Document RaXjRYV5BrQ0OJ0X7D5zv5o6a

AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 30:51 5-528 (1996) Mineral Phases and Some Reexamined Characteristics of the International Union Against Cancer Standard Asbestos Samples here oweir Arabia appe ;ncr-Warwick d Kingdom Norihiko Kohyama, PhD, Yasushi Shinohara, PhD, and Yasunosuke Suzuki, md Standard asbestos samples to be used for biomedical research were first prepared by the International Union Against Cancer (UICC) in 1966 in the United Kingdom and South Africa. Using modern techniques. X-ray diffractometty. analytical transmission electron microscopy, and thermal analysis, ire have now analyzed these UICC samples to determine the mineral compositions (mineral phases) and their respective quantities. UICC chrysotile A (from Zimbabwe) contains 2% fibrous antltophyllite as impurity: chrysotile B (from Can ada) does not contain any fibrous impurities, only non-fibrous minerals. UICC amosite and crocidolite are almost pure. UICC anthophyllite has 20-30% talc as impurity. The chemical compositions and fiber size distributions of the UICC asbestos samples have also been determined. The mean widths of the fibers ofchrysotile A and B are smaller than those of the amphibole fibers. This agrees well with the earlier results which showed the two chrysotile samples to have a larger respirable fraction than the amphibole:. '0 1996 Wiley-lJss. Inc. KEY WORDS: asbestos, standards, mineralogy, X-ray diffractometry, chrysotile, antho phyllite, amosite, crocidolite South Africa I be Veifierlamb . ol the 1976 ,ct- \f A 01923. ..hmJ A* "'ndex. Sci die of S INTRODUCTION The four main types of asbestos (chrysotile A and B. ' crocidolite. amosite. and anthophyllite) were prepared for . standard reference samples by the collaborative efforts of the Llandough Hospital MRC Pneumoconiosis Unit in the United Kingdom and the Pneumoconiosis Research Unit of the National Research Institute for Occupational Diseases in South Africa in 1966 [Timbrell and Rendall. 1972: Rendall, 1970: Timbrell. 1970] following the recommendations made after the Conference on the Biological Effects of As bestos held in New York in 1964. Since then, these asbestos National Institute of Industrial Health. Ministry of Labour Kawasaki, Japan (N.K., Y Sh I Division of Environmental and Occupational Medicine. Mount Sinai School of Medicine of the City University of New York. New York (Y Su i Address reprint requests to Dr Norihiko Kohyama. National Institute of Industnal Health. Ministry of Labour. 21-1 Nagao 6 Chome. Tama-ku. Kawasaki 214 Japan. Accepted for publication March 22. 1996 samples prepared by the International Union Against Can cer (UICC) have been used extensively in many important biomedical studies. The chemical and physical characteristics of the pre pared samples were initially and carefully determined [Ren dall. 1970: Timbrell. 1970]. However, the mineral phases occurring as impurities were not determined and no data have been obtained on the actual asbestos percentage in each sample nor on the types and concentrations of all im purity minerals [Graf et al.. I982|. Recently it has heen argued that chrysotile asbestos contains impurities, especially fibrous tremolite. This is an important argument because some fibrous tremolite has a strong carcinogenic action as shown in animal experimen tation [Wagner et al.. 1973. 19821. Such contamination could adversely affect any experimental results when the UICC asbestos samples had been used. For these reasons, it is important to know the type and quantity of mineral impurities in the UICC asbestos sam ples. especially if different kinds of fibrous minerals them selves exist as such impurities. This paper presents the de- U1996 Wiley-Liss. Inc. 516 Kohyama et al. TABLE I. UICC Standard Asbestos Samples Chrysotile A Chrysotile B Crocidolite Amosite Anthophyllite (from Zimbabwe) (from Canada) (from South Africa) (from South Africa) (from Finland) tails of the mineral phases of each UICC asbestos sample as well as the chemical and physical characteristics using mod em instruments and techniques. MATERIALS AND METHODS Mineral Phases To know the mineral phases, the UICC asbestos sam ples were determined by a conventional X-ray powder dif fractometer (Rint 2200. Rigaku Co., Japan) equipped with a Cu X-ray tube (1.5 kV) and a graphite monochrometer. The powder sample embedded in Al-specimen holder was ana lyzed at a power of 40 kV and 30 mA with a scanning speed of l/min for diffraction angle from 2 to 70 and a step scanning (0.02 interval; 50 and 100 sec exposures for each step) for the detection of amphibole phases. These data were treated by an attached computer. To identify the minor mineral phases effectively, a computer was employed for matching the