of membrane filters for chrysotile was investigated. An electron microprobe method for distinguishing between resin-encapsu lated and free chrysotile and for determining the percentage of free chrysotile particles in dust samples was developed. Sample Preparation and Description Resin plaques containing 0.8 to 18% chrysotile were ground with a power-driven hand grinder equipped with a 7-in. (177.8mm) diameter, 16-gnt abrasive disk simulating the fabrication operations found in boat yards and the automobile industry. Compositions of the plaques are listed in Table 1. The grinding was carried out in a thoroughly cleaned, closed, %V> by 8}4 by 8-ft (2.0 by 2.0 by 2.4-m) room. The plaque, approximately H-in. (12.7-mmj thick by i by 2 ft (073 by 0.6 m), was clamped to a bench 30 in. (76.2 mm) from the floor approximately In the center of the room. The edge was ground for a period of 4-5 min with the grinder rotating in a direction to throw the heavy particles toward the floor. Four petri dishes filled with distilled water with a total surface area of 425 cm1 were used to collect samples of dust as it settled from the air during each grinding operation. Another set was collected over a 2-h period after each grind. Three of the sample dishes in each set were placed in the corners of the room, two at the 5-ft fl.o-m) level and one at the 2-ft (0.6-m) level and the fourth was located on top of the bench, 3 ft (0.9 m) from the grinding area. Each set was combined into one sample and the residues concentrated by evaporation at room temperature. In addition, individual samples were collected before, during, TABLE l--Compositions of chrysoiile-beannQ plagues. Sample No. Resin Weight, Type % Olir> sotile Weight, Type* % Fiber Glass Weight. % Weight, % Lime stone Weight. <tTC* 1 polyester 79. 2 RG-244 0 $ 20.0 2 polyester 99.2 HG-244 0.8 .3 polyester 98.0 RG-244 2.0 4 polyester 96.0 RG-244 4.0 0 epoxy 76.0 RG-144 4.0 jo.o b polyester 35 0 KG-110 5.0 10 0 10.0 40.0 7 polyester 35.0 RG-110 10.0 10 0 10 0 35.0 S polyester 30.D SO-200 18.0 2.0 50.0 1 Typvs RG-110 and RG-I44 are opened fibers. SC3-20O, pelletized; KG-244, silica coated (L'nion Carbide Corporation designations u! re fined California chrysotile. commercial products!. 482 UCC 025651 FAULRtNG ET AL ON DETECTION OF CHRYSOT11.E ASiESTOS 483 and after each grinding operation by aspirating at the rate of 2 litres per min through Type AA membrane filters (0.8 iiDi pore size). Sections of the plaques were polished on both sides to less than 0.25 mm in thickness and examined with transmitted ordinary and polarized light. The chrysotile was present as fibrous particles frequently associated with strain in the resin matrix. A low magnification (X100) examination of concentrated dust samples showed the samples contained not only the material from the plaques but also materials such as fibers of cotton, wood, nylon, and glass (Fig. 1). The presence of extraneous airborne materials emphasizes the importance of distinguishing chrysotile from other fibrous materials. Equipment Description The optical equipment used in the transmitted ordinary and polarized fight studies consisted of a Leitz petrographic micro scope, Kohler-type illumination with a tungsten light source, X45 [0.65 numerical aperture (X.A.)j and X90 (1.32 N.A.) objectives and & positive, X10 eyepiece. The electron microprobe analyzer was an Acton model no. MS-64. Staining Methods Morton and Baker (4! reported that a 1% solution of iodine in glycerine selectively stains chrysotile, serpentine, brucite, and, to a lesser extent, hydromagnesite. The same authors also reported that other organic solvents such as ethyl alcohol, carbon tetrachloride, butyl Carbitol, and tricresyl phosphate may be used. This stain differentiates chrysotile fiber from all other fibrous materials including organic and inorganic fibers and depends on the activity of the iodine and the hydrated magnesia surface of the chrysotile. (Chrysotile is composed of a la/er of magnesium hydroxide on a silica substrate.) Mixtures of polyester, epoxy, mineral wool, calcite, sisal, or any combination, with 0.5% chrysotile and dust samples were treated with sublimed iodine and iodine dissolved in various solvents and were examined with transmitted light at X450 and X900. Of these materials only chrysotile was stained a characteristic orange color and the staining appreciable im proved its visibility. However, some solutions were superior. Figures 2-4 show the colors developed by the chrysotile after being stained by various methods. Optical observations, staining solutions, and the new method for staining with sublimed iodine are given below. Solution 1 Solution 1 (1 g iodine + 100 cm1 glycerine) is an orange-brown