Document 6wqdJ9XD2nXJ9aGqVn3v3XzQ6

ENVIRONMENTAL RESEARCH 31, 32-53 (1983) Characterization of Three Types of Chrysotile Asbestos after Aerosolization Kent E. Pinkerton,*'1 Arnold R. Brody,t Daniel A. McLaurin,$ Bernard Adkins, Jr, Robert W. O'Connor, Philip C. Pratt,* and James D. Crapo* Departments of Pathology and Medicine. Duke University and Durham Veterans Administration Medical Center. Durham, North Carolina 27705; tLaboratory of Pulmonary Function and Toxicology, National Institute of Environmental Health Sciences. iBecton --Dickinson Research Center, and Northrop Services Inc., Research Triangle Park, North Carolina 27711 Received December 22, 1981 Jeffrey Mine and Coalinga Mine chrysotile, two asbestos samples prepared for experi mental research by the National Institute of Environmental Health Sciences, and the UICC B chrysotile reference sample have been characterized in the aerosolized state using gravimetric measurements, light microscopy, scanning electron microscopy, and x-ray energy spectrometry. These methods revealed (1) a greater "respirable" mass fraction in the Jeffrey and UICC B preparations compared to the Coalinga sample, (2) for fibers greater than 5 /am in length and less than 3 /am in diameter, Jeffrey Mine chrysotile contained a significantly greater fraction of fibers longer than 40 /xm in length compared to the UICC B or Coalinga Mine chrysotiles, and (3) Jeffrey and UICC B chrysotile contained no fibers or fiber clusters which exceeded 2 /am in diameter while Coalinga chrysotile contained numer ous fibers and fiber clusters which were greater than 2 /am in diameter. The characterization of these chrysotile preparations in the aerosolized state, in particular the Coalinga Mine chrysotile, demonstrated different fiber length and fiber width distributions when compared with previous characterizations of samples that had been dispersed in a liquid medium by ultrasonification. These observations emphasize the importance of determining the size distribution of fibers in the aerosolized state for inhalation studies and the size distribution of fibers in a liquid suspension for oral ingestion, instillation, or injection studies. Because of differences in length-width distributions, each of the studied chrysotile preparations would be expected to have different patterns of deposition in the alveolar regions of the lung after an inhalation exposure. INTRODUCTION There is increasing evidence that asbestos-induced pulmonary fibrosis neoplasia are due to the physical and chemical characteristics of the inhaled fib (Wagner, 1965: Seaton, 1975: Stanton and Layard, 1978; Pott, 1978). This mak important to use well-characterized fiber preparations in experimental research more clearly identify those factors leading to lung injury. The purpose of paper is to describe two new samples of chrysotile which have been prepared experimental research by the National Institute of Environmental Health Scien (NIEHS). A third chrysotile. UICC B. has also been characterized in this pap ! To whom correspondence and reprint requests should be addressed: Box 3177, Duke Unive Medical Center. Durham. N.C. 27710. 0013-9351 83 53.00 Copvricihi * M Academic Press. Inc \U rights of reproduction in an\ form reserved 32 CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 33 jrsotile represents over 90% of the world's asbestos production and is ubiq- j|in our environment. The properties of fire retardance, chemical inertness, istrength, and flexibility make chrysotile important in numerous applica- ile Asbestos lich as insulation, ceiling tiles, brake linings, and cement products. These ions of asbestos enhance the frequency of inhalation of chrysotile by ifinot directly involved with the mining or processing of this material, K. McLaurin,? ip C. Pratt. tjtlle is a member of the serpentine family of minerals, whereas all other js minerals belong to the amphibole family (Sinclair, 1959). Mechanical , of chrysotile can lead to disruption of the fiber at weak points along the tis disruption can cause the fiber to "open up" into its fibrillar subunits, `urhain Veterans Laboratory of omental Health Services' Inc.. ; new fibers of smaller diameter and/or shorter length, fibers with splayed ^fibers with an uneven diameter along the length, or combinations of the ^'(Assuncao and Corn, 1975). These features make chrysotile more complex Hjlyze in terms of fiber size and number than the amphibole types of asbestos __ fiber-s are basically straight and uniform in diameter. characterization of a chrysotile preparation in terms of particle and fiber nrepared for experinces, and the UICC >solized state using roscopy, and x-ray mass fraction in the 2) for fibers greater uysotile contained a ->ared to the UICC B mtained no fibers or lie contained numer- istribution is influenced by a number of factors. These include (l) the state in |)the chrysotile is found, i.e., in bulk, in suspension, or in an aerosol, (2) the |jjd of sample collection, i.e., on a slide or filter, (3) the manner in which the |fed sample is prepared for examination, i.e., by ultrasonification, transfer or fjijjreparation of the filter or slide, (4) the instrument used to measure the s"'f' and fibers in the sample, i.e., the optical microscope, transmission elec troscope, or scanning electron microscope, and (5) the criteria used for jvfibers, i.e., a minimum fiber length, a 3 to 1 aspect ratio, or characteristics ||\morphology. Ihc characterization - the Coalinga Mine ions when compared a liquid medium bv letermining the size ic size distribution of studies. Because of : preparations would ions of the lung after ^cterization of chrysotile in the aerosolized state has been done for the i|A and B reference samples (Timbrell, 1970a; Beckett, 1973). Two new Pi samples of chrysotile have been prepared in bulk (approximately 1000 of each) and are available for experimental research through the National jjjte of Environmental Health Sciences, Research Triangle Park, North These new preparations (Jeffrey and Coalinga chrysotile) have been Cterized for elemental composition, mineral composition, particle surface |$nd density, and thermal properties using optical emission spectrography, (/diffraction, thermogravimetry, pycnometry, and a number of petrographic $iscopic techniques (Campbell et al., 1980). Particle size analysis has been nonary fibrosis an( samples dispersed in a liquid medium in which measurements of fiber - of the inhaled fibei i'and diameter were made (Wylie, 1979; Campbell et al., 1980; Siegrist and . 