X-ray diffraction (XRD) peaks with the min eral powder diffraction file.1 In this matching analysis, as the peak intensities were expressed in the square root, i.e., small peaks were enlarged and large peaks shrunk in one XRD pattern, all minor phases could be successively iden tified with high precision. XRD patterns for the UICC asbestos samples are shown in Figures 1 and 2. Figure I shows the XRD patterns of chrysotile A and B. of which low diffraction angle parts (5-50 in 20) were enlarged in Figure 1 (bottom) to show the minor mineral phases present in the two chrysotile sam ples. Figure 2 shows the XRD patterns of the UICC am phibole asbestos samples. The peaks of coexisting minerals (impurity minerals) were indicated on the figures of each sample. The precise data of each XRD peak, such as 20. d-spacing, intensity, identification, and Miller index, are omitted here. Table II shows the mineral phases identified and quantified in the five UICC standard asbestos samples by XRD. combined with analytical transmission electron microscopy (ATEM). The detection limit of the XRD anal ysis was about 0.5% in weigh!. Small amounts of other minerals were identified as im purities in chrysotile A (front Zimbabwe) and chrysotile B Published In I he Joint Committee ot Powder Diffrjciton Standards JOPDS' lmernauon.il Centre lor Ditlrustion Djta tICDDi. IW (from Canada). Anthophyllite was found in chrysotile A bj> XRD and was identified as fibrous by ATEM. Since stroi and major XRD peaks of talc are overlapped with soma peaks of anthophyllite, the presence of talc was also exam-! ined by ATEM for chrysotile A and talc was confirmed showing irregular platy shape. The finding of anthophyllif in chrysotile A will be discussed further. No tremoiite pi was detected in either of the chrysotile samples. Although we also examined the chrysotile samples for the presence tremoiite more precisely by XRD analysis employing a step scanning method, no tremoiite peak was detected at the! detection level of 0.1%. The other identified minerals wer$ all non-fibrous minerals, mostly of platy morphology. Quantification of coexisting minerals in the two UICQ? chrysotile samples showed that they exist in very small? amounts judging from the XRD patterns. The amounts ofy each mineral, such as anthophyllite, brucite, and pyroaurite'i in chrysotile A and B, were estimated by an internal stan-.- dard method using pure powdered quartz (from FukushimaJ Japan) as an internal standard. Pure (or mostly pure) sam-| pies of anthophyllite (from Afghanistan; fibrous), brucite (from China; platy) and pyroaurite (from Norway; phty)| were used for the standard minerals allowing the calibration! lines for each mineral to be quantified. The quartz powder* was added into the UICC chrysotile samples and.the satm'; pies of the standard anthophyllite, brucite, and pyroaurite at 10% in weight. The ratio of a peak intensity of each mineral.' to that of quartz was measured for each sample. Then, the calibration lines were made by connecting the intensity ratio.: with the origin of the coordinate axes by a straight line for, each mineral. By this XRD quantification, the quantity ofll anthophyllite in chrysotile A was estimated to be 3% byi weight. However, this percentage also contains a small, amount of talc and the actual content of anthophyllite was.j estimated as approximately 2% on the basis of the ATEMj observation. The other minerals present were also quanti fied. as shown in Table II; the contents of chrysotile in the?; samples of chrysotile A and B were 94% and 92%, respec-] tively. - Some very weak peaks of quartz, estimated at about 1 % or less by weight (Fig. 2). were identified in the amosite and crocidolite samples. No other peaks were found. In the an thophyllite sample, talc, chlorite, and mica (probably phlo- gopite) were identified (Fig. 2). The chlorite and mica were estimated as being about 5%. respectively. It was found that the UICC anthophyllite sample contains substantial talc ac cording to the ATEM observation. In order to estimate this amount of talc, the XRD pattern of the UICC anthophyllite sample was compared with those of other fibrous antho phyllite samples (Figure 3). Fibrous anthophyllite from Af ghanistan is mostly pure, but a very small amount of talc was found by ATEM. Fibrous anthophyllite from Uchida, Japan, also has a small amount of talc, but the amount seemed to be much smaller than that of the UICC antho- m 200 d) to W QO. Wc ? >< phyllit identit sample pe.sk-,. the , fihiou. detea- / exami firmed 'hyllite tepeak though ence of t a step at the Is were :y UICC small urn- of o.iorite il stanishima, :) sambrucite platy) bration lowder e samurite at n> r ' ,.iro me for i>'i! of by . small ite was ATEM quanti in the respec- out 1% ite and the anv- phloa were nd Put aic no i ns ihsllite anlhoim Afof talc 'cb'-'-'. ir 3000L m 2000 & lOOoL UICC Standard Asbestos Samples 517 Chrysotile A Chrysotile B 20 25 30 Diffraction angle (29,CuKa) FIGURE 1. XRD patterns of the UICC chrysotile samples. Ch, chrysotile: An. anthophyllite P. pyroaunte, H, hydromag nesite; B. brucite; Mg, magnetite;noise peak from specimen holder. ' phyllite. It also has a very small amount of vermiculite. The identification and quantification of talc in anthophyllite sample by XRD are very difficult because strong major peaks of 9.4. 