Document 4QB6gqv2KyNmJdOGMXyBBQRaQ

FILE NAME: Henry Company (HC) DATE: 2007 DOC#: HC006 DOCUMENT DESCRIPTION: Journal Article - Simulation Test to Assess Occupational Exposure to Airborne Asbestos from Asphalt-Based Roofing Products Ann. Occup. Hyg., Vol. 51, No. 5, pp. 451-462, 2007 The Author 2007. Published by Oxford University Press on behalf of the British Occupational Hygiene Society doi:10.1093/annhyg/mem020 Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Simulation Tests to Assess Occupational Exposure to Airborne Asbestos from Asphalt-Based Roofing Products FIONNA MOWAT1*, RYAN WEIDLING2 and PATRICK SHEEHAN2 1Health Sciences, Exponent, Inc., 149 Commonwealth Drive, Menlo Park, CA 94025, USA; 2Health Sciences, Exponent, Inc., 500 12th Street, Suite 220 Oakland, CA 94607, USA Received 4 January 2007; in final form 26 March 2007 This study sought to evaluate exposure from specific products to evaluate potential risk from roof repair activities. Five asbestos-containing fibered roof coatings and plastic cements, rep resenting a broad range of these types of products, were tested in exposure simulations. These products were applied to representative roof substrates. Release of asbestos fibers during ap plication and sanding of the product shortly thereafter (wet sanding) were tested initially. Other roof substrates were cured to simulate a product that had been on a rooftop for several months and then were tested to evaluate release of fibers during hand sanding and hand scrap ing activities. Additional tests were also conducted to evaluate asbestos release during product removal from tools and clothing. Two personal (n = 84) and background/clearance (n = 49) samples were collected during each 30-min test and analyzed for total fiber concentration [phase-contrast microscopy (PCM)] and for asbestos fiber count [transmission electron micros copy (TEM)]. PCM concentrations ranged from <0.005 to 0.032 fibers per cubic centimeter (f cc_1). Chrysotile fibers were detected in 28 of 84 personal samples collected. TEM concen trations ranged from <0.0021 to 0.056 f cc_1. Calculated 8-h time-weighted averages (TWAs) ranged from 0.0003 to 0.002 f cc_1 and were comparable to the background TWA concentra tion of 0.0002 f cc_1 measured in this study. Based on these results, it is unlikely that roofers were exposed to airborne asbestos concentrations above the current or historical occupational guidelines during scraping and sanding of these products during roof repair. Keywords: asbestos; curing; exposure simulation; occupational exposure; plastic cement; roof coating IN T R O D U C T IO N Asbestos was used in the past in a variety of roofing products, such as a filler material in asphalt, roofing felts and mastics (Stern et al., 2000), as well as a structural fiber in asphalt- and coal tar-saturated ply sheets and flashing felts (Brzozowski, 1989). In 1994, the National Roofers Contractors Association (NRCA, 1994) estimated that 20% of built-up roof felts and >90% of flashings, coatings, cements and mastics on commercial low-sloped roof jobs con tained asbestos. The use of asbestos in roofing mate rials was discontinued in the early 1980s, with the exception of roof coatings, mastics and cements. These exceptions contain encapsulated asbestos in certain product lines sold today (NRCA, 1994). *Author to whom correspondence should be addressed. Tel: (650) 688-1782; fax: (650) 688-1799; e-mail: fmowat@exponent.com The roof coatings, mastics and cements are widely used on various parts of `roof systems' for water proofing and adhesion, including adhering flashing to roof surfaces or sealing around chimneys, air vents and other penetrations (NRCA, 1994). Some studies report that asphalt-based roof coatings are typically composed of 5-10% asbestos, and the more viscous roofing cements typically contain 15-20% asbestos (ICF Incorporated, 1988). Others note that the amount of asbestos used is proportional to the desired final consistency--that is more fluid coating-like ma terials contained 5-20% asbestos, while heavy mas tics contained 15-45% asbestos (Brzozowski, 1989; NRCA, 1994). In general, the asbestos fibers in asbestos-containing roofing materials are encapsulated in bituminous matrices or resins, thereby binding the fibers and preventing their release (NRCA, 1994). During the course of their work, many roofers may have been 451 452 F. Mowat, R. Weidling and P. Sheehan Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 exposed to asbestos during the repair or tear-off of old roofing materials, including roofing cements and coatings. Very few studies have reported airborne concentrations of asbestos associated with the appli cation, repair or removal of roof coatings or cements. Data do exist for a subset of the coating products-- sprayed-on coatings--and airborne concentrations of asbestos during the spraying of asphalt-based roof coatings were reported to range from 0.003 to 0.3 fibers per cubic centimeter (f cc-1) based on data collected in the 1970s (Anderson et al., 1982). Two specific activities were evaluated regarding the sprayed coatings: spraying asphalt cutback (0.003 0.15 f cc-1 ; sample duration, 342-432 min) and spraying asphalt emulsion (0.01-0.3 f cc-1 ; sample duration not provided). Studies conducted by the NRCA from 1986 to 1991 indicate that removal of asbestos-containing roof flashings, mastics, coatings and cements yielded low asbestos fiber concentrations (range, 0.004 0.027 f cc-1 ; mean, 0.024 f cc-1) (NRCA, 1994). In addition, the NRCA reports that, of the >1500 in dividual monitoring results for built-up roofing, none exceeded 0.1 f cc-1 . These data consisted of several thousand personal breathing-zone and area samples collected between 1986 and 1991. Anderson et al. (1982) also reported airborne asbestos concentrations of 0.1 to 0.6 f cc-1 during tear-out and replacement of roofing materials, but these data were not specific regarding the type of material removed. Finally, in a study of weathered roofing materials, estimated time-weighted averages (TWAs) during building de molition in old Australian buildings ranged from 0.02 to 0.60 f cc-1 (Brown, 1987). Due to the use of fibro sheeting (a corrugated asbestos cement product) and other asbestos-containing materials in the construction of buildings in Australia (NOHSC, 2002), the relevance of these concentrations to the US is unknown. Due to the paucity of exposure