Document mJXewQY2YG59Kqy63rRDwE4d

Asbestos Minerals in S. Speil and J. P. Leineweber i-- \ J i. >> ' . i- i ! r Published in Environmental Research, Volume 2, Number 3, April 1969 Copyright 1969 by Academic Press, Inc. ; DPMC-00207 LAM 006693 ASBESTOS MINERALS IN MODERN TECHNOLOGY S. Speil and J. P. Leineweber Johns-Manvilie Research & Engineering Center Manvilie. New Jersey U.S.A. INTRODUCTION Asbestos is a generic term for a variety-^o-f hydrated silicate . minerals which have nnft^iimnn::aH:r.iL^P f namely, the ability, to_be_spajaejj_jjafco. relatively soff f silky fibers. Although the name is ordinarily associated with those varieties which have technologic importance, it is applicable to all minerals which fit the above description. The term "asbestiform minerals" is perhaps most descriptive. The known varieties of asbestiform minerals can be divided into two main classes on the nf ,.the.i.r-efY&taL strucr__ tures: serpentine and amphiboles. The sole jnembej- of the serpentine class is chrysotlle asbestos^, which is by fhrr~Ehe mo'st common-` Of the "asbestiform minerals. It. accounts for more than 95 per cent of the asbestos fiber prnduoexi-Qd.ay. There are five recognized asbestiform varieties of amphibolej^ci^Dcidolite, amosite, anthophyll i te r-----tremetirfee--and actinojjte~I-----Adrtliough the Amphiboles are common rock-forming miner als, the asbestiform varieties are much less abundant than chrysotile- The physical and chemical properties of the asbestiform min erals can be directly related to their crystal structure and chemical composition. In turn, the physical and chemical properties are responsible for the commercial importance of asbestos. It is understandable, therefore, that great empha sis has been placed on the elucidation of the structure and composition of these important minerals. Several comprehensive reviews on the asbestiform minerals have been published in recent years, including N. W. Hendry(30)/ r. Gaze^25), A. A. Hodgson133) and W. A. Deer, R. A. Howie, and J. Zussman^17). The objective of this paper is to bring this information up to date with particular emphasis on recent developments concerning the physics and chemistry of the asbestiform minerals. In addi tion, the uses of asbestos will be discussed briefly in rela tion to the properties of the individual species. DPMC-05279 LAM 008789 11 .rtutMi-iTHiinafiaaga ii-| inWIffllllWHlIiimilirt.iiiyum.lirP.l 2- - OCCURRENCE V The epigenesis and occurrence of the asbestiform minerals have been the subject of considerable geologic and petrologic re search. Chrysotile and amphibole fibers are found in en tirely different geologic formations. Chrysotile was most probably formed as a result of two separate metamorphic changes in ultrabasic rocks of volcanic origin. The first stage involved the formation of serpentine by the hydrother mal alteration of the original rock. At some later time the chrysotile was formed in cracks and fissures in the rock by recrystallization of the serpentine, again by an aqueous solution and reprecipitation process. In most cases, chryso tile occurs as "cross fibers" which are oriented in a paral lel array across the veins in the serpentine rock as shown in Figure 1. Occasional occurrences of "slip fiber" are found in which the fiber is oriented parallel to the vein as shown in Figure 2. A notable exception to the normal mode of occurrence of chrysotile is the fiber found in the New Idria serpentinite of Western California and at Stragari, Yugoslavians). The New Idria fiber is generally referred to as Coalinga fiber. The great bulk of this deposit consists of soft powdery pellet like agglomerates of chrysotile as shown in Figure 3. The material may be the result of intensive crushing and pulveri zation during or after serpentinization. In addition to its unusual mode of occurrence, Coalinga chrysotile is also un usual from the standpoint of its physical structure. This feature will be discussed in the appropriate section of this paper. The genesis of the amphibole fibers is not as clear-cut as that of chrysotile. Their name, taken from the Greek word amphibolos, meaning ambiguous, is a very apt choice. Figure 4 is a typical example of the mode of occurrence for crocido- lite which is found in the banded ironstones of the Transvaal system of South Africa. They are metamorphized rocks of sed imentary origin, which accounts for the variability in compo sition of the host rocks and consequently of the fiber. The only significant occurrence of amosite is also found in this area. Crocidolite is found in other areas, including Bolivia and Western Australia. Asbestiform anthophyllite is found in many places throughout the world, but there are only a few deposits of commercial importance in Finland and the United States. Tremolite and actinolite are the result of metamorphism of carbonate rocks. They are widely distributed in nature, but of little commer cial significance. Tremolite is a very common contaminant of commercial talc. ' " ~~ -- DPMC-05280 LAM 008790 #> -3- CRYSTAL STRUCTURE Chrysotile. The crystal structure of chrysotile asbestos was first determined by Warren and Bragg (72) an(j iater elucidated by Warren and Herring(69). These investigators determined that the mineral has a layered-type structure similar to the minerals of the kaolinite group. The basis of the structure is an infinite silica sheet (Si205)n in which all the sili ca tetrahedra are pointing in the same direction. Attached to one side of this sheet is a brucite Mg(OH)2 layer in which two out of every three hydroxyls are replaced by the apical oxygens of the silica tetrahedra. The result is a double sheet as shown in Figure 5. The mismatch in the di mensions of the silica and brucite sheets introduces a strain in the structure. Better matching of the layers and relief of the strain can be accomplished in three ways. 1. Substitution of larger ions in the silica sheet or smal ler ions in the brucite sheet. 2. Distortion of the octahedral brucite network or of the tetrahedral silica network. 3. Curvature of the sheet with the brucite layer on the outer surface. Ever since the first electron micrographs were published showing the apparent tubular structure of chrysotile(7, 50, 70), there has been considerable controversy over the mor phology of the fibers. Whittaker(76) t by means of careful X-ray diffraction studies, demonstrated that the lattice was definitely curved. Although he was unable to show whether the structure was a cylindrical arc, a closed circular cylinder, or a cylindrical spiral, he favored a spiral structure. The tubular concept was supported further when Maser, Rice and Klug(44) published the electron micrograph of Figure 6 showing an end-on view of a chrysotile fiber bundle. The fibrils were definitely cy lindrical and included many which appeared- to be pairs of concentric cylinders. A recent paper by K. Yada<81) has furnished what appears to be the final answer to the structure of chrysotile fibrils. By means of high resolution electron microscopy, he was able to observe the actual crystal lattice planes both parallel and perpendicular to the fiber axis. These pictures. Figures 7 and 8, show that most of the fibers have a hollow cylin drical form. The lattice planes have a multispiral arrange- DPMC-05281 LAM 008791 umkm lfP WwoHipi r rrr -4- ment confirming the prediction of Whittaker (76) # -Also appar-* ent in several of Yada's pictures is the presence of crystal lographic dislocations which strongly