1978). This makes il M'i 1980). Particle size analysis of these preparations in the aerosolized state rrimental research ti pnbt done. The purpose of this study is to characterize in the aerosolized state The purpose of thiij fe'ilfcffrey and Coalinga chrysotile preparations along with the UICC B reference ve been prepared foi I- $Eunple. To characterize each preparation, gravimetric dust measurements, optical ntal Health Sciences] jfflefoscopy, scanning electron microscopy (SEM), and x-ray energy spectrometry erized in this paper. "WjCTe used. Information obtained using these techniques has helped identify char acteristics of these asbestos preparations which may influence their deposition >\ M . Duke L'ni\.-rsityi% l$ttem when inhaled as an aerosol and which could potentially contribute to their y to cause pulmonary injury. 34 PINKERTON ET AL.. MATERIALS AMD METHODS Fiber Preparations Chrysotile samples were obtained from the following locations: the Coal Mine in California (Union Carbide), the Jeffrey Mine in Quebec, Cat (Johns-Manville), and the Canadian reference sample prepared by the Inte tional Union Against Cancer (UICC B). A brief history for each preparatic given. Coalinga Mine fiber. Identified as COF-25, this unique form of chrysoti obtained from the New Idria serpentinite mass located in the Diablo rang California. The deposit is unusual in that the fibers are randomly oriented mat, rather than as parallel fibers running in veins, and the mining process is < with bulldozers. The deposit is almost pure chrysotile. The fibers are shoi length and are not of spinning grade. Because of the absence of long fibers in) chrysotile, a great deal of interest has been generated in using this chrysotile "short-range fiber" preparation. Preparation of this short fiber material for experimental research was doni the following manner. The ore from the mine was first screened to remove c taminating rocks. From this point the material was processed in water as as| of I"7c solids and 99% water. The slurry was passed through a grinder as magnetic separator three times to open the fiber bundles iron-containing minerals. Between each grinding the slurry was passed throi); particle size separator (hydroclone) under pressure. The aqueous slurry waj tated inside the hydroclone forming a vortex. Heavy particles escape froiri hydroclone through a side port located near the bottom of the hydroclone. 1 coarse particles were reground and put through the hydroclone again. Theij particles leave the hydroclone through a side port located near the top o] vortex. These particles were fed into a series of hydroclones in which the top ports (overflow port) are progressively smaller to allow for the collection of J and finer chrysotile fibers. The final hydroclone port had an external diameu 25 mm from which the chrysotile preparation was collected. (For industrial) poses the final hydroclone port is usually six inches (personal communical Asbestos Group. Union Carbide. Niagara Falls. N.Y.).) 1 Langer et al. (1978) described in detail a chrysotile sample also obtained! the Coalinga Mine deposit, referred to as Calidria RG-144. The differences tween COF-25 and RG-144 are a result of the differences in the processing o) raw material. COF-25 has a finer particle size than RG-144. This finer particl| is a result of differences in the pressure, vortex characteristics, and overflow): diameters used in the hvdroclone. COF-25 is not derived from RG-144 by fu| grinding or pellet milling. j Jeffrey Mine fiber. Identified as Plastibest-20. this form of chrysotile is a grf! asbestos used by the plastics industry. The fibers run in parallel bundles, orieji crosswise in veins in serpentine rock and for this project were purified by M milling and air separation using standard industrial techniques. The final I preparation by volume is greater than 96% chrysotile (Campbell et If Preparation of the material for experimental research was accomplished by| Mng the material through a hurricane pulverizer three times to open fiber bund! CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 35 %CC B chrysotile reference sample. This preparation is a grade 4 chrysotile K^'^j|g|brell et al., 1968) obtained from eight different Canadian chrysotile mines and according to the proportional production of each mine during the year of >: the Coalingaj jJ/ The approval of this preparation was made in 1966 by the Union Inter- iebec, Canada! ptflMpnale Contre Cancer (UICC) to standardize asbestos samples used in experi- by the Interna- research. The literature is replete with information regarding the physical i preparation is biochemical makeup of this chrysotile preparation (Timbrell et al., 1968; Tim- 1970a, b; Rendall, 1970, 1980; Morgan and Cralley, 1973; Beckett, 1973). We <?f chrysotile is reexamined UICC B chrysotile in the aerosolized state with a twofold pur Diablo range in " , Je: (1) to compare our results for the UICC B fiber size distribution with those y. oriented as a l|H:i|lmd in prior studies in the literature, and (2) to correlate the size distribution of process is done P^^Coalinga and Jeffrey Mine fibers to the UICC chrysotile fiber preparations ;rs are short in liteffijjyjie for experimental inhalation research. No manipulation of these chrysotile ng fibers in thisi i chrysotile as ai iterations was done prior to aerosolizing the fibers. Aerosollzation ph was done inKj^^imodified Timbrell generator (Timbrell, 1968) was used to create a dust cloud feremove con'/dler as a slurry ((grinder and a tajlremove any ||sed through a sMiirry was roii||pe from the! jplone. ThesejM nn. The light" Tjtfte top of thel a 5-m3.stainless steel-exposure chamber. Each asbestos preparation was [(compressed with a plunger in a 2.8 x 9.0-cm cylinder to form a plug. This ^mechanically advanced into the pathway of a blade rotating 1500 rpm to > an aerosol of fibers within the dispersing bowl of the generator. A copper nnecting the dispersing bowl to the exposure chamber facilitated the pas- ij^f aerosolized fibers into the exposure chamber. The concentration of the |fj|ps dust cloud was regulated by adjustment of the airflow through the expowhamber. This resulted in the flow of air being maintained between 200 to 400 "i!