data regarding hand-applied roofing cements and coatings, products that were widely used in past decades, and due to the fact that many of the asbestos exposure data on coat ings and cements are complicated by the fact that un derlying substrates (such as flashing, shingles etc.) could also have contained asbestos, this study was undertaken to generate data on fiber releases and to estimate potential exposures specific to these types of roofing products. Five different roofing products that were historically manufactured and sold by Monsey Products Company (Monsey) and Henry Company (Henry) in the 1950s, 1960s and 1970s were selected as representative of the variety of roof ing cements and coatings. The initial objective of this work was to quantita tively measure the airborne asbestos fibers in the breathing zone of a worker during simulated applica tion, scraping and sanding of five products from rep resentative roof substrates, and the scraping of two of the products from clothing and tools. Because evalu ation of exposure during a roofing job entails poten tial exposure to a variety of substances and products (as in the field during a roof removal), this investiga tion sought to isolate the exposures from specific product types during specific tasks. Exposure simula tion was used due to our inability to separate asbestoscontaining coatings and cements from other asbestos materials used in the construction of roofs. The ultimate objective of the testing was to compare the airborne asbestos concentrations from these tests to past and present Occupational Safety and Health Administration (OSHA) Permissible Exposure Limits (PELs) for asbestos fiber exposures. METHODS Five asbestos-containing commercial roofing products that were sold in the 1950s, 1960s and 1970s were evaluated for their potential release of as bestos fibers under different testing scenarios: three fibered roof coatings (B-10 Fibered Roof Coating, 201 Fibered Roof Coating and D-21 Fibered Alumi num Roof Coating) and two plastic roof cements (C-8 Plastic Roof Cement and C-30 Plastic Roof Cement). A brief description of the five products and their intended use is provided in Table 1, along with their composition, asbestos content and physical/chemical properties. These products were con sidered to be representative of Monsey and Henry plastic roof cement and fibered coating product lines. Because the five products tested are no longer manufactured or sold, representatives from Monsey and Henry mixed and produced the product in June 2005 at Henry's facility in Kimberton, PA, with materials obtained from original suppliers when pos sible, so that these exposure tests could be con ducted. The formulated product met all quality control standards specified for the original products, indicating that the products made in 2005 had the same properties as those manufactured historically. These products were mixed on original equipment still present at the site; therefore, we believe that the type of mixing and degree to which the asbestos fibers were coated during the mixing process were similar to historical conditions. Preparation o f roof substrates The plastic roof cements and fibered roof coating products were applied to three different types of roofs--composition asphalt shingle roofing (termed `asphalt shingle'), bituminous built-up roofing (termed `BUR') and modified bitumen roofing (termed `com posite roofing')--depending on the manufacturer's product specification and intended use. The base of each roof substrate was a 4 x 8-ft (total surface area Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Simulation tests to assess occupational exposure to airborne asbestos 453 Table 1. Characteristics and description of asbestos-containing test products Product Composition Chrysotile asbestos content (%) Description/use C-8 Asphalt, petroleum 15.5 distillate and encapsulated asbestos B-10 Asphalt, petroleum 3.04 distillate and encapsulated asbestos Repairing leaks in asphalt, metal and composition roof surfaces. Used for waterproofing on flashing and as an emergency patching compound that provides good weather resistance. Useful in repairing leaks in shingles, concrete and gutters, and around chimneys, downspouts and eaves. Heavy-bodied asphalt coating, reinforced with encapsulated asbestos fibers, used for surfacing asphalt, metal or composition roofs. Adheres to brick, concrete or steel. C-30 Asphalt, petroleum distillate, 4.3 calcium carbonate and encapsulated asbestos 201 Asphalt, petroleum distillate 4.9 and encapsulated asbestos Repairing leaks in asphalt, metal and composition roof surfaces. Used for waterproofing on flashing and as an emergency patching compound that provides good weather resistance. Useful in repairing leaks in shingles, concrete and gutters, and around chimneys, downspouts and eaves. Heavy-bodied asphalt coating, reinforced with encapsulated asbestos fibers, used for surfacing asphalt, metal or composition roofs. Adheres to brick, concrete or steel. D-21 Asphalt, petroleum distillate, 4.24 aluminum metal, amorphous silica, and encapsulated asbestos Can be used to coat barns, industrial buildings, built-up asphalt roof systems, prepared roll roofing, weatherized galvanized iron and tin. Physical/chemical properties High-grade mastic compound with maximum water and weathering resistance. Soft material that gradually hardens to a coating that is pliable, durable and resistant to water. Cures to a tough, flexible, weather-resistant film. Resists damage from expansion and contraction caused by weather, temperature variation and normal building vibration, due to its ability to retain pliability. High-grade mastic compound with maximum water and weathering resistance. Soft material that gradually hardens to a coating that is pliable, durable and resistant to water. Cures to a tough, flexible, weather-resistant film. Resists damage resulting from expansion and contraction caused by weather, temperature variation and normal building vibration due to its ability to retain pliability. Forms a reflective, durable, flexible and waterproof surface. Produces a bright aluminum roof surface that reflects damaging ultraviolet sunlight. of 32 ft2), 3/4-in.- or 5/8-in.