suggest that the basic structural unit consists of a single magnesia-silica sheet, rather than a double sheet as previously postulated by most authors. Yada's observations also confirm Whittaker's hypo thesis that the basic spiral element consists of five silicamagnesia units with approximately 10 silica-magnesia units forming the 70 A wall of a single fibril. Occasional fibers were observed which were solid rather than hollow. Although these fibers are relatively common in the samples, they are not enough to account for the discrepancy between the measured and calculated densities reported by Pundsack(55) . One final point should be made about the Yada micrographs. The outside surface of most of the fibers shows the presence of highly disorganized or amorphous material. This is the result of damage to the outer layer of fibrils by the elec tron beam under the conditions of observation. There are two other serpentine minerals found in chrysotile- bearing rock: lizardite and antigorite(20). They both have the same chemical composition and the same fundamental sheet structure as chrysotile. The differences between these mine rals reflects the way the strain in the crystal lattice has been relieved. Lizardite, which is the principal constituent of massive serpentine, generally has an extremely fine grained, platy morphology, visible only under the electron microscope in most specimens. Its structure has not yet been elucidated, but X-ray diffraction patterns indicate flat rather than curved sheets. Antigorite, on the other hand, does show evidence of curved sheets. Two of its unit cell dimensions are equal to chrysotile, but the third is much larger and variable. This third (b) dimension can vary from 18.5 A to as large as 100 A, compared to the 9.2 A value for chrysotile. It is believed that the structure consists of undulating sheets the periodicity of which corresponds to the variable unit cell dimension. Amphiboles. The basic crystal form of the amphibole minerals is less complicated than that of the serpentines. The basic structural unit is a double silica chain (Si^Oj^). As in the chrysotile sheets, all of the silica tetrahedra point in one direction. These chains are paired, "back-to-back", with a layer of hydrated cations in between to satisfy the negative charges of the silica chains. The final structure is formed DPMC-05282 LAM 008792 -5- by the stacking of these sandwich ribbons in an- ordered arc ray. A pictorial concept of this structure is shown in Figure 9. The various minerals in the amphibole groups are charac terized by the cations which occur in the structure. The principal cations are magnesium, iron, calcium and sodium. Since the bonding between these ribbons is rather weak, the crystals are easily cleaved parallel to the ribbons along A-A. If the cleavage is very facile, the result is an asbes- tiform mineral. For each variety of asbestiform amphibole, there is a corres ponding massive form with a different mineral name. Normally the asbestiform varieties are not found along with the mas sive counterparts. Undoubtedly, the local geo-chemical con ditions extant at the time of formation contributed to the relative ease of cleavage of any specific deposit and, there fore, to its commercial utility. The massive and asbestiform varieties have the same chemical compositions and X-ray crys tal structures. They can be distinguished by their physical properties and by petrographic examination. CHEMICAL COMPOSITION The chemical composition of commercially available chryso- tiles from various locations are shown in Table 1. For com parison, analyses of lizardite and antigorite are included. In all cases, it is apparent that the composition differs very little from the idealized composition of Mg3(Si2<)5) (OH)4. The impurities which are present may be part of the crystal structure or due to associated minerals. The most common impurity is iron. This can be in the form of ferrous (Fe++) or ferric (Fe+++) ions. It is generally assumed that the Fe+++ can be substituted for silicon in the silica sheets, and the Fe++ can be substituted for the magnesium in the brucite layer. The next most common impurity in chryso- tile is aluminum. Since aluminum can assume either tetra hedral or octahedral coordination it can be substituted in either the silica or brucite layers. Other impurities, gen erally found to be associated with chrysotile in lesser amounts than iron or aluminum, are calcium, chromium, nickel, manganese, sodium and potassium. The ionic radii of the ions commonly associated with chryso tile are given in Table 2. Since these ions vary consider ably in size, they can have an effect on the strains which exist in the chrysotile lattice. Ions which are larger than silicon and smaller than magnesium will tend to relieve the DPMC-05283 LAM 008793 -6- strain wheri substituted in the respective layers-. Aluminum fits this requirement for either layer, since it is inter mediate in size between silicon and magnesium. Ferric iron is larger than silicon so it will help relieve the strain when substituted in the silica layer. Ferrous iron in the brucite layer will increase the strain because of its larger size. Finally, it is improbable the very large ions such as the alkali or alkaline earths can truly be substituted to any significant extent in the brucite layer; and when present in appreciable quantities may exist as "interlayer" cations be tween the primary layers. Various authors have "normalized" chemical analyses of chrysotile, indicating the probable lo cation of the impurity ions in the crystal structure. The chemical composition of the asbestiform amphiboles is more complex than that of chrysotile. The idealized chemical formulas for the various species are given below. In these formulas, when cations are written in parentheses without sub scripts, a variable composition is indicated with the most aoundant species first. Crocidolite Amosite (Na2Fe3++Fe2+++) Si822(OH) - (Fe++ , Mg)7 Sig022(OH) Anthophy Hi te - (Mg, Fe++)? Si822(0H) Tremolite Actinolite - Ca2Mg5 - Ca2(Mg, Fe++)5 Si822(0H) si822(0H) The range of chemical analyses for these varieties of amphi boles are listed in Table 3. Detailed analyses can be found in various publications(17, 23, 32). The considerable varia tion in composition which can occur is readily noted. The actual identification of a particular amphibole species may depend on which of the idealized compositions the sample in question most closely represents. This variability in compo sition is a direct consequence of the fact that the structure can accommodate many different ions in the space between the silica ribbons, and the variable nature of the host rocks can contribute different ions to this structure. Accessory Minerals. The analysis of asbestiform minerals is often complicated by the fact that the samples may contain fragments of the host rock and its associated minerals, and also that other minerals may be intimately intergrown in the fiber bundles. Contamination due to host rock fragments is common in commercial fibers. In the case of chrysotile' as bestos, the most common contaminants are the other serpentine DPMC-05284 LAM 008794 -7- minerals--lizardite and antigorite. While these species have similar overall compositions, the trace element analyses could be influenced by their presence. Other minerals which are found in serpentine masses, and which could be found in commercial chrysotile fibers, are: magnetite, brucite, chro