|> er m-inute. the top exit' 'j^Dust Mass Concentration jjpction of finerj fK diameter of] gmdustrial purfjpftimunication, ^ce the exposure chamber had stabilized (normally after 1 hr of dust genera- ,a gravimetric measurement of the dust concentration within the exposure iber was made. This was accomplished by sampling 100 liters of chamber air ffljPpn through a 0.8-/u,m Nucleopore filter housed within a Gelman filter holder. lUllfe'sample time was 10 min at a flowrate of 10.0 liters per minute. S||btained from (jufferences be- m^wirable Mass Concentration jessing of the mPp|m||Me respirable mass concentration in the exposure chamber for each chrysotile ..Jf-.particle size |(||p|gWation was measured using two different instruments: a Casella sampler and tijGverflow port ||^^|scade impactor. The term "respirable mass concentration" is arbitrary in by further !||M:it was decided by totally different parameters for each apparatus used. Using '('"-m&Casella sampler, a gravimetric estimate of the respiratory mass concentration sdfile is a grade 4 g| ws made by collecting fibers and particles with an equivalent aerodynamic diam>pn(dles, oriented^' ^pr of 7.1 p.m or less on a glass fiber filter. Fibers and particles larger than this mrjfied by roller flilivalent diameter were prevented from depositing on the filter by a multichan- |Jhe final fiber pi 06l horizontal elutriator. Each sample was collected over a 6-hr period at a flow 1 et al.. 19801. rate of 2.5 liters per minute with the Casella sampler placed inside the exposure |ipiished by pas-1 chamber. r< fiber bundles. Using the Cascade impactor. the respirable mass concentration was defined to 36 PINKERTON ET AL. include filters on which the majority (>50%) of fibers collected were less thatf in length. This was determined by examination of the filter by optical mici copy. This respirable mass was obtained by using precutter stages in the Casci impactor to eliminate the longer and thicker fibers and by adjusting the flowrats obtain a fiber size distribution (cut point) of 10 jam or less in length. The volume of air sampled through the Cascade impactor was 200 liters. A t( sample mass was also collected on the same day by sampling 0.2 m3 of chamber at a flowrate of 10.5 liters per minute for 19 min 3 sec using the same setup mH the Cascade impactor. All gravimetric measurements were expressed in mg/j Gravimetric measurements of the respirable mass collected in the Cascade im tor were done for the Jeffrey and Coalinga Mine preparations only, because of known similarity between Jeffrey and UICC B chrysotile. Collection of Fiber Samples The same apparatus used to collect samples for dust concentration surements was also used to collect fiber samples. Samples to be analyzed by microscopy were collected on Millipore-type AAWP membrane filters at a fl rate of 0.1 liters per minute for 3 to 5 min depending upon the chamber conce tion of the chrysotile preparation. Before use, the filters were treated in a solution of Hyamin 2389 and dried at room temperature overnight to prevent s charging of the filter. Samples analyzed by scanning electron microscopy collected on a 0.2-pm Nucleopore filter for 2 sec and 5 sec at a flowrate of 10 li per minute. Preparation of Filters for Examination Light microscopy. After sampling, the filter was placed on a 25 x 75-mm gjj microscope slide, sample side down. The filter was cleared by holding the over an evaporating flask of boiling acetone. After clearing, a drop of glyi triacetate (Permount) was placed on the dissolved filter and a cover slip appliij Electron microscopy. Filters were secured to the polished side of a Gr planchet with carbon paint and were gold coated. The thickness of the gold was 100 to 150 nm. Fiher Characterization Light microscopy. A Beckett G22 graticule was used. At 500x, the mag tion used for sizing and counting, the graticule was a square, 100 pm on The lattice on the graticule had a spacing of 5 pm in one direction and 3 pm perpendicular direction. The following rules were used for fiber characteriza All fibers counted had at least a 3 to 1 length-to-diameter ratio (aspect ratio), fibers 5 pm in length or longer were counted. A fiber bundle which met the 3 aspect ratio requirement was counted as a fiber. Bundles with a diameter gr than 3 pm were not counted. Only fibers whose midpoint was located wi graticule were counted. At least 200 fibers were counted from 20 to 100 ran graticule fields for each filter sample. The characterization of each chrys preparation by light microscopy was based upon measurements from filters lected on a daily basis (5 days-week for 12 months). luectron microscopy. The magnification used for particle sizing and coun CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 37 llected were less than * j filter by optical micro ter stages in the Case adjusting the flowrate iss in length, tor was 200 liters. A to ling 0.2 m3 of chamber ing the same setup mini /ere expressed in mg/m^ ed in the Cascade impa tions only, because of t le. KOOO. To facilitate measurement of long fibers and fiber width, magnificaing from 1200 to 18,000 x were used. Counting and sizing of particles ftme directly on the electron microscope screen. Only particles whose midftjl within the area of the viewing screen were analyzed. This procedure [ly reduced all particles to points, thus causing every particle to have an pbability of being counted. To assure accuracy in measurement, a stan^8 made consisting of latex beads 1.099 ju.m in diameter on a 0.2-/xm Nu K filter. The filter and beads were gold coated and used to verify the magipn prior to each counting session. The area of the filter analyzed consisted ms randomly selected across portions of the filter. A field was randomly ipd at 10,000 magnification and then moved to the right. Several bands ijfiounted per filter and a range of 4 to 8 filters was analyzed for each Ipyie preparation. lust concentration me; es to be analyzed b\ embrane filters at a flow) n the chamber concent: rs were treated in a 0.1 ivemight to prevent statij lectron