4.67. and 3.12 Aof talc are all overlapped with those of anthophyllite. In addition, the orientation effect of fibrous anthophyllite is obviously large. Conversely, the detection and quantification of anthophyllite in talc by XRD are easier as reported IRohl and Langer. 1974], We are trying to develop a precise method for quantification of talcin anthophyllite. but by the present rough estimation, the UICC anthophyllite sample appears to contain over 20% of talc as impurity at least. The shapes of talc particles were mostly irregular platy ones, but some of them showed lath or fibrous shapes. As the fibrous talc could be misunder- X-ray intensity (cps) 518 Kohyama et al. ioo <ao/>, 2? 600 c a> cc FIGURE 2. XRD patterns of the UICC amphibole ashestos samples. C, chlorite; M, mica (phtagopite); A, T, overlapped peak of anthophyllite and talc; Q, quartz. ' `' TABLE II. Mineral Phases of the UICC Standard Asbestos Samples Determined by XRD Analysis and Some Other Methods Minerals Chrysotile Anthophyllite Talc Brucite lizardite Anligorite Pyroaunte" . Hydromagnesile'* Magnetite Total (%) Weight (%) Chrysotile A Chrysotile B 94__ 2 1 -- -- <1 2 -- - 100 92" -- -- 6 Trace" Trace" 2 Trace" Trace" 100 Minerals Amosite Crocidolite Quartz Weight (%) Amosite Crocidolite 99 -- -- 99 <1 <1 too too Minerals Anthophyllite Talc Chlorite Phlogopite Weight (%) AnthaphylliteJ 60 \ 30 i 55 i1 | 1 1 1 too 1 'Including 6-iayer serpentine at approximately t' '`Identified and estimated by ATEM analysis. Trace. <0 5% 'Pyroaunte. Mg5Fe3C0j(0H.CI) 16 4 h20. ptaty mineral "Hydromagnesite. Mg^COjhlOH) 2.4 H,0 ptaty mineral stood as fibrous anthophyllite. the precise identification by ATEM is required. The content of the talc was also esti mated by chemical analysis. Chemical Composition Major elements of the UICC asbestos samples were analyzed by gravimetric determination for SiO: and Fe:0-, + AUOj. colorimetric methods for FeO. and inductively^ coupled plasma-atomic emission spectrometry and atomic absorption spectrometry for TiO:. Fe,03, MnO, MgO. CaO,j Na;,0. and K;0. The other minor elements were determined; by X-ray fluorescence analysis for their presence and quaife! titles. s The results of our analyses of the chemical composition of the l ditionall pressed UICC a brell |l slightly com par Fot close to B wets. IMgi.Ol in the <. utcd to Th chemtc sample erocidi respect A\ ions u phvlliu crystal At small , p hi k U. chioru the a. ioooo 80 0 0 6 000 e 4000 20 00 UICC Standard Asbestos Samples 519 Anthophyllite(Afghan) 40 <%) ophyllite 60 30 5 5 <00 cmely atomic ). CaO. rmined | 15 20 25 Diffraction angle (20,CuKa) FIGURE 3. XRD patterns of the UICC anthophyllite sample and of the other localities. Numbers, d-spacing in A; V, vermiculite; C, chlorite; M, mica (phlogopite). of the UICC asbestos samples are given in Table lII.'Tra- If: ditionally. the cations of silicate minerals have been expi pressed in oxide percent. The chemical compositions of the . UICC asbestos samples were reported on earlier by Timbiell [1970]; since the expression of those results was |. slightly different from the traditional way, they could not be |; compared with our present data. For chrysotile A and B, the chemical compositions are close to ideal: the contents of MgO and FUCK+l in chrysotile B were slightly larger, which may be due to brucite [Mg(OH);] as an impurity. Part of the iron content will be in the chrysotile structure itself but part can also be attrib uted to the impurities, pyroaurite and magnetite. The amosite and crocidolite samples showed a typical chemical composition for these minerals, which means the samples are mostly pure. It is a characteristic of amosite and crocidolite to contain small amounts of Mn and Na ions. respectively. As minor elements, very small amounts of Cr and Ni ions were present in the samples of chrysotile and antho phyllite. These arc considered to replace the Mg ions in the crystal structures. An ideal anthophyllite mainly consists of Si. Mg, and a small amount of Fe for the cations. The anthophyllite sam ple. however, had noticeably high contents of AKO, and K.O, These latter correspond well with the presence of chlorite and mica (phlogopite) as