mite, calcite, magnesite, olivine, pyroxene, tremolite, ac- tinolite, chlorite, talc, and chalcedony. The proportions of these minerals will vary considerably with the location and nature of the deposit. For a specific commercial mine, the nature and content of the impurities in the ore will be rela tively constant. Neighboring deposits in the same serpentine belt may differ considerably in their impurities. Several good publications are available(19, 27, 45) which describe the geology and mineralogy of the various chrysotile pro ducing areas. Other than lizardite and antigorite, the most common miner als associated with chrysotile are magnetite and brucite. Both of these species often are found grown within the fi bers. In many occurrences, the ends of the fiber bundles are capped v/ith a magnetite-rich layer of rock. A mineralogic curiosity which is found in some chrysotile deposits includ ing the Jeffrey mine at Asbestos, Quebec, is a fibrous form of magnesium hydroxide which has been given the name nema- lite(9). These fibers usually occur in bundles which are often several feet long. Reimschussel (60) studied the association of chromium and ni ckel with carefully separated components of ore from the Jef frey mine at Asbestos, Quebec. Using a combination of chem ical dispersion and magnetic separation, the asbestos was di vided into fibrillar chrysotile, serpentine, and magnetic concentrates. All of the chromium was found to be associated with the mag netite phase, most probably as an isomorphous substitute for ferric iron. When chromium appeared in the chrysotile or serpentine fractions, it was associated with the last traces of magnetite which were impossible to remove. Most of the nickel in the ore is also found in the magnetic fraction. A small amount may be associated with the magne tite, but the majority occurs as a separate phase. This phase is the iron-nickel alloy, awaruite, whose composition ranges from FeNi2 to FeNi3. It can be separated by a dif ferential solution method proposed by E. H. Nickel^ . Fin ally, a small amount of nickel, about 0.008 per cent, is found to be present in the chrysotile lattice, most probably as a substitute for magnesium. DPMC-05285 -8- Organic Impurities. The association of benzo(a)pyrene and other organic impurities with asbestiform minerals from var ious sources was investigated by Reimschussel'^0). The total amount of extractable organic matter was determined by long term Soxhlet extraction with cyclohexane. After drying and weighing, the organic residue was analyzed for benzo(a)pyrene by thin layer chromatographic techniques. The fibers examined included North American and African chrysotiles, crocidolite, amosite, and anthophyllite. All samples contained measurable amounts of extractable organic matter, ranging from 40 to 500 parts per million. There was, however, no correlation between the amount of organic matter and the amount of benzo(a)pyrene present. The Canadian and United States fibers (11 chrysotiles and one anthophyllite) contained no detectable benzo(a)pyrene. The detection limit varied with the amount and complexity of composition of the extracted organic matter. It was as low as 0.02 parts per billion and generally below 5 parts per billion. All of the fibers from Africa and Finland, on the other hand, did con tain benzo(a)pyrene. The highest concentration (150 parts per billion) was found in crocidolite from the Cape Province, South Africa. Crocidolite and amosite from the Transvaal Province contained 12 to 18 parts per billion; Rhodesian chrysotile and Finnish anthophyllite contained less than 10 parts per billion. These results are in general agreement with those reported by Harington(2) except that Harington found no benzo(a)pyrene in African chrysotiles. The explanation of this difference may lie in the fact that Harington used virgin samples collected in the field, whereas Reimschussel analyzed typical commer cial products. This could mean that at least some of the benzo(a)pyrene is introduced during the processing or ship ping of the fiber. SURFACE CHARACTERISTICS The surface characteristics of the asbestiform minerals; are very important in relation to their commercial uses and to their interaction with whatever environment they may be ex posed to. Most of the discussion relates to chrysotile be cause the surface characteristics of the asbestiform amphi- bole minerals have received much less attention. The external surface of chrysotile fibers consists of magne sium hydroxide and, therefore, it is not surprising that the fibers behave in some respects as though they were magnesium DPMC-05286 LAM 008796 >nn)niiltfi>riMn*i--iiirtlin -9- hydroxide. For example, Pundsack(53) determined that the pH of a suspension of chrysotile in carbon dioxide-free dis tilled water is 10.33. This compares to a value of 10.37 for a magnesium hydroxide suspension under the same conditions. Pundsack and Reimschussel also determined the "solubility product constants" for various chrysotile fibers(56). The values ranged from 1.0 x 10"H to 3 x 10~12 an(j correspond quite closely to a value of 1.9 x 10~H reported for magne sium hydroxide(64)_ Surface Charges. The electrokinetic behavior of chrysotile is another manifestation of the magnesium hydroxide surface. Martinez and Zucker(^2) studied the effect of pH on the sur face charge, or zeta potential of asbestos ore body minerals by the streaming potential method. Figure 10 shows the com plete pH vs zeta potential curve obtained by these authors. They found the isoelectric point of chrysotile to be 11.8. At lower pH values, the surface charge is positive; above the isoelectric point, the charge becomes negative. They attri bute the sharp increase in potential which is obtained as the pH was lowered from 7 to 3 to removal of hydroxyl groups from the surface and resultant exposure of the magnesium ions. Be low ph 3, the magnesium ions are removed and the silica sur face exposed, accounting for the decrease in zeta potential in this range. The electrokinetic behavior of lizardite, also shown in Fig ure 10, is significantly different from that of chrysotile. It has an isoelectric point of 9.7 with a much smaller charge than chrysotile. Furthermore, the sharp rise in potential, between pH 7 and 3, is not found. This would be in accord with the proposed undulating structure in which both silica and magnesium hydroxide, surfaces are exposed. Chemically, the surface of the amphiboles is similar to that of silica. It is polar in nature, but not as highly polar as chrysotile. The electrokinetic charge is negative and small er in magnitude than the positive charge of chrysotile. Iso electric points have not been well established, although it is presumed that the charge would become positive at very low pH. Most materials have a negative surface charge in aqueous sys tems. Since chrysotile has a positive charge, it will at tract or be attracted to most dispersed materials. This characteristic of chrysotile manifests itself in many of the commercial applications. In addition, the highly reactive surface causes many interesting surface reactions to take place which are intermediate between simple adsorption and true chemical reaction. '.For the sake of better continuity, these interactions will be.discussed along with the chemical properties of the fiber. DPMC-05287 LAM 008797 -10- Surface Area. The specific surface area of chrysotile asbes- tos as determined by gas adsorption measurements has been found to vary considerably with the physical condition