microscopy wei id,at a flowrate of 10 litei ijfc'following three categories were used to characterize the material on the jktfor each of the chrysotile preparations: ffjFiber: (a) At least a 3 to 1 aspect ratio was required, (b) Generally, a fiber flylinder-shaped with a uniform diameter. However, a slight separation of the j|.along any section of the fiber or at the ends was permissible, (c) The ^r. of the fiber was measured at the widest portion of the fiber whether it at the end or at some point along the fiber, (d) Smaller fibrils solidly ijlJito the major body of the fiber constituted a portion of the fiber. It was not mas a separate fiber. The diameter of the fiber would be measured at this ||t constituted the widest portion of the fiber. ~|ier cluster: (a) A fiber mass with at least a 3 to 1 aspect ratio was re- ed!'on a 25 x 75-mm gL_. med by holding the sli ' aikhg, a drop of glycerol and a cover slip applied ofcpshed side of a Gout ttiickness of the gold coi |gjj) It was composed of small fibrils usually oriented in the same direction, ally separated. Separation of fibrils may occur at any point along the the bundle or cluster, (c) The diameter of the cluster was measured at fjtest combined width along the fiber mass. A fiber cluster usually has a HjSriable width than a "fiber." . ^fonfibrous particle: (a) Any particle with less than a 3 to 1 aspect ratio, (b) j||gest dimension was measured as well as the greatest perpendicular mea- Tliit. |r-"floes" or clumps were also occasionally present on the filter consisting Jed fibers and fiber clusters. These were not analyzed because their com lj^t 500 x, the magnific ply prevented accurate separation into individual fibers and fiber clusters, , 100 fjbm on each side| tection and 3 fim in imfiber characterization presentative fibers, clusters, and nonfibrous particles were analyzed for eleuicomposition using x-ray energy spectrometry for each chrysotile preparajSjpach particle was analyzed for the presence of an element by a scoring ifeptio (aspect ratio). Onljj ranging from 0 to 4+. At 10,000x. 50 nonfibrous particles whose mid- which met the 3 to | MjSifell within the viewing area of a randomly selected band across the filter lphvith a diameter greatfl ^analyzed. In addition, the magnesium-to-silicon ratio was determined for all )jint was located within th(| Icicles containing these two elements. llifrom 20 to 100 randoc |||tion of each chrysotil^ RESULTS ;ements from filter coll ML Concentration Measurements ple sizing and countir ft:' total mass concentration and corresponding respirable mass concentration Joheach chrysotile preparation are given in Table 1 using both the Casella sampler 38 PINKERTON ET AL. TABLE 1 Gravimetric Measurements por Each Chrysotile Preparation in the Exposure Chamber Preparation Jeffrey UICC B Coalinga Jeffrey Coalinga (A) Chamber dust mass concentration (mg/m3)" 11.36 = 2.18 10.99 i 2.11 7.76 = 1.46 12.29 -- 3.33" 15.63 = 2.41*' (B) Respirable concentration (mg/m3)" Casella sampler 9.90 1.63 8.32 = 1.75 3.28 s 0.83 Cascade impactor 2.62 r 0.81" 0.801' Ratio b/a 0.871 0.757 0.423 0.213 o.OSl " All data are means SD, n = 240 for each chrysotile preparation. " For Jeffrey all data are means - SD, n = 12. '' For Coalinga the chamber mass concentration is based on three samples and the respirable co; tration is a single sample. and the Cascade impactor. When comparing the ratio of the respirable concentration to the total dust concentration found in the exposure chamber Jeffrey Mine chrysotile and UICC B chrysotile have a relatively high percen of respirable material based upon the Casella sampler measurements (87% fo Jeffrey chrysotile and 76% for the UICC B chrysotile). The ratio of the respi mass concentration to the total mass concentration is significantly lowei Coalinga Mine chrysotile (42%). Using the Cascade impactor a substant smaller fraction of the total dust concentration was found to be respirable. Fol Jeffrey fiber 21% of the total dust concentration was respirable and 5% of the dust concentration was respirable for the Coalinga fiber. Fiber Characterization--Light Microscopy Table 2 lists the percentage of fibers found for each given length interval by | microscopy for each of the three chrysotile preparations. Fibers greater tha ' /u.m in length and less than 3 /am in diameter were found in each preparation, percentage of fibers present from 5 to 30 p,m in length was similar for all aerosolized preparations with greater than 70% of all counted fibers falling this length interval. For the fiber-length intervals of 40-50, 50- 100. and gr than 100 jam, the percentage of fibers contained within each of these length i vals was significantly greater (P < 0.05) for the Jeffrey Mine preparation th either the UICC B preparation or the Coalinga Mine preparation. Fiber Characterization--SEM The length, width, and aspect ratios for the combined fibers and fiber clu are illustrated in Figs. 1--3. Figures 1A --D illustrate fiber length charac'eris Those fibers or fiber clusters having a diameter greater than 0.6 jam are show the crosshatched portions in these figures. This cutoff was chosen because fi CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE TABLE 2 Optical Microscopy Fiber Characterization; Percentage (%) Of All Fibers >5 /xm in Each Size Class" '' 39 Ratio B/A 0.871 0.757 0.423 'he respirable com ! Fiber size (/xm) Coalinga UICC B Jeffrey oo 5-10 10-20 20-30 30-40 40-50 50-100 32.9 7.4 28.8 2.7 17.1 3.4 10.2 2.3 6.0 a: 1.8 3.3 = 1.3 1.5 1.1 31.0 6.8 27.9 2.3 17.6 3.5 11.0 2.5 6.8 1.74 3.9 1.5 1.9 0.8 28.5 6.5' 25.3 2.7*-# 17.2 3.0 11.6 2.5" 8.2 1.8'* 5.7 1.9'# 3.6 1.7'# t data are means SD, n = 52 for each chrysotile preparation. Number of fibers counted per gfe = 1000-2000. H$jijers with a diameter greater than 3 /xm were not included in this study. *$-(< 0.05 when comparing Jeffrey to Coalinga using Duncan's multiple comparison test. 0.05 when comparing UICC B to Coalinga using Duncan's multiple comparison test. # o 05 when comparing Jeffrey to UICC B using Duncan's multiple comparison test. ! respirable massl |r than 0.6 /am have a substantially smaller probability of being respired than are chamber, the* high percentage ents (87% for thflj > of the respirable icantly lower foil >r a