-thick plywood board. Figure 1 shows typical examples of the three roof types that were constructed. Repair activities were not measured in the field because it was impossible to determine the types of materials used in original roof construction and the composition of these materi als; therefore, representative roofs were built using nonasbestos-containing products, to allow for isolation of exposure to the roof coatings and cements alone. The asphalt shingle roofing entailed a composition asphalt shingle roofing system over a 3/4-in.-thick plywood board with a representative masonry chimney (common veneer brick laid using commercial adhe sive) and metal flashing (noncorrosive, nonstaining, 26-gauge galvanized metal). CertainTeed Landmark 30 fiberglass-based asphalt shingles were laid on a `breather-type' underlayment (Fontana VulcaSeal 30), per the manufacturer's specifications. Products C-8 and C-30 were applied to the asphalt shingle roofs. The BUR entailed a base plus three-ply, smooth surfaced fiberglass built-up roof membrane over a 5/8-in.-thick plywood board. A layer of light weight, asphalt-coated, fiberglass base felt [Johns Manville (JM) PermaPly 28] was placed on the ply wood and covered by a two-component, solvent-free, elastomeric cold-application adhesive (JM MBR Bonding Adhesive), consisting of an asphalt base material and an activator. A primer (Henry 104 Quickdry) was used over the base ply, upon which a coated fiberglass mat ply sheet with a light film of asphalt (JM GlasPly IV) was placed to cover the cold-application adhesive layer. Products B-10 and 201 were applied to the BUR substrates. The composite roofing entailed a composition smooth cap sheet and base ply over a 5/8-in.-thick plywood board. A rosin sheathing paper (Ratan Red 3.3-lb rosin paper) was used to cover the plywood and was covered with a high-quality cold-process adhesive (Siplast PA 311 Adhesive). A modified bitumen base ply consisting of a lightweight random fibrous mat impregnated and coated with highquality styrene-butadiene-styrene modified bitumen (Siplast Paradiene 20) was applied to the adhesive. The cap sheet, a smooth-surface cap sheet impreg nated and coated with ceramic granules (Siplast Paradiene 30) was fastened on top. None of the materials used to construct the roof sub strates contained asbestos, with the exception of the product under study that was applied to the fully con structed roof panel. All roofs were constructed in late 454 F. Mowat, R. Weidling and P. Sheehan riod of 14 days, to accelerate the volatilization of sol vents and the curing of the product, consistent with ASTM (1981, 1994) standards. Full cure was deter mined based on successive weighing of the smaller samples. Product-coated roof panels were heated to 220 5F for 14 days. Eight thermocouples placed within the oven and connected to an external digital read-out ensured that temperature was maintained in the desired range for the full curing period. This cur ing process ensured that the test substrates were all similar in condition and representative of the product after several months of natural curing on roofs under ambient conditions. Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Fig. 1. Typical examples of the three constructed roof types: (A) composition asphalt shingle roofing (asphalt shingle), (B) modified bitumen roofing (composite roofing) and (C) BUR. June 2005 by professional roofers from Western Roof ing Service, based in San Francisco, CA. The profes sional roofers applied the plastic roof cement products with a trowel to the shingles, flashing and masonry, using the manufacturer's specifications for coverage (trowel product to a thickness of 3/8-1/2 in.). The fibered roof coatings were applied with a roller (two coats, with a 24- or 48-h drying period between coats). The roof substrates were allowed to dry outdoors for at least 1 week prior to being placed in the curing oven. Curing process Once constructed, the roof substrates were weighed and placed in a custom-built oven for a pe Testing protocol The purpose of this exposure testing was to collect representative air samples in the breathing zone of a worker during application, hand scraping and sanding of the five products on representative roof substrates, clothing and tools under controlled condi tions and to determine the total fiber concentration and asbestos fiber counts in those samples following OSHA and National Institute of Occupational Safety and Health (NIOSH) procedures. Six different activ ities were evaluated in the testing program: (i) appli cation, (ii) wet sanding, (iii) removal from laundered clothing, (iv) removal from soiled tools, (v) hand sanding and (vi) hand scraping. The first four activi ties used wet and uncured product, while the latter two involved cured products. Each test was con ducted for 30 min, either in triplicate (for C-8 Plastic Roof Cement and B-10 Fibered Roof Coating) or as single-exposure tests (201 Fibered Roof Coating, C-30 Plastic Roof Cement, and D-21 Fibered Aluminum Roof Coating). These tests are described in detail in Table 2. Abatement technicians from MARCOR Environmental performed all removal activities. Testing facility All exposure tests were carried out in two adjacent 12 x 16 x 8-ft testing chambers (clean rooms) that were constructed at the Forensic Analytical Special ties, Inc., facility in Hayward, CA, in a large indoor room with a roll-up garage door. The test chambers were constructed of high-quality plywood and were lined with polyethylene sheeting to prevent migra tion of fibers into the room from outside, to con tain any fibers released during testing and to allow for easier cleanup between exposure tests. A high efficiency particulate air (HEPA)-filtered negative air unit regulated at ~200-400 ft3 min-1 was used, with the exhaust channeled through an opening of 10-12 in. diameter. Use of the HEPA air unit resulted in a ventilation rate in the testing chambers of 6-18 air changes h-1 , depending on the replicate, which was thought to adequately represent outdoor Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Simulation tests to assess occupational exposure to airborne asbestos 455 Table 2. Test activities performed on roof products Activity Application Wet sanding Hand scraping Hand sanding Removal from laundered clothing (clothing test) Removal from soiled tools (tool test) Description of task This test was used to quantitatively estimate the release of asbestos fibers (if any) during the application of B-10 Fibered Roof Coating and C-8 Plastic Roof Cement to the surface of representative roof substrates using either