of the fibers. For example, Pundsack(54) reported that pulling fibers from a block of crude Jeffrey chrysotile fiber with tweezers gave products with surface areas, as determined by nitrogen adsorption, ranging from 4 to 12 sq meters/gram (m^/g) depending on how thoroughly the fibers were pulled apart. When the fibers were opened further in a Wiley Mill, the comparable surface area was more than 30 m2/g. Values in excess of 50 m2/g were obtained when the fibers were soaked in an Aerosol OT solution to separate individual fibrils. ' Naumann and Dresher(48) studied the surface areas of various chrysotiles in more detail. They also found a considerable variation in surface area with the degree of fiber opening for most fibers. The two exceptions were the chrysotile from New Idria (Coalinga) and Stragari. In these cases, there was very little variation with the degree of opening. Theoreti cal surface areas of the chemically dispersed fibers calcu lated from measurements of fibril diameter distributions were in good agreement with the measured values, as shown be low. ' Fibers from Average Observed SA Calculated SA55 Fibril Diameter Canadian Grade 7R 50 m2/g 55 m2/g 375 A New Idria 78 m2/g 76 m2/g 275 A The authors propose, that for those fibers whose surface areas are sensitive to the degree of opening, the voids be tween fibers are partially filled with solid material, and therefore not accessible to nitrogen or other gases. In the case of the New Idria fibers, these interfibril voids are available for adsorption. In either case, the intrafibril voids are not available. Several workers 48, 54) have attempted measurements of the pore size distribution of various asbestos fibers, in cluding chrysotile and amphiboles. .The methods have included water vapor adsorption, nitrogen absorption and mercury pene tration. Most nitrogen adsorption'' results show a peak in the vicinity of 20 angstroms which has been interpreted as a measure of the radius of the pores within the fibrils. Harris (29), however, points out that this may be an artifact of the measuring method and is open to serious question. DPMC-05288 lam 008798 irrr'n- iBiii ' .ifilrlrifirrijlinliiW I, -11- The surface area of amphibole asbestos is considerably lower' than chrysotile; it does not exhibit any unusual porosity. Patterson and Thompson(52) reported that sawh blocks of Wit- tenoom crocidolite have surface areas around 5 m2/g. This increases to between 7 and 8 on teasing fibers from the block. The fully fiberized material had a value of 14,8 m2/g. Adsorption. The adsorption of various materials on the sur face of chrysotile has been studied from both the liquid and vapor states. Young and Healy (84) studied the adsorption of several vapors on chrysotile. They found that nitrogen, ar gon, carbon monoxide, acetylene, n-butane, trimethyl amine, and dimethyl amine all gave surface areas of 9.7 m2/g on grade 7R Canadian fiber. Ammonia and water vapor, however, gave surface areas of 17.6 m2/g for the same sample. The dif ference could not be explained in terms of chemisorption or other specific interactions because all the isotherms were completely reversible. Their conclusion was that the ex tremely polar water and ammonia molecules could be adsorbed on portions of the surface which are not available to less polar molecules. They further concluded that some of the pores in chrysotile may be plugged with water, and that these plugs are permeable to polar vapors only. Young and Healy also report a similar anomolous sorption of water vapor on antigorite, a non-fibrous serpentine. Antho- phyllite and tremolite, the only amphiboles studied, did not exhibit this behavior. The adsorption of various organic compounds on chrysotile from both the liquid and vapor phases is currently being studied by Weeks and Leineweber(75)_ The fiber used in this study was specially air cleaned to remove most of the non- fibrous material and extracted with carbon tetrachloride to remove organic contaminants. Ethanol, benzene, and hexane all exhibit normal isotherms on chrysotile. The surface areas, estimated from the isotherms are ethanol, 18.8 m /g; benzene, 11.2 m2/g; and hexane, 9.6 m2/g. The nitrogen sur face area is 21.2 m2/g. The corresponding heats of adsorp tion are ethanol, 13 kcal/g; benzene, 11 kcal/g; and hexane 9 kcal/g. Zettlemoyer, et al<85), reported 16 kcal/g for the heat of adsorption of water on chrysotile. These adsorption data support the obvious premise that the polar surface of chrysotile: has a greater affinity for polar molecules than for non-polar, in general agreement with the findings of Young and Healy1. DPMC-05289 LAM 008799 -12- Adsorption from solution is complicated by the- concur- rent adsorption of solute and of solvent, making the inter pretation of the isotherms quite difficult. Weeks and Leine- weber studied the following binary systems on the same fiber used for the vapor adsorption studies: benzene - ethanol; benzene - tert-butanol; benzene - hexane; benzene - naphtha lene; benzene - anthracene; hexane - naphthalene. Typical adsorption isotherms obtained from these systems are shown in Figure 11. These isotherms of concentration change are plots of the "apparent" or differential adsorption of the "solute" (the component whos.e concentration is indicated on the ab scissa) vs concentration. These concentration changes were determined by means of a differential refractometer. The ex tremes in the types of isotherms obtained are illustrated by the benzene-ethanol system and the hexane-benzene system. In the former, the apparent adsorption of ethanol is positive at low concentrations and negative at high concentrations, whereas in the latter system, the apparent adsorption of hex ane is negative throughout. A qualitative interpretation of these effects is that the affinity of the surface for ethanol and benzene are essentially equal, while benzene is more strongly adsorbed than hexane. The affinity of the surface for the compounds studied can be listed in the following de creasing order: x ethanol = butanol = benzene ^-naphthalene> anthracene > hexane Quantitative interpretation of these isotherms by the methods proposed by Kipling and Tester(36) was impossible because sufficient information was not available to be able to ex tract the individual isotherms. CHEMICAL CHARACTERISTICS Asbestos has often been touted as the "indestructible min eral". In reality, this is far from the case. As far back as 1885(67) the reactivity of chrysotile with acids was recog nized and in 1890 Clark and Schneider(16) found that chryso tile was the most susceptible to acid attack of all the ser pentine minerals. Nagy and Bates(47) and Nagy (46) confirmed this conclusion with electron microscopic and X-ray diffrac tion studies on ac,id-treated chrysotile and antigorite. After treatment with IN HCl for one hour at 100C, the chrysotile X-ray diffraction pattern completely disappeared while that of antigorite was relatively unchanged. Electron micrographs of the reaction products showed that chrysotile was very severely etched and had lost its tubular morphology. Faust and Nagy(21, 22) studied the differential solubility of chrysotile DPMC-05290 LAM 008800 IH*ngft>rmiii'\Tirib&itV T jj'i i lifim 'uSiss^eim UHiMi rnmrnKmmmikammab^mmm and serpentine in more detail. They confirmed that chrysotile is almost completely destroyed in IN HC1 for one hour at 95C, while antigorite is almost untouched under the same condi tions. The reactivity of lizardite is intermediate