substantially) cspirable. Forth| nd 5% of the tot |'ifibers with smaller diameters (Pooley and Clark, 1979). Figure 1A shows yth distribution for combined fibers and fiber clusters of Jeffrey Mine Jle using a normal, numerical distribution plot. Figure IB illustrates the pdistribution for Jeffrey chrysotile using a log scale. The length distribution Is, and fiber clusters for UICC B chrysotile, illustrated in Fig. 1C, is very jpo that of Jeffrey Mine chrysotile. Combined fiber and fiber cluster length ihga Mine chrysotile, seen in Fig. ID, shows a distribution with many Exceeding 30 /am in length. ffboth the Jeffrey and the UICC B chrysotile, approximately 75% of the piped fibers and fiber clusters, were less than 5 /am in length, while less than Pthe combined fibers and fiber clusters from the Coalinga chrysotile were ,h interval by lighl s greater than 1 Iran 5 /am in length. At least 92% of the combined fibers and fiber clusters in Trey and UICC B aerosols were less than 10 ,u.m in length, but only 66% of i preparation. The] milar for all three] hers failing withii abined fibers and fiber clusters in the Coalinga chrysotile aerosol were less |Jp) /am. '|Tog distribution for combined fiber and fiber cluster width is similar for -100, and greati these length inter! eparation than folj Mine chrysotile and UICC B chrysotile as seen in Figs. 2A and 2B. Few $)pr fiber clusters exceed 0.6 /am in diameter. In contrast, (Fig. 2C) numerous ||Vand fiber clusters exceed 0.6 /am in width in the Coalinga Mine chrysotile >n. Ww'.'JAit9Ws . Miration. e log aspect ratio for the Jeffrey Mine chrysotile and UICC B chrysotile show IU^,piajority of the fibers and fiber clusters having an aspect ratio less than 100:1 and fiber clusters' mis?-1 3A and 3B). However, both the Jeffrey and UICC B chrysotile have some th characteristics.! ?)rs exceeding this aspect ratio. The Coalinga Mine chrysotile had no fibers or fim are show a byj clusters exceeding an aspect ratio of 100:1 (Fig. 30. 'sCn because fibers1 !> JftTables 3-5 contain a detailed description of the length distribution of fibers and Mr .* M ts; 40 PINKERTON ET AL. A JEFFREY MIME CHRYSOTILE 400 r UICC S CHRYSOTILE JEFFREY MINE CHRYSOTILE 01 05 1 5 10 50 100 LENGTH Uml D. COALINGA MINE CHRYSOTILE 1 5 10 50 100 LENGTH ((im) 5 10 50 100 LENGTH | fjm) Fig. I. Frequency distribulion of length for combined fibers and fiber clusters in each o| aerosolized chrysotiie preparations. (A) Jeffrey Mine chrysotile plotted on a linear scale for lei (B-D) Chrysotile preparations plotted on a log scale for length. fiber clusters, each as a separate category. The fibers and fiber clusters are] pressed as numbers counted per length interval, the percentage of the total fo in each length interval, the cumulative percentage, and the mean diamete fibers or fiber clusters in each length interval. The fibers and fiber clusters for preparation are also expressed in terms of their respective aspect ratio Tables 3-5. A comparison of Tables 3-5 demonstrates that the mean diameter of fiber fiber clusters for any given length interval is similar for the Jeffrey and UIO preparations. There was no fiber or fiber cluster measured in either the Jeffrej UICC B preparation that exceeded 2 jam in diameter. The aspect ratio increi for both fibers and fiber clusters with increasing length in the Jeffrey and UlCj preparations. The Coalinga preparation has a similar mean fiber and fiber clul diameter below the 5-ju.m length interval. However, for fibers and fiber clusl longer than 5 /j.m in length, the mean diameter in the Coalinga preparatio] significantly greater for both fibers and fiber clusters compared to the Jeffrey UICC B preparations. In general, as the fiber or fiber cluster length increase the Coalinga preparation, the width also increased. The presence of "thick'' fil CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 41 -E JEFFREY MINE CHRYSOTILE UICC B CHRYSOTILE COALINGA MINE CHRYSOTILE n = 1054 n=1013 n= 1050 00 OTILE 005 005 02 06 1 2 10 WIDTH (Mm) 2. Frequency distribution of log width for combined fibers and fiber clusters in each ||L|pp|jjDlized chrysotile preparation. 1 5 1000 ii each of thejji ale for length.' *"'l||`thick" fiber clusters in the Coalinga preparation is reflected in the smaller P|t ratios which in any length interval seldom exceeded 60:1. P? results of fiber counting by optical microscopy cannot be expected to fate closely with those obtained by scanning electron microscopy. The !!& for fiber counting by optical microscopy includes all fibers less than 3 /am ers are ex total found liameter of ors for eac.i t ratios in JEFFREY MINE CHRYSOTILE w W m m &. 1KSito: UICC B CHRYSOTILE COALINGA MINE CHRYSOTILE of fibers or id U1CC B Jeffrey or > increased id UICC B ber cluster mt iiboF er clusters paration is effrey and creased in 1. # ck" fibers 10:1 toon ASPECT RATIO 1000-1 10:1 100:1 ASPECT RATIO 10:t lOO'l ASPECT RATIO 1C0C 1 Frequency distribution of aspect ratio for combined fibers and fiber clusters in each ^Aerosolized chrysotile preparation. (// - 1054. 1013. and 1050. respectively.) TABLE 3 Pa h ik i b Siz e D is t r ib u t io n of Je ffr e y M in e C h r y s o it u -: in the A ero so lized S i a i e ILK): 1 -1 9 9 :1 200:1-499:1 100:1-199:1 200:1-499:1 >500:1 Nonfibrous particle, n Number TABLE 5 i i u i :Pa k Size D is t r ib u t io n o i: Co a u n g a M in e Ch r y s o t iu - in the A ero so lized S p a ie CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 45 ^er and greater than 5 /am in length. The criteria for fiber-fiber cluster ffcy scanning electron microscopy includes fibers of all lengths and diamjng as they possess at least a 3 to 1 aspect ratio. By these standards, 75% '' r-fiber clusters in the aerosolized Jeffrey and UICC B preparations y scanning electron microscopy would not have been counted by optical |jy. For the aerosolized Coalinga preparation, 60% of the fiber-fiber ypuld not have been counted by optical microscopy (10% of the total r cluster number exceeded the 3-/am-diameter limit and 50% of the total ' r cluster number was less than 5 /am in length). Particle Analysis ^ments detected in the 50 randomly selected nonfibrous particles for each e preparation are shown in Table 6. One half of the nonfibrous particles 1 in the Jeffrey and UICC B