roller brushes (coating) or a trowel (cement). This test was used to quantitatively estimate the release of asbestos fibers (if any) during the wet sanding of uncured products (B-10 Fibered Roof Coating and C-8 Plastic Roof Cement) from the surfaces of representative roof substrates using either coarse sandpaper or a wire brush. These tests were conducted after the application test, where coated roof substrates were allowed to dry in ambient air for 1 week. This test was used to quantitatively estimate the release of asbestos fibers (if any) during the hand scraping of all five cured products from the surface of representative roof substrates using a long-handled putty knife. This activity might be performed by roofers in removing this product during a small roof repair or emergency patch. The approximate surface area of roof coating that was removed was 10 ft2 for the plastic roof cement products and 32 ft2 for the fibered roof coatings. This test was used to quantitatively estimate the release of asbestos fibers (if any) during the sanding of all five cured products from the surfaces of representative roof substrates using 60-grit sandpaper. Hand sanding represents an aggressive manipulation of these products and a `worst-case' test of fiber release and potential exposure. The approximate surface area of roof coating that was removed was 10 ft2 for the plastic roof cement products and 32 ft2 for the fibered roof coatings. This test was used to quantitatively estimate the release of asbestos fibers (if any) during the hand scraping and picking of C-8 Plastic Roof Cement and B-10 Fibered Roof Coating from representative laundered clothing (that had been smeared or soiled with the product prior to laundering) using a plastic scraper or blunt butter knife. These activities were conducted to represent attempted removal of the product from workers' clothing following laundering. This test was used to quantitatively estimate the release of asbestos fibers (if any) during the hand scraping and picking of C-8 Plastic Roof Cement and B-10 Fibered Roof Coating from representative soiled tools (that had been dipped and coated with product and left to dry outdoors) using a flathead screwdriver. These activities were conducted to represent removal of the product from workers' tools coated with residual product from past application activities. conditions where roofers might conduct their work (indoor air exchanges are typically > 4 changes h_1, as described by the ASHRAE (2001). A custom-built holder was used to hold the product-coated roof panels during testing. This holder was composed of wooden two-by-fours and angled at a 4 : 12 pitch, to simulate a common roof angle and ensure that the worker performing the removal activities was in an appropriate body position for roofing work. A diagram of the testing room setup is shown in Fig. 2. Prior to testing, the roof surface was thor oughly cleaned with a HEPA vacuum to remove any extraneous dust and to ensure that results would be specific to the product tested. Test design/air sampling data During the testing of the five roofing products, air samples were collected to analyze for the presence of airborne asbestos fibers. The exposure tests were per formed in triplicate for C-8 Plastic Roof Cement and B-10 Fibered Roof Coating. Single-exposure tests were conducted for 201 Fibered Roof Coating, C-30 Plastic Roof Cement and D-21 Fibered Aluminum Roof Coating. During each replicate of each test, two personal samples were collected from the breathing zone of the worker performing the removal activity. Personal samples (rather than area samples) were collected because these were thought to best repre sent a worker's exposure. These samples were col lected on 0.8-pm (pore size) mixed cellulose ester 456 F. Mowat, R. Weidling and P. Sheehan Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 (MCE) filters with a diameter of 25 mm, housed in a carbon-impregnated polyethylene cassette, for phase-contrast microscopy (PCM) and transmission electron microscopy (TEM) analyses. Personal air samples for fiber analysis were collected at a rate of 15 l min-1 for 30 min, using electrically powered vacuum pumps with flexible tubing to attach the filter cassettes to the pump. Airflow rates for the sampling trains were calibrated before and after each sampling event. A relatively high airflow rate was selected to ensure that an adequately low detection limit was achieved. Background air samples were collected in the test ing chamber to provide information on ambient con centrations of asbestos fibers in the room air prior to each test and clearance air samples were collected in the testing chamber after each replicate. These sam ples were collected before and after each exposure test. Background and clearance samples were col lected using the perimeter pumps located in the test ing room and were collected on 0.8-pm (pore size) MCE filters, 25 mm in diameter. Samples were col lected for 2 h at a flow rate of 15 l min-1, correspond ing to an air volume > 1500 l. Samples collected during testing were analyzed according to standard NIOSH methods. The personal samples collected during sampling were analyzed for fiber concentrations by PCM using NIOSH Method 7400, and for asbestos concentrations by TEM using NIOSH Method 7402 (NIOSH, 1994a,b). Quality control samples encompassed blanks (10%), repli cates (10%) and duplicates (5%). Both replicates and duplicates were run on the PCM samples; only duplicates were run on the TEM samples. Forty grid openings were counted on both the initial and dupli cate filters, where the duplicate involved analysis of a new filter prepared by a different analyst from the analyst who prepared the initial filter. (For some sam ples, it is possible that a different grid was counted because two grids were prepared per sample.) The acceptable range of fiber counts was determined us ing a 95% confidence interval of the Poisson distribu tion of the fiber results. TWA concentrations For samples in which a detectable concentration of asbestos was measured, total fiber concentration ob tained by PCM was converted to a PCM-equivalent (PCM-E) asbestos concentration using the asbestosto-total fiber ratio by TEM. The PCM-E represents the fraction of PCM total fibers that is estimated to be asbestos fibers, as opposed to other nonasbestos fibers such as cotton or wood fibers. The PCM-E concentrations were then