between that of chrysotile and antigorite. Badollet(2f 3) summarized the available information on the stability of asbestiform minerals. Strong acids decompose chrysotile rapidly with the removal of all MgO and a total weight loss of 60 per cent. The residue which remains after acid attack consists of amorphous silica which retains a very fragile fibrous morphology. In contrast to the sensitivity of chrysotile, the amphibole fibers are much more resistant to acids. There are, however, significant differences between these fibers. The data in Table 4 shows that anthophyllite, crocidolite, and tremolite are significantly more resistant to acid attack than amosite and actinolite. Twenty-two days at room temperature had es sentially the same effect as the 2 hr reflux exposure. All of the fibers were relatively stable in 25 per cent sodium hy droxide solutions. The high solubility of actinolite was at tributed to impurities in the sample. Hiscock(33) m0re recently studied the rate of decomposition of asbestiform fibers in boiling 4N hydrochloric acid. He found the following relative order of stability as shown in Figure 12. tremolite > anthophyllite> crocidolite>actinolite > amosite chrysotile After an initial rapid weight loss, the rate of attack de creases radically with tremolite and anthophyllite showing ex tremely low rates. Recently Weeks studied the rate of de composition of chrysotile fibers less than one micron in dia meter in 0.12N hydrochloric acid at 37C. Decomposition was determined by analysis for magnesium which had gone into solu tion. The decomposition is relatively slow under these con ditions as shown in Figure 13. The straight line indicates that the decomposition is diffusion controlled. This con clusion is justifiable because extraction of the magnesium from the chrysotile structure leaves a residue of silica through which both the acid and magnesium ions must diffuse for the decomposition to continue. DPMC-05291 LAM 008801 "'V  '~iMrrr '^if*tir>rfaiaa^aatMMaa -14- It is also interesting to note that extrapolation to zero time v indicates an initial decomposition of approximately 6 per cent. This decomposition probably represents the immediate dissolution of the surface magnesium hydroxide. X-ray dif fraction analysis of the 6-1/2 hr reaction product showed no difference from the original material in spite of the fact that it was more than 50 per cent decomposed. The acid appa rently attacks the surface of the fibrils leaving an unreacted core which diffracts X-rays the same as chrysotile. This is in accord with the observation that chrysotile fibers can be readily fibrillized by agitation in weak acid. Electron mi crographs of the reaction product show an etched surface on typical chrysotile tubes. The residual silica on these par tially decomposed fibrils is probably responsible for this etched effect. The very rapid initial reaction of the surface hydroxyls was also noted by Pundsack^6) during the stepwise titration of cnrysotile with 0.5N hydrochloric acid at 100C. Ample time was allowed for equilibrium to be attained after each incre ment of acid. The titration curve. Figure 14, exhibits sev eral inflection points. The first increments of acid cause a sharp decrease in pH from an initial value of 10 to about 6.8. This point corresponds to the reaction of about 4.5 per cent of the fiber with acid. After the first inflection the curve levels off until about 67 per cent of the fiber is reacted and finally tails off to a pH of 2.3 as the reaction is completed. Pundsack estimates that about 7 per cent of the total hydrox yls exist on the surface, and that 67 per. cent decomposition represents reaction of all the structural hydroxyls. The final 33 per cent reaction represents reaction with the silica gel formed by the process. The chemical reactivity of crocidolite has been studied in considerable detail by Thompson(68) using fibers in the dia meter range of 0.05 to 0.15 microns. In a Soxhlet extractor, water removed 4 per cent of the silica and 6 per cent of the sodium ions. These values correspond to a depth of attack of 1 and 1-1/2 unit cells, respectively. Similarly, in boiling alkali the attack is limited to the surface layer. In 5N hydrochloric acid at 100c, 20 per cent of the struc ture is disrupted within three minutes, corresponding to a 57 A penetration. After three minutes, the rate drops rapidly with less than twice the above penetration after 6 hr, agree ing generally with Hiscock's findings ( 33). Thompson proposes that the formation of a tough coating of polymerized silica protects the crystals from further attack. This hypothesis is suoported by the fact that the fibers are again susceptible to acid attack if the silica layer is removed by reaction with alkali. DPMC-05292 LAM 008802 In the presence of 0.2N EDTA at pH 5.5 and 100c, the rate of* decomposition of the fiber appears to be diffusion controlled. In this case at least a portion of the silica becomes dis persed in the reaction medium. Thompson concludes that the amphibole structure is not intrinsically resistant to acid at tack. The apparent resistance is a result of the protection afforded by the silica layer. The lesser resistance of amo- site to acid attack can possibly be attributed to faults in the structure. The resistance of the asbestiform minerals to attack by re agents other than acids is generally considered excellent at temperatures up to 100C, but deteriorates rapidly at higher temperatures. Ball and Taylor(6) studied the reactions of chrysotile with several materials under hydrothermal condi tions . Recent studies by J. C. Yang(83) indicate that reaction be tween chrysotile and calcium hydroxide is detectable in two days at 230C. Similarly Reimschussel '60) foun(j that chryso tile was completely decomposed in concentrated potassium hy droxide at 200C within 24 hr. Thompson(69) showed that cro- cidolite is attacked by potassium lOOOc and studied the hydrothermal other chemicals. or sodium hydroxide above reaction with a variety of Under certain conditions, the reaction of chrysotile asbestos with weak acids can be limited to the surface of the fibers. Pundsack and Reimschussel(57) demonstrated that fatty acids and other weak organic acids when dissolved in non-aqueous solvents react with the fiber surface to form a monomolecular "chemisorbed" layer. The amount of chemisorbed acid is af fected by the presence of sorbed water on the fiber. This is illustrated below. Chemisorption of Oleic Acid from Benzene by Chrysotile Mg, acid sorbed/g, fiber Fiber conditioned for 48 hr at 185C 29% RH | 100% RH 10.8 7.4 0.5 DPMC-05293 lam 008803 -16- Equilibrium in these reactions is established very rapidly probably within seconds. It was also noted that above a con centration of 0.005 moles of oleic acid/kg of benzene, the amount of acid chemisorbed by the fiber is constant and has a value of about 1 per cent by weight of the fiber. This value could vary with the exposed surface area of the fiber. Even after fifteen extractions with hot benzene, 80 per cent or more of the original oleic acid remains fixed on the fiber surface. Fiber which contains chemisorbed fatty acid is some what hydrophobic and markedly organophillic. The hydrophobic character of the fiber is of a limited nature, since the fiber can be wet by vigorous stirring in water. In general, organic compositions possessing acidic functional groups dissolved in non-polar or slightly polar solvents, such as benzene and methyl ethyl ketone, exhibit