preparation samples demonstrated the presJ{piagnesium and silicon only. The magnesium-to-silicon ratio for fibers and Sters analyzed in all three preparations ranged from 0.62 to 1.34. Although ^`ans of identification, particles with a Mg-to-Si ratio in this range are likely gent fragments of chrysotile. The percentage of particles possessing the ^-to-Si ratio as chrysotile is 40% in the Jeffrey Mine preparation, 52% in JG B reference sample, and 20% in the Coalinga Mine preparation. The ls,!with low magnesium-to-silicon ratios of 0.30 to 0.40, found only in the fine aerosolized sample, may represent talc particles. DISCUSSION the chrysotile preparations designed for experimental research and zed in this study have not been previously studied in the aerosolized JTC B chrysotile has been studied extensively by both Timbrell (1970a) jlfctt (1973) in the aerosolized state. To fit the Coalinga Mine and Jeffrey pies into the spectrum of chrysotile preparations available for experiJfahalation research, UICC B chrysotile was used as a reference to link the j,study to the work carried out by other investigators, in particular, the ation studies of Timbrell and Beckett. Although differences exist in the f< of fiber collection from the dust chambers and in the preparation and |s of samples, the studies of Timbrell and Beckett on UICC B are compara|the present study (Table 7). By light microscopy we found a greater proporJ|the fibers falling into longer fiber length intervals than did the studies of and Timbrell. This difference may reflect differences in fiber preparation ..(counting techniques. By electron microscopy, the results of this study for chrysotile are nearly identical to those of Timbrell for the distribution of |jfengths. Beckett (1973) used SEM to determine fiber length and evaluated |.fibers longer than 5 fx,m. When our present results were recalculated and |pssed in a similar fashion (Table 7), we found a similar pattern, but proper ty fewer long fibers than did Beckett. Some of these differences may be due erences in the manner in which the aerosols were generated and the samples fccted. e Jeffrey Mine and Coalinga Mine chrysotiles have been elegantly analyzed 46 TABLE 6 Elemental Composition of the Nonfibrous Particles in each Chrysotile Preparation" Percentage occurrence Elements delected NaMgAlSiCa NaAl NaAISiKCaFe Mg MgAl MgAISi MgAISiKFe MgAlSiCaFe MgAlSiCrFe MgAISi Fe MgSi ratio Mg to Si = 0.3-0.4 Mg to Si = 0.6- 1.1 Mg to Si = 2.0-10 MgSiCa MgSiFe MgK Al AlSi AlSiCa AlSiK AlSiCr AlSiFe AIK AlKCa AlCr SiCr CaCr Cr Fe None Jeffrey 2 -- 2 2 4 14 -- -- 6 50 (6) (40) (4) -- -- -- 6 T -- -- -- -- -- -- -- 2 2 UICC B -- -- -- 2 -- 30 -- 2 -- --,, 54 (--) (52) (2) -- 2 4 -- ----- __ -- -- __ -- __ -- __ 4 Coalinga __ 6 ---. 4 -- 10 -- -- 2 2 20 (20 H -- 6 -- 14 4 --! 2f 4 T --i n' 14'< n n 4" __ ) -- ' 50 random nonfibrous particles were analyzed for each chrysotile preparation. ' May represent a contaminant from the rotary blade used to aerosolize the preparation. by others using the sample preparation technique of fiber dispersion in a medium (Wylie, 1979; Campbell et al.. 1980; Siegrist and Wylie, 1980). This of analysis is satisfactory for studies in which the preparations are to be ingi or injected in suspension. However, these studies do not provide an adeqj characterization of the asbestos preparations for inhalation studies since the1 cess of aerosolization is generally less efficient in fiber dispersion. A case injj is the Coalinga Mine chrysotile preparation. This asbestos preparation characterized by Campbell ci al. (1980) using fiber samples which were dispi in a liquid medium by ultrasonification for 10 min. The samples were analyze transmission electron microscopy. They found that 2.1% of the total chrysj tl'ICLES IN EACH 10 CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 47 TABLE 7 Fiber Length Distribution--Optical Microscopy Dispersion Percentage longer than stated length (pm) non Method Study 10 20 40 Aerosol Present 100 68.2 40.5 12.6 Aerosol Beckett0 100 32 11 5 Aerosol Timbrell1' 100 22.3 U 1.7 Alcohol Timbrell1' 100 46.9 21.5 7.0 Celloidin Timbrell6 100 53.6 16.9 3.8 Fiber Length Distribution--Electron Microscopy Instrument Study Timbrell6 Percentage longer than stated length (pm) 0.2 1 10 20 100 73.6 54.4 27.0 9.6 Present 100 88.2 64.2 25.5 6.7 Present 100 85.0 50.7 24.9 8.0 5 10 20 30 50 100 45 12 6 100 26.4 7.4 2.3 0.8 100 32.1 10.8 5.5 2.7 jure 2a in Beckett, 1973. fifed from Tables 6, 7, or 8 in Timbrell, 1970a. 1 ition. Phe preparation m Iglispersion in a liquid! pi, 1980). This type| Isfilre to be ingested1 foyide an adequate! Wes since the pro-* |jj|t: A case in point ||preparation was were dispersed fere analyzed by Iftotal chrysoule mi Jps analyzed were greater than 10 /xm in length and had a mean diameter of ffi.m. In the present study using aerosolized samples drawn directly upon ibpore filters and subsequently gold coated, it was found that 34.0% of the luted fibers and fiber clusters.or 28.1% of all particles (fibers, fiber clusters, Ijonfibrous particles) in the Coalinga preparation were greater than 10 /xm in jpi and had a mean diameter of 2.92 /xm. In a subsequent study by Siegrist and (1980), the Coalinga preparation (referred to as short-range chrysotile) was faeterized for particle size distribution by both transmission and scanning Iprpn microscopy. The samples were prepared by hand swirling in distilled Jer and dishwashing liquid (for SEM analysis) and by ultrasonification in water glO min (for TEM analysis). The results demonstrated that by both SEM and |M more than 90-95% of the particles were less than 10 /xm in length and more 95% of the particles were less than 1 /xm in diameter. In our study using fosolized samples 71.9% of the total number of particles were less than 10 /xm in [lgth and 68% of the total number of particles were less than 1 /xm in diameter, 'ff^. The differences between these two methods of sample preparation and analysis CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 49 50 PINKERTON ET AL. suggest that either ultrasonification breaks down fiber bundles into smaller dies and fibrils or that fibers tend to cluster in the aerosolized state. Other i tigators have shown that ultrasonification can result in the breakdow