used to calculate 8-h TWA exposure concentrations in an attempt to estimate the airborne concentration during a typical workday. Eight-hour TWAs were calculated to (i) estimate po tential exposure of workers conducting hand scrap ing and sanding activities for 30-min durations during the 8-h workday and (ii) to allow for direct comparison to current and historical occupational exposure limits or guidelines. Given the fact that the activities tested in this exposure simulation were not likely to have been performed by workers using these products, particularly for the duration of time tested, 8-h TWAs were calculated only for 30 min of each activity in a given workday. The estimated 8-h TWA was calculated using the following equation: ^ (ci)(ti) 8 ' (1) where: n = the total number of activities evaluated in a par ticular scenario (for our calculations, one of the six activities was tested), ci = PCM-E asbestos concentration measured during activity i (in fibers per cubic centimeter) and ti = duration of activity i performed by a worker (in hours). When calculating the 8-h TWA, the sum of all ti must equal 8 h. In our calculations of 8-h TWAs, the activity was considered to have occurred for 30 min; for the remaining 7.5 h of the 8-h workday wherein no scraping or sanding activity was per formed, the airborne asbestos background PCM-E concentration estimated for this specific exposure simulation was used in the TWA calculation. This value is a nonzero PCM-E background concentration of 0.0002 f cc-1 , based on the background sample for this study, in which a single asbestos fiber was identified. RESULTS Total fibers During the simulated application, wet sanding, hand scraping, hand sanding, clothing and tool tests, total airborne fiber concentrations measured by PCM (all fibers, including nonasbestos) ranged from <0.005 to 0.032 f cc-1 (Table 3). The detection lim its in this study ranged from 0.003 to 0.006 f cc-1 . Total fibers (which are composed of nonasbestos and asbestos fibers) were detected in 38 of the 84 per sonal samples (Table 3). PCM total fiber counts for personal samples ranged from 0 to 29.5 fibers. How ever, only a relatively small fraction of these were as bestos fibers, as evidenced by the TEM asbestos fiber counts (see below). The remaining nonasbestos fibers likely represented cotton fibers from the workers' clothing, wood fibers present in the test room or other fibers present in air. The raw PCM total fiber Simulation tests to assess occupational exposure to airborne asbestos 457 Table 3. Analysis of samples using PCM (NIOSH Method 7400) Test activity Product Number of samples Number of detects PCM total fiber concentrations (f cc ) Arithmetic mean Standard deviation Application C-8 6 2 0.007 B-10 6 1 0.006 Wet sanding C-8 6 0 0.006 B-10 6 3 0.01 Hand scraping C-8 6 5 0.008 B-10 6 2 0.009 C-30 2 2 0.008 201 2 2 0.02 D-21 2 2 0.01 Hand sanding C-8 6 3 0.007 B-10 6 1 0.008 C-30 2 0 0.006 201 2 2 0.02 D-21 2 1 0.008 Clothing C-8 6 6 0.01 B-10 6 3 0.01 Tools C-8 6 1 0.006 B-10 6 2 0.006 Total 84 38 0.002 Not applicable Not applicable 0.01 0.002 0.004 0.0007 0.003 0.004 0.002 0.004 Not applicable 0.01 0.003 0.003 0.006 0.0006 0.0004 Minimum ND (<0.006) ND (<0.006) ND (<0.006) ND (<0.006) ND (<0.006) ND (<0.006) 0.007 0.016 0.009 ND (<0.006) ND (<0.006) ND (<0.006) 0.012 0.006 0.008 ND (<0.006) ND (<0.005) ND (<0.006) ND, not detected. Maximum 0.009 0.006 ND (<0.007) 0.032 0.011 0.015 0.008 0.02 0.015 0.01 0.015 ND (<0.006) 0.027 0.01 0.017 0.019 0.007 0.007 Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 concentrations were adjusted based on TEM analyses to derive a PCM-E asbestos concentration that would enable calculation of an 8-h TWA. Asbestos fibers No asbestos fibers were found in any of the tool ex posure tests or during the application or wet sanding tests. During the simulated hand scraping, sanding, clothing and tool tests, airborne asbestos fiber concentrations measured by TEM ranged from <0.0021 to 0.056 f cc-1 . Asbestos fibers were de tected in 28 of 84 personal samples, and the number of asbestos fibers found ranged from 0 to 26 (Table 4). All detected asbestos fibers were identified as chrysotile fibers. The samples that contained > 10 asbestos fibers were collected during the hand sanding test on cured material. Representative micrographs of the samples used to determine asbestos counts in the hand scraping and hand sanding of cured product (tasks that produced the highest counts) are provided in Fig. 3. Overall, most of the samples showed fibers attached to `chunks' of matrix (Fig. 3A,C), while a few others showed free, curly fibers (Fig. 3B). In general, the hand scraping test gener ated larger pieces of matrix compared to sanding; in the sanding test, the observed length of fibers was typically shorter than that for the scraping test. For the scraping tests, typical fiber lengths ranged from 15 to 40 pm, with widths of 1-3 pm. Fiber lengths for the sanding tests were 7-18 pm, with all widths > 1 pm. Background, concentrations For all replicates, total fiber concentrations mea sured by PCM in background/clearance samples ranged from below the detection limit of <0.002 to 0.009 f cc-1 (Table 5). In all six replicates for appli cation, total fiber concentrations ranged from 0.003 to 0.007 f cc-1 . For wet sanding, total fiber concen trations in the six replicates ranged from 0.002 to 0.007 f cc-1 . In all nine replicates for hand scraping, total fiber concentrations ranged from 0.002 to 0.009 f cc-1 . For hand sanding, total fiber concentrations in the nine replicates ranged from 0.002 to 0.006 f cc-1 . For the clothing test, total fiber concentrations for the six replicates ranged from below the detection limit (<0.002 f cc-1) to 0.005 f cc-1 . Finally, for the tool test, total fiber concentrations for the six replicates ranged from 0.003 to 0.007 f cc-1 . For the PCM fiber counts, 1-22 fibers (total fibers, including nonas bestos) were identified in the background/clearance samples. As part of the exposure simulation, 49 background/ clearance samples were collected. Forty-two of these samples were obtained between the periods of time when the exposure simulation experiments were being performed (e.g. immediately before and after each replicate test), and the remaining background samples were taken prior to conducting the exposure tests. Of the 42 exposure-related samples, two back ground samples