impurities. The amount of the actual hydroxyl water, which was estimated to be 3.65%', was also larger than that of anthophyllite (calculated as 2.3% for the ideal crystal structure). The estimation of the actual hydroxyl water is described below.2 These results reflect the fact that the UICC anthophyllite sample contains talc, chlorite, and phlogopite as impurities. These minerals' ideal amounts of hydroxyl water are larger than that of anthophyllite, i.e.. 4.7%, 13%, and 4.3%, respectively. Us ing these values of ideal hydroxyl water, the contents of each mineral can be calculated as anthophyllite 60%. talc 30%. chlorite 5%. and mica 5% by the following equation: 2.3(%) A + 4.7<%> T + I3(%) x C + 4.3(%) x M = 3.65(%i where A. T. C. and M are the fractions of anthophyllite. talc, chlorite, and mica, respectively, and T = 1.0 -- A - C - VI. as the fractions of chlorite (C) and mica (M) had been \\ the iron in '.lie original anttiophyllile -.ample was all oxidize,! when heated at I lOO'C (ignition) to measure the content ot H;Q(+i. the value of hydroxyl water was measured smaller by the weight gain following the oxidation. To subtract the weight gam and obtain the actual content of hydroxy I water, the total iron was initially expressed by a feme form as in the parentheses in Table III. then only the feme iron was recalculated into ferrous form and the weight difference lb.32 - 5 75 = 0 was added to the initial value of 3.65rr producing 4.22^. Finally, the actual content of hydroxyl wa ter could be estimated to be 3.65c3- by subtracting the value of SO'. 0 57c(. which was recalculated from SO*.0.71^ and was considered to be contained tn the initial weight loss of H*Of+rt as a gas component, from that of HjO(+):4 22^ 520 Kohyama et al. TABLE III. Chemical Compositions of the UICC Standard Asbestos Samples' Chrysotile A Chrysotile B Amosite Crocidoiite Anthophyllite Si02 Ti02 ai203 Fe203 FeO MnO MgO CaO Na20 K20 H20(+) h20(-) Minor elements SO/ CIJ Cr203 COj03 NiO CuO ZnO Total (%) 39.89 0.02 0.76 1.97 0.49 0.06 42.60 0.33 Trace Trace 12.58 0.87 0.31 0.05 0.23 Trace 0.21 ND Trace 100.01 38.10 Trace 0.40 2.39 1.14 0.06 43.26 0.17 0.02 0.02 13.67 0.66 50.53 ND 0.55 1.90 35.34 1 82 6.43 0.51 0.02 0.27 2.32 0.20 48.84 0.02 0.06 19.07 19,95 0.11 2.32 1.08 5.58 0.06 2.33 0.34 0.13 0.25 0.07 Trace 0.12 ND Trace 100.08 ' 0.24 0.03 0.03 Trace Trace ND ND ND Trace Trace ND ND ND ND 99.93 '1 ' 99.76 55.31 0.03 0.99 (6.32) 5.75 0.17 31 15 0.29 0.03 0.43 (3 65)4.22 1.31 0.71 0.01 0.13 Trace 0.14 Trace 0.03 100.00 "Data obtained by wet chemical analyses except for anthophyllite and minor elements, which were obtained by X-ray (luotescence spectrometry. Trace. <0.01%, HD, not detected (<0 001); numbers in parentheses, see text. H20M and H20(-). weight loss by ignition up to 1,100C and 110C. respectively 'These values included in Ha0(*/--) and excluded from the total. obtained earlier by XRD analyses as 0.05. respectively (Ta ble II). T = (0.9 - A). From this, the above equation can be written; 2.3(5H A + 4.7(<7c)(0.9 - A) + W'rl x 0 05 + 4.3(5H x 0.05 = 3.65(9!-), Therefore, the fraction of anthophyllite (A) can be calcu lated to be 0.60. and that of talc 0.30. This calculation has a relatively large error, but can be used to estimate the approximate content of talc. Thermal Analysis Thermal analysis can indicate various kinds of thermal reactions of minerals, such as dehydration of absorption water, surface water, and interlayer water (occurring at ap proximately 100-140'C). dehydroxidation of structural wa ter (OH), oxidation of ferrous iron, phase transformation, etc., which are unique for each mineral. The differential thermal analysis tDTA). thermal sravimetre (TO), and de rivative thermal gravimetry (DTG) were simultaneously done using a DTA-TG apparatus (Thermoflex TAS300, Rigaku Co.) for the UICC asbestos samples. The heating rate and sample weight were 20C/min and approximately 20 mg, respectively. To prevent the oxidation of ferrous iron contained in the amosite and crocidoiite, the thermal anal yses were also done in an N: environment in addition to the normal analyses in air. The results are shown in Figure 4. In the DTG curves of chrysotile A and B. the peaks of 665C and 650C show the dehydroxidation of OH water in chrysotile. Exothermic peaks of 837^ and 83 UC in the DTA curves of chrysotile A and B show the recrystalliza tion to forsterite and/or enstatite from the amorphous state after the dehydroxidation. The other peaks in the DTG curves would indicate that minor minerals coexisted: 260C (pyroauritei. 372C (brucitei. 