chrysotile into smaller fibers (Spurny et al., 1980). In addition, clustering offft caused by the aerosolization does not seem likely since the Jeffrey and Ulffl chrysotile preparations collected in the same manner as the Coalinga chryJ have distinctly different particle size distributions from that seen ini aerosolized Coalinga preparation. The illustration of each fiber preparation o| Nucleopore filters (Fig. 4) also show a distinct fiber morphology for the Co chrysotile compared to the Jeffrey and UICC B chrysotile. Some of the differences between each of the chrysotile preparations were reflected by the gravimetric measurements taken of each chrysotile aerosol exposure chambers. Comparisons made between each chrysotile preparatio' the ratio of respirable mass concentration to total mass concentration demons] a distinct difference between the Coalinga chrysotile and the other two pre| tions of Jeffrey and UICC B chrysotile. It would appear that compared ti Jeffrey and UICC B fibers the Coalinga preparation contains a greater propoj of particles which are captured in the elutriator system of the Casella sample! in the precutter stages of the Cascade impactor. These fibers and particle! captured because of their greater mass. Using scanning electron microscopy we found that the range of diamete fibers and fiber clusters was greater for the Coalinga preparation than fo1 Jeffrey and UICC B preparations in the aerosolized state. The trend of incre' diameter with increasing length was found only in the Coalinga preparatio bers and fiber clusters greater than 10 /u,m in length with diameters exceeding were present in the aerosolized Coalinga preparation while in the Jeffre UICC B preparations no fiber or fiber cluster of any length exceeded 2 diameter. The characterization of chrysotile in the aerosolized state in terms of length and fiber diameter is paramount in understanding the nature of any ration used for inhalation studies. Although the physical characteristics of at may not be the only factor in causing lung injury, fiber diameter and, to ai degree, fiber length play key roles in the potential for a fiber to reach the alv regions of the lung where injury as a result of asbestos inhalation appears more severe. The potential for each of these chrysotile preparations to cause lung injuK inhalation can be best assessed by a review of what is known about the ph| characteristics a fiber must possess in order to reach the alveolar portions ' lung. Spherical particles below a certain diameter (approximately 3-4 jam] reach the alveolar airspaces of the lung (Lippman and Albert, 1969). Larger} cles are eliminated by deposition in the upper airways primarily thrl sedimentation and impaction as a result of their greater mass. Timbrell. Harri' Fraser have examined the more complex aerodynamic properties of fibe comparing the deposition pattern of fibers to those of artificial spheres < fiml) 1965). In general they found that fibers have a similar deposition pattern to I spheres of three to four times greater diameter. This would suggest that CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 51 iles into smaller bi ed state. Other in\ n the breakdown <n, clustering of fibi Jeffrey and UICC e Coalinga chrysoi m that seen in >er preparation on, logy for the Coalin ater than 1 /xm in diameter would not be likely to reach the alveolar ^although the curvature of the fiber should also be taken into consideral^jrell, 1970b). Additional studies have shown that an equivalent fiber |i(>f 0.5 to 2.0 /xm results in deposition within respiratory bronchioles and jalveolar ducts, while fibers of smaller equivalent diameters result in Reposition in the more distal portions of the lung (Harris and Fraser, iiiey and Clark (1979) measured the diameter of chrysotile fibers recovijung specimens after tissue digestion. They found that only 0.06% of the a diameter exceeding 0.5 /xm. Although it is possible that these eparations were all ysotile aerosol in tl sotile preparation, ntration demonstral e other two prepant hat compared to th| a greater proportion Casella sampler ant ;rs and particles are'] fibers have fragmented longitudinally while in vivo (Suzuki and Churg, |fese results suggest that few fibers with a diameter greater than 0.5 /xm alveolar regions of the lung. If this limit in fiber diameter is true for lung deposition and the length of the fiber plays the major role in fijjsis and cell injury, the potential for the three asbestos preparations reported |study to cause lung injury by inhalation will be different. Only the Jeffrey ICC B chrysotile preparations have fibers which are less than 0.6 /xm in 6r when length is greater than 30 /xm. The Coalinga Mine chrysotile does Skess this property; instead, with increasing fiber length, fiber diameter also fjs. Most fibers longer than 20 /xm in length in the aerosolized Coalinga inge of diameters q| aration than for the! e trend of increasing) nga preparation. Pi ters exceeding 2 in the Jeffrey and1 i exceeded 2 /xm in] jtjons are also greater than 0.6 /xm in diameter. Thus, in an inhalation jtore fibers of greater length distribution would be deposited in the alveolar M the lungs for the Jeffrey and UICC B chrysotile preparations than for the ^ chrysotile preparation. ipoint one may ask the question, what constitutes a short-range fiber prepa re have seen that the Coalinga preparation originally thought to be a shorteparation contains numerous long fibers in the aerosolized state. The I'to this question should be based on the potential of a fiber to reach the 5 Q>in terms of fiber lamre of any prepaa&wjristics of a fiber i|`and, to a lesser 1 |^|ich the alveolar I '' "M appears to be regions of lung. In other words, is the fiber respirable? The work of mators who have studied the aerodynamic properties of fibers (Timbrell, |jl973) and the physical dimensions of respired fibers (Pooley and Clark, $ would suggest that the Coalinga chrysotile in the aerosolized state repre short-fiber preparation since only fibers less than 30 /xm in length are likely | respirable. In contrast, the Jeffrey and UICC B chrysotile preparations lin fibers greater than 30 /xm in length possessing a diameter which would t penetration of the fiber into the alveolar regions of the lung. Based on these ping injury by 1 | the physical ||yations, the Coalinga Mine chrysotile preparation can be considered to repSt a "shorter" fiber preparation in the aerosolized state. ibrtions of the |}summary, the three chrysotile preparations characterized in the aerosolized |3-4 /xm) can 5'in this study demonstrated distinct properties. Gravimetric measurements of Larger partiIttrily throug_h chrysotile preparation revealed that both the Jeffrey and UICC B preparas have a significantly greater portion of the total chamber dust concentration if, Harris and E|?