were identified by TEM analysis as containing a single asbestos fiber. TEM asbestos fiber concentrations were below the analytical sensitivity 458 F. Mowat, R. Weidling and P. Sheehan Table 4. TEM asbestos fiber counts (NIOSH Method 7402) and PCM-E Test activity Application Wet sanding Hand scraping Hand Sanding Clothing Tools Total Product C-8 B-10 C-8 B-10 C-8 B-10 C-30 201 D-21 C-8 B-10 C-30 201 D-21 C-8 B-10 C-8 B-10 Number of samples 6 6 6 6 6 6a 2 2 2 6 6b 2 2 2 6 6 6 6 84 Number of samples with detected asbestos fiber counts 0 0 0 0 2 2 1 2 2 5 3 2 2 2 3 2 0 0 28 TEM asbestos fiber counts ND ND ND ND 1, 2 1, 2 1 8, 9 2, 4 3, 3, 6, 12, 13 1, 2, 10 12, 15 24, 26 13, 21 1, 1, 4 2, 5 ND ND F/f Minimum NA NA NA NA 0 0 0 1 0.57 0 0 0.8 1 0.91 0 0 NA NA Maximum NA NA NA NA 0.5 1 0.33 1 0.67 0.86 0.67 0.83 1 0.93 0.67 0.5 NA NA PCM-E (f cc- 1) Arithmetic Minimum mean NA NA NA NA 0.0014 0.0014 0.0012 0.018 0.0073 0.0040 0.0038 0.0049 0.020 0.0074 0.0024 0.0010 NA NA NA NA NA NA 0 0 0 0.02 0.006 0 0 0.0048 0.012 0.0056 0 0 NA NA Maximum NA NA NA NA 0.0055 0.006 0.0023 0.016 0.0086 0.0076 0.006 0.005 0.027 0.0091 0.011 0.0035 NA NA No asbestos fibers were found in any tool test, so PCM-E and TWAs were not calculated. F/f, percentage of asbestos fibers to total fibers (by TEM); NA, not applicable; ND, not detected; PCM-E, product of F/f and PCM fiber concentration. aOne sample overloaded; data not included in analysis. bTwo samples overloaded; data not included in analysis. Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 (0.0005-0.0009 f cc-1 , depending on replicate) for all but these two of the 42 samples, where the TEM asbestos fiber concentration for both was at the limit of analytical sensitivity of 0.0008 f cc-1 (Table 5). One asbestos fiber was detected in the background sample for the first sanding test of B-10 Fibered Roof Coating and the background sample for the second sanding test of product 201 Fibered Roof Coating. Zero asbestos fibers were found in all other background samples. Based on the fact that ~95% of the samples had zero asbestos fibers detected, the sample with the lower asbestos concen tration was used to estimate PCM-E background con centrations for the 8-h TWA calculations. This value is a nonzero PCM-E background concentration of 0.0002 f cc-1. Quality assurance Due to the low fiber count in this study, there is a wide variability associated with the acceptable range of fiber counts. For all but one sample, results of the duplicate analysis indicated a `pass'. For that sample, the initial sample contained six fibers; the duplicate sample contained one fiber. The acceptable range was 2-13 fibers. The most likely explanation for the variability in fiber count between the initial and duplicate samples was the heterogeneous distri bution of particulate matter on the filters. Due to the nature of the roofing material and the activities evaluated, the filter contained a mixture of fine dust, fibers and large, black chunks of matrix, some of which also may have contained fibers. These matrix chunks resulted in uneven concentrations of fibers in certain locations on the filter and contributed to the difference in fiber count between the initial and duplicate sample. To ensure that the initial fiber count was accurate, the second analyst recounted the fibers on the first sample. This recounting con firmed the finding of six fibers in the initial sample; thus, the finding of a single fiber in the duplicate sam ple is likely due solely to the heterogeneity of partic ulate matter on the filter and differences in the grid openings counted by the analysts. TWA concentrations Many of the results indicated asbestos concentra tions at or below the detection limit; therefore, 8-h TWAs were not calculated for these data. The esti mated 8-h TWAs ranged from 0.0003 to 0.002 f cc-1 based on 30 min of activity in an 8-h workday for three of the four activities evaluated in this expo sure simulation; 8-h TWAs were not calculated for the application, wet sanding or tool tests because no asbestos fibers were identified in any of these Simulation tests to assess occupational exposure to airborne asbestos 459 samples (Fig. 4). The 8-h TWA estimates are compa rable to the PCM-E background concentration of 0.0002 f cc-1 measured in this study. In addition, airborne asbestos TWA concentrations in this study were similar to ambient outdoor air concentra tions reported in the literature (e.g. Selikoff and Hammond, 1968; Selikoff et al., 1972; Nicholson and Pundsack, 1973; Chesson et al., 1985; Bignon, 1989; Corn, 1994). The reported concentrations range from nondetect to ~0.0003 f cc-1 in rural air and to 0.003 f cc-1 in urban air. Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Fig. 3. Example photomicrographs of debris from scraping and sanding activities. DISCUSSION The abundant natural supply, low cost, fireproof nature, chemical inertness, ease of mixing and rein forcement properties of asbestos led to its use in a va riety of applications, including the roofing industry (Brzozowski, 1989). With regard to mastics and coat ings specifically, asbestos allows the mixture to re main in a consistent state; therefore, if the mastic is applied to a vertical or sloping surface such as a roof, the prevention of running or sagging due to the con sistency is a beneficial property provided by the as bestos content (Hodgson, 1985; Brzozowski, 1989). In addition to controlling viscosity, asbestos provides mechanical properties once the product has cured and improves weathering effects (Brzozowski, 1989). OSHA regulates asbestos-containing cements, coatings and mastics and flashings as `intact inciden tal materials', which are less stringently regulated than other types of asbestos-containing materials due to their nonfriable nature and because the asbes tos is encapsulated in matrix material that prevents the release of airborne fibers (Good, 1997). This is particularly evident for asbestos-containing roof coatings, mastics and cements, which even today, are sold legally in some hardware stores. This simulation provides the first data specific to hand-applied roof coatings, mastics and cement products. The products tested were cured to simulate the presence of these products on a roof for several months, after volatilization of solvents and other chemicals. Although these products were fully cured, they remained pliable, per the manufacturer's