a strong tendency either to chemisorb or to slowly react with chrysotile. Long- chain aliphatic acids, such as stearic acid, oleic acid, and palmitic acid are chemisorbed by dry fiber. Aromatic-type acids, such as benzoic acid and related compounds, are also chemisorbed as are dibasic aliphatic acids, such as adipic acid. Although the unsaturated six-carbon sorbic acid ap pears to be chemisorbed by chrysotile, the related shorter carbon chain acrylic, crotonic acids show some evidence of slow reaction with the dry chrysotile even in non-polar sol vents and there is a tendency for the adsorbed layer to show an affinity for water. Maleic acid reacts with the bulk fi ber. If the fiber contains adsorbed water, the interaction of an organic acid in benzene or MEK solutions differs markedly from that with the dry fiber. The long chain aliphatic acids, e.g., stearic, oleic, when dissolved in non-polar or slightly polar solvents, show little or no tendency to sorb on fibers containing adsorbed water. Acids, which have some slight af finity for water, such as adipic, benzoic or sorbic, react with the bulk fiber structure instead of chemisorbing as they do on dry fibers. The reaction of chrysotile with certain anionic wetting agents, such as Aerosol OT, is peculiar in that they cause the fiber bundles to separate into ultimate fibrils. This re action is accompanied by strong chemisorption of the agents with permanent modification of the surface(51). When chrysotile fiber is decomposed by strong hydrochloric acid in the presence of chlorotrimethy1 silane(24) a very in teresting reaction occurs. Normally, the decomposition by acid leaves a residue of amorphous silica which polymerizes in DPMC-05294 LAM 008804 -17- the shape of the original fiber. The chlorotrimethyl silane' however, reacts with the silica sheet as the magnesium is re moved and prevents further polymerization of the silica. The final product of the reaction is in effect an organo-silicon polymer in sheet form. The sheets are rolled into hollow tubes and exhibit typical chrysotile morphology under the electron microscope. They also show a 15 A spacing between the layers by X-ray diffraction. In organic solvents, the polymer swells and the X-ray diffraction pattern disappears indicating that the tubular sheets have unrolled. This prod uct is interesting confirmation that the basic chrysotile structure is a spiral rather than a closed concentric cylin der. Reaction with Water. in addition to being vulnerable to at- tack by acids, and, under certain circumstances alkalies, as bestos fibers are also subject to attack by water. Pro longed extraction of chrysotile with water has been studied by Holt and Clark(31) and Reimschussel(60) . Holt reported that when chrysotile is extracted with boiling water, the solution contained both magnesium and orthosilic acid. He proposed that the chrysotile decomposed by loss of magnesium ions leaving a residue of colloidal silica. He also suggests that the colloidal silica is hydrolyzed to orthosilicic acid. Reimschussel1s results confirm that chrysotile is decomposed by water. He found that the concentration of magnesium in the extract was relatively high during the first three to four hours of Soxhlet extraction and then began to decrease. The decrease in magnesium concentration was accompanied by the formation of a precipitate of amorphous magnesium silicate. After the initial rapid reaction, magnesium and silica are removed in amounts proportional to the chrysotile composition. Whether the removal of silica proceeds by way of solution or by the formation of colloidal silica is yet to be determined. There is no doubt, however, that chrysotile is slowly "solu ble" in water under conditions of continuous extraction. For crocidolite, Thompson(68) reports that 4 per cent of the silica and 6 per cent of the sodium are removed by Soxhlet ex traction with water. In this case, the action was equivalent to that of alkalies at corresponding temperatures. SYNTHESIS Chrysotile. Numerous investigators have studied of chrysotile and serpentine minerals. These motivated either by pure theoretical interest or to grow large synthetic crystals. It is quite the synthesis studies were by the desire probable that DPMC-05295 LAM 008805 chrysotile was synthesized as early as 1927(34) but before the v advent of electron microscopy and sophisticated X-ray dif fraction techniques, it was not possible to distinguish among the minerals of the serpentine group. The mere fact that the chemical analysis of the reaction products matched that of serpentine could not be considered distinctly diagnostic for chrysotile. Chrysotile has only been synthesized under hydrothermal con- ditons. Bowen and Tuttle(10) demonstrated that chrysotile is formed at temperatures up to 500C and at pressures up to 40,000 psi. Yang(2), formed chrysotile at temperatures as low as 115C. Since a lower limit of temperature stability for chrysotile has not been established, it is conceivable that it could be formed in an aqueous system with the correct composition under ambient conditions, particularly within geo logic time periods. There are several interesting features which have been noted for synthetic chrysotile. First, regardless of the conditions employed, or the trace impurities and mineralizers which have been added, only single chrysotile fibrils have been synthe sized. No method has been found for synthesizing the large bundles of parallel fibers which occur in nature. Secondly, if the product formed during the early stages of synthesis is examined under the electron microscope, one can see evidence of sheets in the process of rolling into tubes. Roy and Roy(63) carried out an extensive study of the synthe sis of serpentine minerals with magnesium and silicon substi tuted wholly or partially by other ions. Replacement of mag nesium by similar-sized nickel ions yielded a product which was either platy or tubular, depending on other factors, such as the presence of sodium chloride in the reaction mixture . Increasing the size of the tetrahedral layer by substituting silicon with germanium resulted in a serpentine mineral which formed large, platy, hexagonal crystals. Similarly, partial substitution of aluminum in both layers yielded platy "alu minum serpentine" (Mg5Al) (AlSi303Q) (OH) 4. Another ser pentine phase was also synthesized from nickel and germanium Ni3Ge205 (OH)2 which also had a platy structure. It was not possible to produce serpentine-type phases substituting man ganese, zinc, cobalt, iron, chromium or gallium for magnesium. This, however, does not preclude the possibility of trace amounts of these ions being present in natural or synthetic materials. The authors concluded that the tubular structure of chrysotile is not only a consequence of the ion sizes, but other external influences as well. DPMC-05296 LAM 008806 -19- Amphiboles. The amphibole minerals can be synthesized by' either pyrogenic or hydrothermal methods. Recent studies at the Institute for Silicate Chemistry, Leningrad, are probably the most comprehensive in this field. Amphiboles containing fluorine substituted for hydroxyl were produced by heating a mixture of oxides and fluorides to 900-1100C in tightly closed platinum or ceramic vessels<26) . The mineral phase and the morphological characteristics of the products depended on the composition of the initial mixtures, the fluorine content and the temperature. The isomorphous series of fibrous fluoro-amphiboles included a wide