color and stains chrysotile a pale orange; polyester, pale yellow; and epoxy, dark red-brown. (No photomicrographs are shown.) Before staining, polyester and epoxy are transparent and without color. Color changes in the other materials were not noted. The stained chrysotile was easily distinguished in the water-collected, air-dried samples. Solution 2 Iodine is appreciably more soluble in a solution of ethyl alcohol and glycerine than in pure glycerine. Solution 2 (2.5 g iodine -f 50 cm3 ethyl alcohol + 50 cm1 glycerine) produced darker colors than Solution 1, resulting in chrysotile being more easily detected (Fig. 2). Solution 3 The orange-stained chrysotile is easily noted in Solution 3 (1 g iodine + 100 cm3 benzene) because of its contrast with the pale violet color of the liquid (Fig, 3). Iodine is violet in nonbasic solvents such as benzene and orange or brown in basic solvents such as ethyl alcohol and ethyl alcohol plus glycerine. After the solution was placed on the specimen, a drop of Crown immersion oil was added to decrease the evaporation rate of the benzene. Small fibers of chrysotile were detected that would probably have been overlooked if stained with Solution 1 or 2. Other materials were also examined (sisal, cotton, wood, synthetics, and glass), and only the epoxy and polyester were stained. Chrysotile is rapidly stained in any of the three solutions; however, about l/i h is required for the epoxy and at least 6 h for the polyester to develop the indicated colors: dark red-brown FIG. 1--hurt coitected in water after grinding 0.8% 6 plaque: (left) ordinary light and (right) polarized light. *' w UCC 025652 484 JOURNAL Of TESTING AND EVALUATION and yellow, respectively. One week later the specimens were reexamined. The characteristic color of the stained chrysociie, polyester, and epoxy intensified with time. Solution 3 was also carefully applied to sections of membrane filters containing reference and grinding dust samples. The filter was placed on Crown immersion oil to make it transparent and examined with transmitted light. Although Crown immersion oil was used throughout this investigation, other types of im mersion oil should be equally effective. The NIOSH standard phthalate-oxalate-membrane solution could not be used since it bleaches the stain. Sublimed Iodine Chrysotile reference standards and air-dried dust samples were placed in covered containers with iodine. On standing at room temperature, the iodine sublimed and stained the free chrysotile orange, and after several hours the chrysotile inside the polyester resin was also stained. For example, Fig. 4 shows orange-colored chrysotile fibers inside a pale yellow polyester particle. The color of the chrysotile faded on standing at room temperature (<1 h), particularly when it was not encapsulated in resin. This was avoided by mounting the stained sample in immersion oil and using a cover glass. Chrysotile deposited on membrane filters was also stained by exposure to the iodine atmosphere. Although the chrysotile stained in 20 min, an exposure time of about an hour produced optimum contrast between the chrysotile and the filter. After staining, the filter was placed on one to two drops of immersion oil and covered with a cover glass and the chrysotile identified. Stained chrysotile is also more easily detected with phase con trast microscopy than nonstained. Microscopic Analyses The water-collected dust samples previously described were concentrated by evaporation. Five slides of each sample stained with Solution 3 were examined at X450 with transmitted ordinary and polarized light. A field typical of the dust from a LOTc chrysolite-bearing sample as it appears with parallel polars is shown in Fig. 5. This field contains sisal, fibrous glass, resin, and calcite but not chrysotile. The chrysotile in the dust samples was generally in the form of agglomerated masses of fibers rather than fibers or bundles of fibers; however, it was easily distinguished from the other phases by the orange stain. The agglomerated nonfibrous-shaped chrysotile particles are probably generated by the chopping or shredding of bundles of fibers during the grinding operation since only fibrous chrysotile was detected in thin sections of the plaques. Although other investigators )J] indicated that chryso tile curls up in water, an optical examination of the chrysotile used in this work did not show any tendency to change shape, even after immersion for extended periods. As additional identi fication characteristics, chrysotile has weak birefringence and a shimmery appearance when rotated in polarized light with crossed polars. The other materials present in the dust were identified as follows: sisal, by its lack of