design, particularly on hot days. The present study did not evaluate the effect of long-term weathering on the roofing products (as would be expected on a roof that is typically in service for >30 years); however, stud ies to evaluate the potential for fiber release from roofs that have been subjected to weathering condi tions (e.g. sun and rain) for long periods of time will be undertaken. It is possible that weathering of the roofing material may result in the matrix drying out and perhaps becoming more brittle. An Australian study of asbestos cement roof `cladding' materials in 40-year-old buildings indicated the potential for 460 F. Mowat, R. Weidling and P. Sheehan Table 5. Background/clearance samples by PCM and TEM during test activities Test activity Application Wet sanding Hand scraping Hand sanding Clothing Tools Total Product C-8 B-10 C-8 B-10 C-8 B-10 C-30 201 D-21 C-8 B-10 C-30 201 D-21 C-8 B-10 C-8 B-10 Number of samples 3 3 3 3 3 3 1 1 1 3 3 1 1 1 3 3 3 3 42 PCM total fiber concentrations (f cc x) Number of detects Minimum Maximum 3 0.003 0.004 3 0.003 0.007 3 0.002 0.005 3 0.003 0.007 3 0.005 0.007 3 0.002 0.008 1 0.008 0.008 1 0.009 0.009 1 0.003 0.003 3 0.004 0.005 3 0.003 0.006 1 0.003 0.003 1 0.002 0.002 1 0.004 0.004 2 ND (<0.002) 0.005 3 0.003 0.005 3 0.003 0.006 3 0.003 0.007 41 ND, all values were not detected. TEM fiber concentrations (f cc x) Number of Minimum detects Maximum 0 ND (<0.0007) ND (<0.0008) 0 ND (<0.0007) ND (<0.0008) 0 ND (<0.0007) ND (<0.0007) 0 ND (<0.0007) ND (<0.0008) 0 ND (<0.0007) ND (<0.0009) 0 ND (<0.0005) ND (<0.0008) 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0009) ND (<0.0009) 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0007) ND (<0.0008) 1 ND (<0.0008) 0.0008 0 ND (<0.0009) ND (<0.0009) 1 0.0008 0.0008 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0008) ND (<0.0008) 0 ND (<0.0007) ND (<0.008) 2 Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 Fig. 4. Estimated 8-h TWA concentrations. a thin layer, or `fleece', of asbestos fibers to become exposed as the roofing surfaces weather (Brown, 1987). Exposures of <0.1 to 0.22 f cc_1 were mea sured when the roof surface was painted or cleaned with a water jet for several hours. During roof re placement (including unfastening, removal and dis posal of roof sheets), concentrations ranged from below the limit of detection to 0.32 f cc_1, with esti mated TWAs ranging from 0.03 to 0.21 f cc_1. Once long-term weathering is completed on representative roof substrates, these results will be reported. The results of this exposure simulation indicate that workers involved in application, wet sanding, hand scraping and hand sanding of these cured prod ucts from roof panels, laundered clothing and soiled tools for an extended period of time would not have been exposed to asbestos concentrations above the current applicable occupational health standards. Al though a relatively high air exchange rate was used, possibly resulting in dilution of the resulting airborne asbestos fiber concentration, this airflow was compa rable to what roofers would experience outdoors, where wind and other factors affect exposure. Some of the activities tested (e.g. hand scraping, hand sand ing) would not be typical, or even expected, practice in the use of this product, based on the manufac turer's specifications; however, they may have been conducted for quick, small-scale repairs. It is un likely that hand scraping or sanding would ever be conducted continuously for the 30-min time period tested in this study, but these results indicate poten tial exposure for worst-case scenarios of fiber release for these small-scale activities. Per the manufacturer's specifications, the plastic roof cement products were used to repair leaks and as a patching compound for small repairs. These products were single-coat products; as such, long term sanding or scraping of the C-8 and C-30 products between coats was unlikely, although small-scale sanding may have been done to `scuff' existing product. For the fibered roof coatings, the products Simulation tests to assess occupational exposure to airborne asbestos 461 Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021 were generally applied to an entire roof. A roof in need of repair likely would be removed in large sec tions or in its entirety; therefore, the coating typically would not be scraped or sanded in the process of re moving it. These products were two-coat products that did not require sanding between coats. The NRCA states that BURs are usually removed with power roof cutters that cut through the felts and un derlying insulation to separate the roof into smaller sections that are more easily handled (NRCA, 1994). These sections are then pulled up using prying tools, `power roof cutters' or `power tear-off ma chines'. In effecting small repairs, the NRCA states that BURs could be removed using manual tools and methods, such as chopping or cutting through the plies with axes, hatchets or utility knives. They note that roof coatings, mastics and cements are re moved by hand methods, usually together with the components of the roof (e.g. plies, flashing, shingles etc.) to which they are adhered (NRCA, 1994). How ever, assuming an exposure scenario in which a worker manipulates these products in a manner that was evaluated in this study, airborne asbestos con centrations would not exceed current or historical occupational exposure limits. More aggressive cutting, tear-out and sawing ac tivities were not conducted because it would be diffi cult to conduct these activities for a full 30 min (to ensure adequate sampling) on a 4 x 8-ft roof sub strate. Exposure resulting from removal activities in the field was not measured because it was not possi ble to determine the composition and type of under lying products (e.g. shingles, plies etc.) that were used in combination with the cements and coatings; as such, we could not isolate potential asbestos expo sures to the products of interest. Although this inves tigation did not evaluate roof tear-out activities, the NRCA reports that removal of BURs with power tools produces exposures well below 0.1 f cc-1 (NRCA, 1994). These included projects in which 63 personal breathing-zone samples were collected where there were no dust suppression or collector controls (e.g. wetting the roof, misting the work area and HEPA vacuuming or other dust removal activity). Although