58()C and 560C (probably chlorite), and (.anustorite). FIGURE 4. DTA and DTG curves of the UICC standard asbestos samples. In air, analyzed in air environment: in N, analyzed m N, environment: A and B. chrysotile A and B. respectively isly (00, ting tcly iron nalthe 4, s of r in the 17.1- )C ihl> 1* 522 Kohyamaetal. The DTA and DTG curves of the amosite did not show Transmission Electron Microscopic any peaks in the air environment, but in the N2 environment (TEM) Observation the endothermic peak and weight loss peak were detected at 820C in the respective curves. The oxidation of ferrous ATEM analyses for the UICC asbestos samples were, iron in the amosite occurred at a relatively low temperature done using a TEM (H-8000, Hitachi Co., Japan) equipped of approximately 400-500C and continued gradually up to with an energy dispersive X-ray spectrometer (EDX, Kevex- 1,000C or more in air. The DTG curve shows a broad Co., USA) at an accelerating voltage of 200 kV. The UICCj weight gain curve between approximately 500 and 1,000C. asbestos samples were dispersed in pure water and a droplet - No mineral impurity was detected in the amosite sample by of the dispersion was put on an Ni TEM grid with collodion':; the thermal analysis. The final products of the amosite, as film coated by carbon. determined by XRD after heating to l ,200C in an air en The TEM photographs for UICC asbestos samples are : vironment, were quartz and cristobalite (Si03), magnetite shown in Figure 5a-e. It can be seen that chiysotiie A and ' (FeO FeiOj), hematite (Fe2Oj), and enstatite (MgSiO,), B consist of very thin curly fibers (fibrils) as well as thicker . and in an N-, environment were quartz, cristobalite. magne fibers (fiber bundles). In both samples, platy minerals were . tite, and enstatite. also found, which were identified to be lizardite, antigorite,' The DTA and DTG curves of crocidolite analyzed in pyroaurite, brucite, chlorite, and talc as minor impurities on;~ both air and an N, environment were slightly different from the basis of their selected area electron diffraction (SAED)' each other. In air, the DTG curve showed weight loss peaks patterns and EDX spectra. In chrysotile A, fibrous antho at 505C and 690C and a weight gain peak at 946C, while phyllite was also commonly found by TEM observation as the DTA curve showed endothermic peaks at only 680C impurity fibers. To observe the presence of anthophyllite and 946C. The low sensitivity of DTA is the reason why fibers more clearly and the ratio to talc, the chrysotile fibers the endothermic peak at approximately 500C was not de in chrysotile A were digested by a strong acid (HC1) and an ', tected. In Ni, the DTG curve showed two weight loss peaks alkaline (NaOH) solution, by which the minor minerals at 436C and 670C which became slightly lower than those such as anthophyllite and talc were concentrated in the re-: of in air. The weight gain did not occur at 946C as in air. mainder as shown in Figure 6. The SAED pattern and EDX The DTA curve showed four endothermic peaks at 430C. spectrum in Figure 6 proved the coexisting fibers are typical 640C, 821C, and 919C. The first two endothermic peaks anthophyllite. showing an example of the chemical compo would show the dehydroxidation of hydroxyl water of the sition as Si02 56.5%, MgO 32.4%. and FeO 10.1%, which crocidolite. The oxidation of the ferrous iron would occur at is also similar to that of UICC anthophyllite. The platy about 946C, corresponding to the weight gain and trans particles seen in Figure 6 were talc and the amount seemed formation of crystal structure. The endothermic peak of the to be much smaller than the anthophyllite fibers. The fi DTA at 946C would be due to the transformation. In N,. brous anthophyllite mostly consisted of relatively large fi the transformation occurred at 9I9C, but the oxidation did bers. such as 1-50 pm in length and 0.1-3 (im in diameter not occur abruptly but proceeded gradually and the DTG (Fig. 6). , curve did not show a clear weight gain peak at approxi In contrast to curly fibers of chrysotile, fibrous amphib- ' mately 919C. An endothermic peak newly appeared at oles usually show a rectilinear shape of the particles as seen 821C in the DTA curve in N-,. The cause of the peak cannot in Figure 5c-e, which are UICC amosite, crocidolite, and be explained at the present. No mineral impurity was de anthophyllite, respectively. The amosite and crocidolite tected in the crocidolite sample by the thermal analysis. The samples appeared almost pure by TEM observation, but the final mineral products when heated in an air environment anthophyllite sample contained many platy minerals, which were quartz, cristobalite. and hematite. When heated in an were mostly talc with some mica and chlorite. The talc N, environment, magnetite was found in addition to the particles occurred mostly as irregular plates, but fibrous talc other products. was also found. Fibrous talc can be differentiated from