staining and high birefringence; epoxy and polyester, by the colors produced on staining (red brown and pale yellow): and fibrous glass, by its distinctive shape. Epoxy, polyester, and fibrous glass are isotropic in polar ized light. Calcite has high birefringence and a refractive index of about 1.66 in the ordinary direction. Stained fibrous chrysotile particles with a ratio of length to diameter (L/D) >3 and more than 3 urn in length were counted on five slides of each concentrated dust-sample. The results are shown in Table 2. The number of free fibers increased with chrysotile content of the polyester resin plaques: no fiber at 0.8%, four at 2%, and a maximum of ten in the 4-18% level. No free fibers were found in the dust from the 4% epoxy resin plaque. This sampling and counting procedure was not used to establish a quantitative measure of airborne chrysotile fibers. However, the results indicate that only a few fibers of chrysotile were released by grinding chrysotile-bearing thermosetting resins and that epoxy resin, which "wets" the asbestos surface more readily than polyester, possibly releases fewer fibers than polyester resin. Specimens collected on membrane filters for various times TABLE 2--Comparison of optical and electron microprobe analysis of chrysotile asbestos in dust samples. Microscopic Electron Probe Examination Method d'.ain Grinding Period during Collection Method water Collection Time, min 7 Particle Shape L. D > 3 Particle Length, nm > 3 stain NIOSH NIOSH NIOSH NIOSH Method 1 after before during during* after during water membrane membrane membrane membrane water 90 60 7 7 90 L/D > 3: L/D > 3 L- D > 3 L/D > 3 L/D > 3 5 nonfibroiLs > 3 > ft > ft > ft > 5 > 3 Method 1 during water 5 fibrous >3 Method 1 after water 120 uonfibrous >3 Method 2 Method 2 Method 2 after during during water membrane membrane 120 fibrous 7 nonfibrous 7 fibrous >3 >3 >3 sample No l 2 3 4 5 6 7 S No. fUxcs/5 slides 00 00 40 10 3 00 00 48 8S 0 0.0! 0.01 0.01 0.06 0 0 0.03 No. fibers.'cm1 of aiT 0 0 0.2 1.2 0.1 1.2 0 09 0 0 0 0.7 0.1 0.6 0 0.3 0 0 0.01 0.02 0.06 0.02 0 0.06 3 0 No. particles/1000 dust particles 04 0 606 7 28 0 0 10 2 0 1 1 1 0 2 Breathing tone. *-* w V-' xj UCC 025653 FAULTING ET Al ON DETECTION OF CHRYSOTIlE ASBESTOS 485 FIG- 2--Chry*ottie and epoxy stained with xodine-gllKcrine-eUcahol solution {Solution }: (left) XI00 and (right) X500. FIG. 3--CAryeo^tie tiauwd with iodine-fcenzn solution (Solxtfion 3) : (left) X 100 and (right) X500. FIG- 4--Chrysotxlc, free and encapsulated, in polyester resin, stained with tubiimtd iodine ( FIG. 5--Typical dust aampf* collected during grinding of polyester ream. cArysoftie < 10%), glass, sisal, and calcite plague in iodine-benzene solution `.Solutions, X500\. /' --- O \J UCC 025654 486 were analyzed by counting particles longer than 5 din with i/D > 3 by the NIOSH phase contrast method. Fibrous glass and other materials that satisfied the size criteria but were clearly not asbestos were excluded. These counts include fibers other than ehrysotile that cannot be identified by this technique. The results, listed in Table 2, indicate a trend similar to that of the staining data. The correlation of the two methods is par ticularly evident in the number of fibers detected in the dust from the 4% chrysotile-polyester plaque. JOURNAL OF TESTING AND EVALUATION Evaporated Coatings Carbon was vacuum deposited on chrysotile-bearing mem brane filters. Sections of these filters were then mounted on either the phthalate-oxalate mounting medium or immersion oil and examined. As shown in Fig. 6, carbon emphasizes the difference in texture of the filter and ehrysotile, improving the detection of small ehrysotile fibers. In polarized light and crossed polars, a thin deposit of carbon increases the apparent bire fringence of ehrysotile. A polarized light examination is essential for distinguishing ehrysotile from other fibrous materials in samples prepared by this method. A similar textural emphasis but not apparent increase in birefringence was achieved with evaporated aluminum. Fraser (6'] reports that the conical shape of the pores in Type AA cellulose ester filters restricts the penetration of particles to a depth of 15 to 20 m, and Mercer [7], to a depth of 50 i>m. In a discussion of a paper by Thomas [Si, Green [9] reported that the penetration depends on several other factors, including air flow through the filter. If the carbon evaporation method is to be of value, it is necessary that the particles detected at X450 be located on or very near the surface of the filter. To determine the depth of penetration, photomicrographs were taken at X450 of the same area on a carbon-coated ehrysotile sample with ordinary and polarized light