||$|fich is respirable compared to the Coalinga preparation. By light microscopy it ||bf fibers by !&s found that for fibers greater than 5 /xm in length, the Jeffrey preparation in the ||P (Timbrell. |S/(?0s01i2ed state has a significantly greater fraction of fibers exceeding 40 /xm in f^^jngth than does UICC B chrysotile or Coalinga chrysotile. Finally, by scanning fthat fibers p^fM^ectron microscopy it was found that the Jeffrey and UICC B preparations pos- ',/rV 57 PINKERTON ET AL. sess fibers and fiber clusters which cross a large length range (the longest sured was 150 /xm), but none which exceeded 2 /xm in diameter. The Coali preparation also possessed fibers and fiber clusters across a similar length rai but many exceeded 2 /xm in diameter. In terms of respirability, the Jeffrey CJICC B aerosolized fibers constitute a mixed short-range and long-range fiber pr< ration, while the Coalinga aerosolized fibers represent a somewhat short-fiber pi ration in which very long respirable fibers are not present. In applying the abf fiber characterization studies to experimental research it is essential to rememl that further manipulation of these chrysotile preparations by grinding (Langei /.. 1978) or by fiber separation techniques may alter the fiber size distribm from that presented in this paper. ACKNOWLEDGMENTS This work was supported in part by NIEHS Contracts NOl-ES-0-0004 and N01-4-21-64-BCD| - REFERENCES Assuncao, J.. and Corn, M. (1975). The effects of milling on diameters and lengths of fibrous glas; chrysotile asbestos fibers. Amer. Ind. Hyg. Assoc. J. 36, 811-819. Beckett, S. T. (1973). The evaluation of airborne asbestos fibres using a scanning electron micros* Ami. Ocaip. Hyg. 16, 405-408. Campbell, W. J., Huggins, C. W., and Wylie. A. G. (1980). "Chemical and Physical Characteri: of Amosite, Chrysotile, and Nonfibrous Tremolite for Oral Ingestion Studies by the Natj Institute of Environmental Health Sciences." U.S. Bur. Mines Rep. Invest. 8452. Harris. R. L., and Fraser, D. A. (1976). A model for deposition of fibers in the human respir| system. Amer. Ind. Hyg. Assoc. J. 37, 73. Langer. A. M.. Wolff, M. S., Rohl, A. N., and Selikoff, I. F. (1978). Variation of propertl] chrysotile asbestos subjected to milling. J. Toxicol. Environ. Health 4, 173- 188. Lippman, M., and Albert, R. E. (1969). The effect of particle size on the regional deposition ofii aerosols in the human respiratory tract. Amer. Ind. Hyg. Assoc. J. 30, 257-275. Morgan, A., and Cralley, L. J. (1973). Chemical characteristics of asbestos and associated trai ments. hi "Biological Effects of Asbestos" (J. C. Wagner. Ed.), pp. 113-118. IARC Scie] Publication, Lyon. Pooley. F. D.. and Clark, N. (1979). Fiber dimensions and aspect ratio of crocidolite, chrysotil] amosite particles detected in lung tissue specimens. Ann. N.Y. Acad. Sci. 330, 7U-716. Pott, F. (1978). Some aspects on the-dosimetry of the carcinogenic potency of asbestos and| fibrous dusts. Staub-Reinhalt Luft 38, 486-490. Rendall, R. E. G. (1970). The data sheets on the chemical and physical properties of the UICCsl reference samples. In "Pneumoconiosis: Proceedings of the International Conference. Johaf burg. 1969" (H. A. Shapiro. Ed.), pp. 23-27. Oxford Univ. Press. London. TM Rendall, R. E. G. (1980). Physical and chemical characteristics of UICC reference sampl^j "Biological Effects of Mineral Fibres, (J. C. Wagner, Ed.). Vol. I. pp. 87-96. IARC Sci( Publication No. 30. IARC. Lyon. Seaton, A. (1975). Asbestosis. hi "Occupational Lung Diseases" (Morgan. C. Keith, and A. Si Eds.). Saunders. Philadelphia. Siegrist. H. G.. and Wylie. A. G. (1980). Characterizing and discriminating the shape of asl particles. Environ. Res. 23. 348-361. Sinclair. W. E. (1959). "Asbestos--Its Origin, Production, and Utilization." Mining public London. Spurny. K. R.. Opiela. H.. and Weiss. G. (19801. On the milling and ultrasonic treatmeii; ->f fibr biological and analytical applications*. In "Biological Effects of Mineral Fibres" t.l. C. Wt Ed.), pp. 931-936. (ARC Scientific Publication No. 30. IARC. Lvon. ngth range (the longest mi n in diameter. The Coalii across a similar length rani respirability, the Jeffrey ige and long-range fiber prej a somewhat short-fiber prq elsent. In applying the aboti ;&it is essential to remembi ilions by grinding (Langer er- the fiber size distributiol i*O;0CO4 and NOI-4-21-64-RCD. CHARACTERISTICS OF AEROSOLIZED CHRYSOTILE 53 T-and Layard, M. (1978). "The Carcinogenicity of Fibrous Minerals. Workshop on ftPefinitions and Measurement Methods," pp. 143-151. NBS Special Publication 506. j^jand Churg, J. (1969). Structure and development of the asbestos body. Atner. J. Pathol. \965)- The inhalation of fibrous dusts. Ann. N.Y. Acad. Sci. 132, 255-273. p68). A simple dispenser for generating dust clouds from standard reference samples of ' yi. Occup. Hyg. 11, 273 -281. 70a). Characteristics of the International Union Against Cancer standard reference Tasbestos. In "Pneumoconiosis: Proceedings of the International Conference, Johanjg9" (H. A. Shapiro, Ed.), pp. 28-36. Oxford Univ. Press, London. &}970b). The inhalation of fibers. In "Pneumoconiosis: Proceedings of the International &ce, Johannesburg, 1969" (H. A. Shapiro, Ed.), pp. 3-9. Oxford Univ. Press, London. !'(1973)- Physical factors as etiological mechanisms. In "Biological Effects of Asbestos" jtjhgner, Ed.), pp. 295-303. IARC Scientific Publication, Lyon. fl'foibson, J. C., and Webster, I. (1968). UICC standard reference samples of asbestos. Int. Wdr 3, 406-408. jt'C. (1965). The sequelae of exposure to asbestos dust. Ann. N.Y. Acad. Sci. 132, 691. |pG. (1979)- Fiber length and aspect ratio of some selected asbestos samples. Ann. N. Y. Acad. 30, 605 - 610.