variety of cations com prising a mixture of Na or other mono or divalent metal plus either Mg++, Fe++, Cr+++, Cu++, Co++, Ni++, Mn++, or Cd++. Within the reaction mix, crystals of 0.5 to 1.0 millimeters in length, and 0.1 to 2.0 microns in diameter were formed with crystals as long as 20 millimeters on the surface of the reac tion mass. The physical properties of the synthetic fluoro- amphiboles were stated to be superior to the best natural va rieties with tensile strength of 20,000 to 40,000 kg/cm2. Thermal decomposition did not occur until the temperature exceeded 900C. Hydroxy amphiboles containing combinations of Na or Ca plus either Mg++, Fe++, Co++ or Ni++ were formed under hydrothermal conditions(39). Typical reaction conditions were 350-600OC with pressures ranging from 300 to 2000 atmospheres. After reaction times of 6 hours to 3 days, fibers as long as 4 mil limeters were obtained with diameters ranging from 0.1 to 1.0 micron. The hydrothermal fibers were too small to measure their mechanical properties. Their thermal stability was of the same order of magnitude as the natural materials. PHYSICAL PROPERTIES Tensile Strength. Asbestos is used primarily as a reinforcing fiber. Its tensile strength is, therefore, of prime signifi cance. The measurement of tensile properties is complicated by the combination of short length and small diameters re sulting in a wide range of values from the same type of fiber as reported by Badollet^^ . Zukowski and Gaze'^^1 showed a strong dependence of strength on fiber length with maximum values of 61,000 kg/cm2 and 58,000 kg/cm2 for crocidolite and chrysotile, respectively, with a fiber length of approximately 2 mm. Motion pictures showed that fiber failure normally oc curred by rupture of weak interfibrillar bonds, rather than true tensile failure of a fiber. Using a newly-developed micro-tensi ]e machine, comparable data were obtained by Burman on a variety of asbestos fibers 4 mm long and 10 to 20 microns in diameter(33). Crocidolite and chrysotile had the greatest DPMC-05297 LAM 008807 i-rTT-r HrpMWivrt fs^n,, - -m -. -,- |- J --- -.2. V0< strength followed closely by amosite, with the other amphi- boles significantly weaker. More recent work by Burmand^) in Table 5 casts doubt on the validity of assigning specific tensile strength values to each asbestiform mineral, although the same relative order is maintained. All fibers appear to have strengths less than the theoretical value of over 100,000 kg/cm^ attributable to silicate chain structures. The fact that chrysotile and some of the amphiboles give so nearly the same values has led Whittaker(77) to suggest that the fiber strength is affected more by the crystal imperfections intro duced during fiber formation rather than by the atomic ar rangement or structure of the fiber. This is consistent with the hypothesis that amphibole fibers are built up of overlap ping crystallites held together by H-O-H coordinate linkages or other cohesive forces. Chrysotile fibers, on the other hand, are bundles of fundamental fibrils of relatively con stant diameter, but varying length. The difference between different chrysotiles is best exemplified by Coalinga fiber (Figure 15) where most fibers consist of overlapping fibrils approximately 0.5-2 long cemented together, and Jeffrey chrysotile which has long individual fibrils (Figure 16). Such a concept can also explain the gradual time-dependent loss in strength of asbestos fibers with increasing tempera ture below the decomposition temperature of the crystal. Small losses of water in the early stages of dehydration for both chrysotile and amphiboles are probably associated with reduced edgewise bonding between the crystallites or fibrils comprising the asbestos fibers, resulting in reduced "tensile strength" of the fiber. Harshness. The term harshness is related to`the flexural modulus of the asbestiform fibers. The high modulus amphi boles are all generally harsh and relatively stiff even in the finest fibers shown in the electron micrograph of Figure 17. Although most chrysotile fibers are soft, semi-harsh fibers are also commercially available. Harsh chrysotile is not us ually commercially significant. The difference in the appear ance of fiber masses is well illustrated by the photomicro graphs of Figures 18 and 19. Harsh fibers yield an open, bulky, fast-filtering, mass. The flexible soft fibers form stringy, dense masses with slow fil tration characteristics. This specific attribute of soft chrysotile is often a serious disadvantage in wet processing techniques employed in the manufacture of asbestos-cement products. Semi-harsh chrysotile fibers, when economically available, or crocidolite can partially replace soft chryso- DPMC-05298 LAM 008808 21'- - V tile to improve filtration. Chemical techniques mayalso be used to accomplish the same objective either by adding poly electrolytes (1) or by treating the fiber with sodium sili cate^) . Considerable research has been devoted to correlate harshness with fundamental physical or chemical factors. Woodroofe(80) has indicated a relationship with the water content of the fiber. In line with this, Badollet and StreibM) have pa tented a technique for increasing the harshness of chrysotile by flash calcining in the range of 500C to drive off part of the chemically combined water. More recently, hanger and Kerr(37) have indicated a correlation with fine mineral in tergrowths in the bundles of fibrils comprising chrysotile. Another hypothesis is that harshness is related to the rela tive contents of the two crystallographic forms, clino-chryso- tile and ortho-chrysotile in the fiber from a source (Whittaker and Zussman (77)). specific Thermal Decomposition. Asbestos minerals, despite their rela tively high fusion temperature, are completely decomposed at temperatures of 1000C or lower, depending on the mineral species. The course of the thermal decomposition can be followed by three different but inter-related techniques: differential thermal analysis (DTA), thermogravimetric analysis (TGA), and static dehydration. The actual decomposition phenomena in amphiboles are extreme ly complicated and depend on the type of atmosphere and par ticularly on the specific amphibole involved. Typical curves for the behavior of crocidolite in air are given in Figure 20 from Hodgson(33), The first chemical change, corresponding to an apparent loss of water, occurs at 420C with the final decomposition of the amphibole into a pyroxene mineral, cris- tobalite, and. iron oxide occurring at 900C. This water is formed by migration of protons which are oxidized at the sur face by oxygen in the air. The other amphiboles lose water primarily by condensation of hydroxyl ions. Oxidation of divalent iron has a profound ef fect on the thermal behavior of these minerals and this be comes very evident by comparing the reactions in air and in an inert atmosphere. Both the dehvdroxylation temperature and decomposition temperature appear to increase with increased MgO content in the different amphibole species. Hodgson(32) gives a detailed experimental and theoretical review of these DPMC-05299 LAM 008809 Ha itip -22- phenomena, which include loss of physically combined water, loss of chemically combined water, and breakdown into ultimate decomposition products. These thermal analysis techniques have proven particularly useful in categorizing amphibole sam ples from different locales. For example, Bolivian crocido- lite exhibits characteristics of both