at two positions of brightness (90 deg apart-- crossed polars) and compared. The particles detected with ordinary light were on and within 2 am of the surface and with polarized light and crossed polars, on and throughout the filter. FIG. 7--Chryeelilc on CEM filter--carbon coaled: (A) local ion after coating and IB) location after tranepoeing. No particles located more than 2 *<m below the surface of the filter were found. Particle Retention on Membrane Filters During the optical examination of the carbon- and aluminumcoated ehrysotile samples collected on membrane filters, the particles frequently moved. For example. Fig. 7 shows the transposition of a carbon-coated ehrysotile particle that occurred when the sample was prepared for microscopic examination. The retentive power of membrane filters has been attributed to electrostatic charges developed by the passage of air through the filter. It was of interest to establish whether the evaporated coatings or sublimed iodine dissipates these charges, resulting in the particles being more sensitive to handling, or if particle movement also occurs on noncoated filters but is not recognized. Filters containing ehrysotile in the as-deposited state and after staining with sublimed iodine were examined in reflected light. A comparison of photomicrographs taken of the same areas before and after gently tapping the back of a slide supporting the stained and unstained ehrysotile sections showed several particles were removed. Similar experiments were made using ft gentle flow of air for about 1 s across the surface of the filter sections. The amount of stained and unstained ehrysotile re moved from the filters appeared to be about the same. Filters containing somewhat large amounts of ehrysotile were divided, and half of each filter was subjected to various me chanical treatments. The other half was the control sample. To emphasize differences, the treatments were more severe than dust-bearing filters would normally receive during transporta tion and handling. Some of the chrysotile-bearing filters were carbon coated to determine the effect of carbon. The effect of an evaporated coating of aluminum was assumed to be similar to that of carbon and therefore was not analyzed. The same physical treatment for the carbon-coated and carbon-free filters was ensured by placing halves of each filter in a small plastic container (5 cm diameter by 1 cm) and rolled on the floor about 4 ft (1.2 m). Some filter halves were cut with scissors and others were dropped five times in a container about 4 ft (1.2 m). The filter sections were free to move in the con- ^ U xj UCC 025655 FAUIIUNG ET At ON DETECTION OF CHRYSOTILE ASiESTOS 487 tainere. The magnesium contents of the control and treated samples shown in Table 3 were determined by atomic absorption. These data indicate the carbon deposit has a tendency to secure the chrysotile to the filter and emphasize the importance of careful handling. Electron Microprobe Analyses The dust samples collected in water and on membrane filters were analyzed to identify chrysotile and to distinguish between resin-encapsulated and free chrysotile on a quantitative basis. Free chrysotile particles refer only to a free surface exposed to the electron beam; the particle could be partially encapsulated. The methods and results are discussed below. Method 1 Method 1 detects free chrysotile and chrysotile encapsulated in thin coatings of resin in water-collected dust samples. The dust samples concentrated by air evaporation, water-dispersed chrysotile, and silica-coated chrysotile* reference samples were placed on polished copper specimen mounts, air-dried, and 1 Calidria Asbestos RG-244, Union Carbide Corp.. Niagara Falla, N.Y. carbon coated. Standard reference specimens were also prepared by compacting them when dry under pressure and carbon coating. The intensities of the MgK. and SiK. X-rays generated in the reference specimens were measured, and ratios of these intensities representing the relative amounts of magnesium and silicon were established. A similar method could be used to analyze samples collected on membrane filters. Whether the samples were pre pared from a wet or dry state did not affect the results. The electron beam was positioned on 1000 dust particles by scanning a randomly selected area. Each particle was analyzed for magnesium and silicon, and the quanta of MgK. and SiK. generated were measured. Derived relative amounts of mag nesium and silicon were compared with similar data from standards and the chrysotile particles were identified. Whenever chrysotile was detected, the shape of the particle and any associated orange cathodoluminescence, characteristic of the resin phase, were