the majority of the samples were of short duration (usually <120-180 min), all samples were below the current OSHA PEL. The NRCA states that, although dust is generated during removal, most of the dust was likely nonasbestiform material such as mineral or glass fibers found in the insulation under lying the BURs. This was confirmed in several proj ects where TEM analysis indicated no asbestos fibers present during roof removal (NRCA, 1994). TWAs are generally the best indicator of worker exposure based on the nature and duration of the worker's activity over a typical 8-h day. In this study, 8-h TWAs were calculated assuming 30 min of activ ity in a given workday. For the five roofing products evaluated in the present study, calculated 8-h TWAs ranged from 0.0003 to 0.002 f cc-1 (based on 30 min of activity), all of which are well below the cur rent OSHA PEL of 0.1 f cc-1 . Specifically, 8-h TWA results were 100-fold below the current OSHA PEL of 0.1 f cc-1, which has been in place since 1994, and 100- to 1000-fold below historical exposure lim its in the workplace. In addition, many of the TEM asbestos concentrations and 8-h TWAs were close to the background concentrations of asbestos present in ambient air at the test site and reported historically in ambient air in US cities (e.g. Selikoff and Hammond 1968; Selikoff et al., 1972; Nicholson and Pundsack, 1973; Chesson et al., 1985; Bignon, 1989; Corn, 1994). In conclusion, this simulation study indicates a low probability that asbestos will become airborne during manipulation of coating- and cement-type roofing products such as those tested. The results indicate an exceedingly small asbestos exposure associated with the repair of these products, even during the un likely 30-min application, hand scraping and sanding of these products from roof substrates. FU N D IN G Monsey Products Company and Henry Company. Acknowledgements--Two of the authors (F.M. and P.S.) have been designated as experts in litigation involving these prod ucts. We appreciate the material sciences work performed by Chris Scott who aided in reformulating the five products; Rachel De Guzman, architect, for aiding in design of the roof structures; Western Roofing for their expertise in constructing and coating the roof substrates; expertise in industrial hygiene and sampling design from Jeff Hicks; testing assistance from Riley Sheehan; and data analysis assistance from Kathy Dyson. We also acknowledge Forensic Analytical Specialties, Inc., particularly Gustavo Delgado, Jim Flores, Mark Floyd and Brian Ramsell, for providing the indoor test chamber facility and industrial hygiene oversight and for sample analysis. We gratefully thank MARCOR Environmental for constructing the clean rooms and conducting the product removal activities. REFERENCES Anderson PH, Grant MA, Mclnnes RG et al. (1982) Analysis of fiber release from certain asbestos products. Draft final report. Prepared for the U.S. Environmental Protection Agency, Office of Pesticides and Toxic Substances Chemical Control Division, December, Contract 68-01-5960. ASHRAE. (2001) Ventilation for acceptable indoor air quality. American National Standards Institute (ANSI)/ASHRAE Standard 62-2001. Atlanta, GA: American Society of Heat ing, Refrigerating and Air-Conditioning. ASTM. (1981) Standard test methods for mastics and coatings used with thermal insulation. ASTM Method 461-81. West Conshohocken, PA: American Society for Testing and Materials. ASTM. (1994) Standard practice for making and curing test specimens of mastic thermal insulation coatings. ASTM Method 419-94. West Conshohocken, PA: American Society for Testing and Materials. 462 F. Mowat, R. Weidling and P. Sheehan Brown SK. (1987) Asbestos exposure during renovation and demolition of asbestos-cement clad buildings. Am Ind Hyg Assoc; 48: 478-86. Bignon J. (1989) Mineral fibres in the non-occupational envi ronment. In: Bignon J, et al. Non-occupational exposure to mineral fibres. IARC Scientific Publication No. 90. Lyon, France: International Agency for Research on Cancer, pp. 3-29. Brzozowski KJ. (1989) Trouble overhead: the problem of asbes tos on roofs. The Construction Specifier, November 55-59. Chesson J, Margeson DP, Ogden J et al. (1985) Evaluation of asbestos abatement techniques. Phase 1: removal. EPA560/5-85-109. Washington, D.C.: U.S. Environmental Pro tection Agency. Corn M. (1994) Airborne concentrations of asbestos in non occupational environments. Ann Occup Hyg; 38: 495-502. Good C. (1997) No evidence, no regulation. Professional Roofing; November 22-4. Hodgson AA. (1985) Alternative to asbestos and asbestos products. Berkshire, UK: Anjalena Publications. pp. 165-6. ICF Incorporated. (1988) Asbestos exposure assessment. Prepared for Dr. Kin Wong, Chemical Engineering Branch, U.S. Environmental Protection Agency, Office of Pesticides and Toxic Substances; March 21, EPA-OTS Report No. 000527589. Nicholson WJ, Pundsack FL. (1973) Asbestos in the environ ment. In: Bogovski P, Timbrell V, Gilson JC, Wagner JC, editors. Biological effects of asbestos. Proceedings of a working conference held at International Agency for Research on Cancer, Vol. 8. IARC Scientific Publication. pp. 126-32. NIOSH. (1994a) Asbestos and other fibers by PCM. NIOSH method 7400, manual of analytical methods. 4th edn. Atlanta, GA: National Institute of Occupational Safety and Health. NIOSH. (1994b) Asbestos by TEM. NIOSH method 7402, NIOSH manual of analytical methods. 4th edn. Atlanta, GA: National Institute of Occupational Safety and Health. NOHSC. (2002) Handling of asbestos-cement (fibro) prod ucts. Chapter 9. In: Code of practice for the safe removal of asbestos. Available at http://www.eastfremantle.wa.gov. au/html/pdf/AsbestosRemoval.pdf. Accessed 24 February 2006. NRCA. (1994) ``Objective data'' demonstration for certain roofing materials and operations under OSHA's 1994 asbes tos standard, December 14. Rosemont, IL: National Roofers Contractors Association. Selikoff IJ, Hammond EC, III. (1968) Community effects of nonoccupational environmental asbestos exposure. Am J Public Health; 58: 1658-66. Selikoff IJ, Nicholson WJ, Langer AM. (1972) Asbestos air pollution. Arch Environ Health; 25: 1-13. Stern FB, Ruder AM, Chen G. (2000) Proportionate mortality among unionized roofers and waterproofers. Am J Ind Med; 37: 478-92. Downloaded from https://academic.oup.eom/annweh/article/51/5/451/201118 by guest on 06 May 2021