fi The thermal analysis for anthophyllite showed complex brous anthophyllite by SAED or EDX analysis. Of the three curves of DTA and DTG disclosing some coexisting min examples of amphibole asbestos, fine fibers were common erals. In the DTG curve, the peaks at 585C. 805C. 919C. in the crocidolite sample. and l.060C would be derived from the dehydroxidation of structural water of chlorite, mica, anthophyllite. and talc, Fiber Size Distributions respectively. The DTA curse was mostly consistent with the DTG curve. The exothermic peak at approximately 1.060C The fibers in all of the UICC asbestos samples obvi would be due to the transformation of the crystal structure. ously differ widely in aspect ratio (length and width), from The final mineral products after heating to 1.200C were very short and thin to very long and thick. To obtain the quartz, cristobalite. and enstatite. fiber size distribution, all of the fibers in a sample from 524 Kohyama et al. brane lowu techr were photi timet i r mi I :w l liter the v prim nifie' magi of 111 to ft that ; cou 11 disir fiber ir.otr - iW .1 si ol lit large and cido pies. pare is fa sanii the' fIC trasi do|n FIGURE 6. TEM pnotograph of the remainder of the chrysotile A sample chemically digested, showing anthophyllite fibers (a) and platty talc particles it) The SAED pattern and EDX spectrum show that the fibers are a typical anthophyllite Ni" in EDX soectrjm. noise from Ni TEM grid. t 1 j .. ' l short to long as well as from thin to thick must he examined bv electron microscopy at a proper magnification. Some new TEMs having a very high resolution at low magnifica tions dess than xIOOi have heen developed recently. Two such I [-.Ms having this ahilitv. I .EM-2000 'Topcon Co.. Japan) and H-8000. were used in the present study for ihe measurement of fiber sixes. The five L'ICC asbestos samples were well dispersed in distilled water and the proper portions were filtered on cel lulose ester membrane filter (Millipore filter") The mem me.: ano ftgu pap< the hull shot >se ur>' -in -'.lv UICC Standard Asbestos Samples 525 ibrane filter samples were transferred onto TEM grids fol The mean fiber widths of chrysotile A and B were less lowing low temperature ashing and carbon extraction than those of the amphiboles; i.e., the mean fiber widths of techniques [Kohyama and Suzuki, 1991 ]. The TEM samples the three amphibole samples were 0.31-0.38 pm, but 0.17 were observed using a TEM (LEM-2000 or H-8000), and the and 0.15 pm for chrysotile A and B. Timbrell [1970] re .photographs were taken at magnifications of a few hundred ported that the two chrysotile samples (chrysotile A and B) \ times. The negatives were enlarged on the print by 10 times would be more respirable than the amphiboles. The differ 'or more. The prints, therefore, had total magnifications of a ences of the mean fiber widths between the chrysotile and ifew thousand times. Fiber lengths were measured using ei- amphiboles, as measured in our study, seem to correlate 'ther a ruler or a scale magnifier of xIO magnification, and well with the ratio of respirable and non-respirable fibers the widths were measured by the scale magnifier on the reported previously [Timbrell, 1970]. Cyclone measure prints. When the fiber size was measured by the scale mag ment of the ratio of the respirable fraction in the anthophyl nifier, the size could be detected down to 0.01 (im at a total lite was reported by Timbrell [1970] as being a little bit magnification of a few ten thousand times. The total number greater than in the other amphiboles, but the present data of fibers counted for the size measurement ranged from 300 showed that the anthophyllite was greater in length and to 500 fibers for each sample. The benefit of this method is equal or thicker in width compared with the other amphib that almost all fibers on a very wide field of the TEM sample oles. Judging from these observations, the previous respi could be observed on one print. rable fraction would have been affected by talc particles The data for fiber sizes were plotted on a log-normal contaminating the anthophyllite sample. The present find distribution chart for each sample (Fig. 7a-e). The measured ings of the anthophyllite were obtained only from antho fiber size parameters such as geometrical mean and geo phyllite fibers, these fibers having been differentiated from metrical standard deviation (SD) for length and width are talc particles (mostly platy and some fibrous) by their char shown in Table IV. The mean lengths of the chrysotile A acteristic morphology, SAED pattern, and EDX spectrum. and B and crocidolite sample are slightly shorter than those of the amosite and anthophyllite samples. These also have a DISCUSSION large SD, indicating large fiber length