noted. Frequently, electron microprobe images were taken. The images in Fig. 8 are representative of most of the chrysotile-bearing particles detected. The same or nearly the same locations of the white areas on the electron images, distinctive of the tesin phase, and the areas enriched in magnesium and FIG- 8--Chrysotile particles m Just samples frirm 1% ckrysotiU-poigeslcr plague itop) collected JunriQ grinding, X.125 and (bottom) collected after grinding, XdtX): (left to right; electron image, magnesium, and silicon Chrysotile particles are emphasized with arrows. UCC 025656 488 JOURNAL OF TESTING AND EVALUATION TABLE 3--Maffnetium corticate of control and treated tamptes. Sample Treatment Reference Sample After rolling 4 ft (1.2 m) Reference Sample After cutting with Scissors Reference Sample After dropping 4 ft (1.2 m) (5 times) UncoateU % Mg % Mg Removed 0.085 0.039 0.12 0.09 0 062 54. 1 25 0.042 32.3 Carbon-Coated % Mg % Mg Removed 0.056 0.051 0.16 0.14 8.9 12.5 silicon on the X-ray scanning images, characteristic of chrvsotile, show the clow; association of chrvsotile and resin. The chrysotile particles detected during these analyses were usually equidimensional or lamellar rather than fibrous, as a result of grinding. For example, see Fig. S, especially the mag nesium and silicon X-ray scanning images. Figure 9 is typical of the dust sample collected from the 10% chrysotde, polyester, glass, and calcite plaque. The phases indicated by numbers on the electron scanning image were identified as follows: No. 1, glass; N'o. 2, chrysotile, resin, and calcite conglomerate; and No. 3, calcite. The numbers of chryso tile particles per 1000 dust particles detected by this method are presented in Table 2. One set of the samples analyzed by this method was collected in w'ater for 5 min during.grinding and the second set, for 2 h after grinding. The maximum number of fibrous chrysotile particles detected, either free or encapsulated in a thin resin coating, was 2. Throughout these analyses, it was difficult to find fibrous chrysotile particles. For example, 28 nonfibrous but no fibrous, free, or encapsulated chrysotile particles were de tected in a dust sample from a 10% chrysotile-bearing plaque. Method S -Method 2 was developed to detect only free chrysotile or chrysotile with a surface exposed to the electron beam in dust samples collected on membrane filters. It could also be applied to dust samples collected by other methods. A standard sample of pure chrysotile, Calidria asbestos RG-114,1 and a dust sample containing only resin-encapsulated chrysotile were collected on membrane filters. Sections of these two reference filters and the dust-bearing filters to be analyzed were secured to a specimen mount and carbon coated. The coat ing was deposited sufficiently thick to prevent the detection of the MgK,, and SiK X-rays generated in the encapsulated chrysotile standard but thin enough for easy detection when produced in the free chrysotile standard. In a variation of this method, the carbon coating thickness was adjusted to detect only the SiK,, X-rays from the encapsu lated chrysotile standard, The intensity of the SiK. X-rays is three to four times greater than that of the MgK,, X-rays Union Carbide Corp., Niagara Falls, N.Y. FIG 'J--Dubi rodectcd during grinding of pAj+^ter plaque containing 10%. chrysotile.. 10% glass. 10% sisal, and 35% ralcite. (Arrow I) a glass fiber' (Arrow a rhryso{ile-bearuig particle, i Arrow A) aunte '.Top, :eft to night) electron \mage, magnesium, and aluminum; ^bottom. left to right) calcium and sfficyn. UCC 025657 RAdRIHG ET At OH DETECTION Of CHRYSOTIIE ASBESTOS 489 Thus, a thickness of the carbon coating can be deposited that will cut off any indications of the MgK X-rays and still permit the detection of the SiK,, X-rays. Such a carbon coating allows easy detection of both -WgK. and SiK. X-rays generated from free chrysotile and thus discriminates between free and encapsu lated chrysotile. After adjustment of the carbon coating thickness, the dust bearing membrane specimens on the same mount were analyzed. The electron beam was positioned on 1000 particles ' and the intensities of the MgK. and SiK. generated were measured. Ratios of these intensities were compared to similar ratios derived from the standard reference sample. The electron micro probe was operated at lower than normal specimen current and accelerating voltage, 50 nA and 15 kV, and the dwell time of the beam on each particle minimised to avoid melting or decompos ing the filter. Images typical of the free chrysotile standard prepared by this method are shown in Fig. 10, and a chrysotile particle associated with a trace of calcite in a dust sample, is shown in Fig. 11. The number of free chrysotile particles detected per 1000 dust particles is shown in Table 2. The free fibrous chrysotile particles range from 0 to 2 per 1000 dust particles and the nonfibrous from 0 to 7. Summary Dust samples generated by grinding polyester and epoxy resin plaques containing up to 18% chrysotile asbestos were collected by settling in water and on membrane filters. Samples were examined optically and with the electron microprobe. The dustbearing membrane filters were also analyzed by the NIOSH phase contrast method. Staining procedures for identifying and increasing the optical detectability of chrysotile asbestos are presented. These pro cedures using iodine in solution or as a vapor are specific for the hydrated magnesia surface and distinguish chrysotile from all other commonly encountered fibrous materials. They can be applied to membrane filters but require the use of immersion oil to clear the filter. Extensive optical examination of the samples collected in water, although not a quantitative measure of airborne chryso tile dust levels, shows clearly that free fibrous chrysotile is rare. There appears to be a trend with chrysotile content: no fibers at 0.8% chrysotile in the resin, a few at 2% chrysotile, and a low level at 4-18% chrysotile. The membrane filter samples analyzed by the NIOSH phase contrast method show similar trends. Vacuum deposition of carbon and aluminum improves the detection of chrysotile by emphasizing the textural differences between chrysotile and the supporting membrane filter. Carbon deposition also increased the apparent birefringence of the chrysotile. Studies were designed to compare the membrane filter reten tion of stained, carbon-coated, and as-deposited chrysotile. Results established that staining by sublimed iodine has no effect, whereas vacuum-deposited carbon has a tendency to secure the chrysotile to the filter. Movement of carbon-coated chrysotile on the filters was difficult to avoid but easily detected during the optical examinations. As-deposited chrysotile on membrane filters is probably more easily moved during han dling than generally recognized, but, because of difficulties associated with its detection, overlooked. An electron microprobe method was applied to water-collected FIG. 10--Chrysotile collected on CBM filler (Method ). Chrysotile particles ore emphasised with arrows: ,'top) electron image, (middle) magnesium, and (bottom) silicon samples to measure the number of chrysotile particles per 1000 dust particles that were either encapsulated in a thin coating of resin or had a surface exposed to the electron beam. It was found that the chrysotile was usually present as somewhat equidimensional, resin-encapsulated particles that were fre quently conglomerated with other materials: A total of 2 fibrous UCC 025658 ^ o ij 490 JOURNAL OF TESTING ANO EVALUATION FIG. 11--Dust collected on (.'EM (filer orinding of polyester plaque containing 10% chrysolite. 10% sisal. 10% glass, and 35% calcxte Method ?) Ckrysottie particles are mnphasued u-ih arrovrs. (Top left.) electron image, (top right' magnesium, (bottom left) calrium, and (bottonn right) silicon. rhry?otile particles was found in 4 dust samples (4000 particles counted) from plaques containing 2 to 10% asbestos. A new met nod was developed to distinguish between resitiencapsulated ehrysotile and chrysotile particles with a free surface exposed to the electron beam. This method wa3 applied to count the number of chry.sotile particles with a surface exposed to the electron beam in situ on membrane filters. A ckno'.dcdgrrunis The authors would like to acknowledge the assistance of K. S. Malizie, for his electron inieroprobe analyses, and K. M. Hamner, for the atomic absorption analyses. References [rj Edwards. G. H. and Lynch, J. R., A nook of Occupational Hygiene, Vol. 11. So. 1. 1!H58, pp. 1 -6. [i) Crossmon, G.. J/'icrvcftemtcav Journal, Vol. 10, 19(56, pp. 273-2S5. [i] Julian. Y.and MrCrone. \V. C., The Microscope. Vol. 18, No. 1. 1970. pp. 1-L0. (4) Morton, M. and Baker. W. G., Transactions of the Canadian Institute of Mining and Metallurgy. Vol. 44, No. 354. 1941, pp. 515-523 (5] Wall, II., unpublished International Report. Occupational Health Services. Environmental Health Branch. Ontario Department of Health, June 1900. [6\ Fraser. D. A., Archives of Industrial Health. Vol. II, No. 8. 1953, pp 412-419. I?) Mercer. T. T.. .4 retires of Industrial Hygiene and Occupational Medicine, Vol, Hi. No. 5. 1934. pp. 372-379. (5| Thomas. D. H . .1`rumal of the. Institute of Heat and Ventilation Engineers, Vol. 25>, May 1952, pp 35-55. [91 Green, H. L., Journal of the Institute of Heat and Ventilation Engineers. Vol. 20. May 1952. pp. 55-70. UCC 025659 ^ ^ C v/ O'