distributions. Length and width measurements were repeated on amosite and cro The UICC asbestos samples have been very useful to cidolite samples from a different portion of our UICC sam manyexperimental studies on the health effects of asbestos. ples. The results of these second measurements are shown in Well-characterized and easily available asbestos standards parentheses in Table IV. The variability in width and length permit the comparison and assessment of results obtained i, fairly large, which is to be expected from such mineral by different groups. The reliability of the UICC asbestos simples. Considering the observed variability, however, samples has long been accepted. The samples were well tnese results clearly show similar size distributions for characterized for some aspects when they were first intro UICC amosite and UICC crocidolite. This is in direct con duced. However, complete and precise characterization of trast with the phonographs of samples of amosite and croci the mineral phases was not available. Our findings from dolite pictured on page 296 of "Biological Effects of As state-of-the-art procedures may give rise to some confusion. bestos," IARC Scientific Publication No. 8 [Bogovski et For example, there was no tremolite found in chrysotile A: al,, 1973]. The IARC publication suggests a severalfold it would appear that in some reports [Wagner et al,, 1986], thinner size distribution for crocidolite compared with the fibrous anthophyllite found in chrysotile A was believed amosite. to be fibrous tremolite. Another aspect concerns the purity The fiber lengths of all five UICC samples had been of each sample. Chrysotile A and B samples were found to measured earlier by Rendall [1970] and Timbrell [1970], have fiber contents of 94 and 92%. respectively; they con and for comparison those data were also plotted in the same tained other minerals as impurities. The UICC anthophyllite figures. The mean fiber widths are newly reported in this sample now appears- to be low in fiber content, approxi paper. Concerning the length distributions for all samples, mately 60%. These observations may require reevaluations. the Rendall data (curve 1 in Fig. 7a-e). which combined We detected some minor mineral phases at approxi both the data of optical microscopy (OM) and TEM. gave mately 0.5% levels in the UICC asbestos samples in this shorter lengths compared to those found in the present anal study, and the quantification limit was estimated as approx yses. The 1970 Timbrell data on length by TEM for the imately 1%. For amphiboles in chrysotile A and B. we airborne fibers (curve 3 in Fig. 7a-e) were also slightly examined at the 0.1% level using a highly sensitive XRD shorter than those of ours for amphibole samples (but of analysis. Chlorite minerals coexisted in both chrysotile A similar length for chrysotile A and B samples). It is likely and B. and hydromagnesite in chrysotile B was close to the that the data of OM (curves 2 and 4 in Fig. 7a-e). which detection limit, indicated as "trace" in Table II. Chlorite would measure fibers longer than 3 pm only, were biased to particles were also identified by the ATEM observation. express a longer size range than those of the TEM. The thermal analysis using DTG clearly detected the weight 526 Kohyama et al. loss peak due to the dehydroxidation of hydroxyl water of chlorite at approximately 580C. It has been demonstrated that DTG analysis is very effective for the quantification of minerals such as chlorite, brucite. and pyroaurite. when they have a lot of hydroxyl water and the detection limit some times exceeds that of XRD. Since DTG analysis is less useful for qualitative analysis than XRD analysis. DTG analysis is best utilized in conjunction with XRD and/or ATEM. Chrysotile A from the Zimbabwe mine contains fibrous ;j| anthophyllite estimated at approximately 2%. Based on our: unpublished data, most of the commercial chrysotile sam- pies I conta the l the 2 coult Visit seem babv tivel; anthi phyl obse injec and hanc phyl c.lie min; a re crea nie.v min. lion date obs; man of p live. Km The nae chi\ UICC Standard Asbestos Samples 527 TABLE IV. Fiber Size Parameters ot the UICC Standard findings, acknowledged to be based on sparse data, sup Asbestos Samples' ported the patterns of asbestosis, non-malignant respiratory disease, malignant mesothelioma, and lung cancer that have Length Width Mean (pm) SO Mean (pm) SO been reported elsewhere. Therefore, it is of interest to de termine whether the incidence of lung cancer and mesothe lioma in the Zimbabwe chrysotile miners and millers is Chrysotile A Chrysotile B Amosite Crocidolite Anthophyllite 2.4 2.6 4.3 (2.8) 2.5 (2.9) 4.2 2.3 2.3 3.3 (1.9) 2.0 (1.9) 3.3 0.17 0.15 0.31 (0.30) 0.33 (0.25) 0.38 1.8 1.8 1.9 (1.6) 2.1 (1.6) 1.8 significantly higher than among the other chrysotile miners who work with uncontaminated fibers. If a high incidence is observed, the anthophyllite fibers contaminating the chrysotile could have contributed to the incidence of lung cancer and/or mesothelioma. If the high incidence is not observed, it may be concluded that such a small contami nation of anthophyllite fibers has no effect on the incidence of lung cancer and/or mesothelioma. Therefore, further ep 'Mean, geometrical mean; SO. geometrical standard deviation. Mumtiers in parentheses represent the second measurement. idemiological study of the miners and millers in the Zim babwe chrysotile mine is necessary to assess the effect of such a low level contamination of amphibole asbestos (tremolite or anthophyllite) in chrysotile. Fibrous anthophyllite was found in chrysotile A, but no pies from Zimbabwe (which were milled, and sent to Japan) amphibole asbestos was found in chrysotile B at the detec contained fibrous anthophyllite in approximate amounts of tion level of 0.1%. Wagner [1974] reported that most of the the UICC chrysotile A. This means that the chrysotile from asbestos-related diseases, such as asbestosis, lung cancer, the Zimbabwe mine contained anthophyllite fibers which and mesothelioma, occurred in rats when they inhaled could not be separated by an industrial milling process. chrysotile B but no mesothelioma occurred due to chrysotile Visibly, the raw chrysotile fibers from the Zimbabwe mine A; i.e., chrysotile A (containing anthophyllite fibers as im seem to be harsh like amosite fibers. This implies that Zim purity) could not induce mesothelioma. It is interesting to babwe chrysotile could be geologically produced at rela speculate'as to why this difference occurred'. Generally, tively higher temperatures, in which both chrysotile and such discrepancy is supposed to be within experimental er i anthophyllite could exist as stable states. ror because the incidence of mesothelioma is very low as In some studies of carcinogenicity of fibrous antho induced by inhalation experiments in animals. However, at phyllite, extremely high incidences of mesothelioma were least it can be said that the effect of the fibrous anthophyllite observed in animal experiments in which this fiber type was occurring in such a small amount was not larger than that of injected directly into the peritoneal cavity of rats [Fukuda the chrysotile itself, which was the main component of the and Kohyama, 1925; Adachi et al., 1995], On the other sample. hand, information on cancer among employees of antho Kohyama and Suzuki (1991 ] reported the size distribu phyllite mines is relatively limited. In Finland, pleural cal tion and contents of asbestos fibers (mostly chrysotile and cification was first found in habitants near an anthophyllite amosite) in the lungs of North American insulation workers. mine by Kiviluoto 11960]. Meurman et al. [1974] reported The chrysotile and amosite that the workers used were a relatively high incidence of asbestosis. a significant in mostly imported from Canada and South Africa, respec crease in incidence of lung cancer, but no incidence of tively [Selikoff et al.. 1979]. The Canadian chrysotile and mesothelioma in the workers of a Finnish anthophyllite South African amosite could be considered to be almost the mine. Recently, the incidence of lung cancer and mesothe same as UICC chrysotile B and amosite. respectively. When lioma among the anthophyllite mine workers has been up we compared the size distributions of the two kinds of as dated; 3 pleural and 1 peritoneal mesothelioma have been bestos fibers found in the lungs with those of UICC observed among 503 deaths of the 999-miner cohort [Meur chrysotile B and amosite. it was clearly shown that both man et al.. 1994; Karjalainen et al.. 1994], A high incidence kinds of asbestos fibers in the lungs were shorter and thin of pleural plaque was also found in the residents who had ner than UICC chrysotile B and amosite. approximately one lived near an old anthophyllite mine and milling factory in half in length and one third in width. This phenomenon can Kumamoto prefecture. Japan [Hiraoka et al.. 1990. 1992]. be explained by a size selection after inhalation; i.e.. the The mine and factory had been operated for about 20 years, larger fibers were selectively trapped at the upper tracheal since the early 1940s. tree and relatively fine fibers were deposited in the lung In 1991. Cullen and Baloyi [1991 ] assessed the data for parenchyma. Moreov er. fiber separation from fiber bundles asbestos-related diseases that had been found among into fibrils and dissolution of fibers probably occurred in the chrysotile miners in Zimbabwe from 1980 to 1990. Their case of the inhaled chrysotile.