anthophyllite and cro- cidolite, while the Transvaal species is often an intimate mixture of crocidolite and amosite. v The decomposition of chrysotile is much simpler and indepen dent of atmosphere. Under dynamic heating conditions (DTA), dehydroxylation occurs at approximately 650C and formation of forsterite and silica, about 810C, as shown in Figure 21. In static dehydration experiments, the initial water loss below 500C is time-dependent with no detectable change in the X-ray diffraction pattern (12, 80). Martinez(4U has summarized the various theories for the atomic rearrangement during the de hydration and formation of forsterite. The simplest approach to describing the decomposition above 500C (static) is to consider it as a three-step process: the solid residue of the first step 600C) is a slightly hydrated amorphous mag nesium silicate with a minor amount of poorly crystallized forsterite. This is followed by the formation of well-cry stallized forsterite with some residual amorphous material at 600 to 1000C. Finally, heating above 1100C yields a mixture of enstatite and forsterite. This generalized scheme has been confirmed by comparisons of infrared spectra of heated chrysotile with those of mixtures of pure synthetic minerals(61). Many investigators have studied the thermal decomposition of chrysotile from a wide variety of sources, and all have been found to yield essentially the same DTA curves. Recent studies with an extremely sensitive duPont Differential Ther mal Analyzer have disclosed significant differences between fiber from different sources. For many chrysotiles, the 650 dehydroxylation peak is really a doublet'-^". This doublet phenomenon was first observed by Martinez(40) in a mixture of two samples of the same fiber which had been subjected to dif ferent degrees of intensive grinding. The dependence of dehy droxylation temperature on particle size in the analogous min eral kaolinite was first reported by Speil(66). it has not yet been determined whether these recently observed doublet peaks in specific chrysotiles correspond to two fiber diameter populations or to some other phenomenon. DPMC-05300 LAM 008810 ITHTTiii mum rtfcMMwaiai -23- Mechanical Disintegration. For effective reinforcement, the asbestiform mineral should be fiberized to the degree required by the specific application. Mechanical milling or attrition is the basic method of fiberizing asbestos minerals. Ideally, the fiber bundles should be opened without reducing the fiber length. In practice, the fibers are shortened to a degree controlled not only by the severity of the mechanical action, but even more, by the brittleness or harshness of the mineral. The soft chrysotiles show minimum length disintegration during opening while, for the same mechanical attrition, the semi- harsh and harsh chrysotiles are shortened significantly. Amphiboles are even more susceptible to length attrition by mechanical impact, and are usually given their final opening by the ultimate consumer, often in the actual mixing or pro cessing operations. Normally as the fibers become more open, the additional energy required for further opening increases rapidly, imposing a practical limit on the degree of subdivision attainable in commercial milled products. Recently, laboratory grinding tests with an intensive dry grinder* showed that it is pos sible to actually destroy the structure of chrysotile so that it is no longer identifiable by either X-ray diffraction or by the electron micrograph of Figure 22^0). Jeffrey chrysotile heated to 700C for 1 hour yields an amorphous material with exactly the same appearance when viewed by the electron micro scope. This suggests that the changes observed with intensive grinding were actually caused by momentary localized temper ature surges in a fibril as it absorbed the tremendous impact energy. To substantiate this hypothesis, chrysotile was sub jected to prolonged dry ball milling which yielded a similar appearing amorphous mass. Wet milling, which precluded the possibility of attaining localized high temperatures, produced short ultimate fibrils which maintained their crystalline form and were easily identifiable as chrysotile. These observations have been used to explain the results of controlled brake-wear tests performed at the Johns-Manvi1le Research and Engineering Center(^5). Wear dust was collected from passenger car brake linings subjected to a series of stops simulating normal traffic and highway driving. Air borne dust and debris in the brake drums were collected sepa rately and analyzed completely. The composition of the inor ganic fraction of these residues matched that of the original brake lining which contained a total of 70 per cent chrys otile. No chrysotile was observable in the wear dust by *Spex Mixer/Mill, manufactured by Spex Industries, Inc., 3800 Park Avenue, Metuchen, New Jersey 08840 DPMC-05301 LAM 008811 -24- either X-ray or optical microscopy. An electron micrograph of the dust is shown in Figure 23. The resemblance between this and chrysotile, which had been mechanically (and, presumably, thermally) decomposed is apparent. Despite the fact that thermocouple measurements of the brake drum temperature did not exceed 275C, theoretical calculations indicated that the temperatures at localized points of contact exceeded 1000C. Minute fragments of chrysotile fibrils could be observed in some of the electron micrographs and a modified point count technique was devised to determine the chrysotile content. Conservative estimates of the chrysotile content by this me thod showed that more than 98 per cent of the chrysotile is completely destroyed during normal brake usage. Similar stud ies by the U. S. Public Health Service substantiated these re sults for normal driving conditions(38). v IDENTIFICATION The positive identification of the asbestiform minerals is highly dependent upon such factors as the physical form, the presence of contaminants, prior mechanical, thermal or chem ical treatment, etc. The variable nature of the amphiboles makes it necessary to classify a particular sample as "most closely resembling" a specific mineral species. It is obvious that, even if sufficient material is available, simple chemical analysis is not sufficient to characterize an asbestiform mineral because all asbestiform minerals have mas sive counterparts with the same chemical composition. One must, therefore, rely on a combination of methods. The methods most frequently used are chemical analysis, petro graphic microscopy, X-ray diffraction, electron microscopy, electron diffraction, and differential thermal analysis. Perhaps the most reliable method of identification for par ticles down to a few microns in diameter is petrographic microscopy. Using this technique, the fibrous nature of the species is evident and the optical properties can be used to determine which species is present. Among the amphiboles however, the variable composition may also effect the optical properties, so that again positive identification may be dif ficult. In such cases, X-ray diffraction, and chemical anal ysis, or DTA can be used for confirmation. If the size is below the practical working limit for optical microscopy, the electron microscope is the proper tool. This instrument will only give information about the size and shape so that positive identification is not usually possible for DPMC-05302 LAM 008812 y A-y -25- specific amphiboles. The tubular appearance 'of chrysoti?le under the electron microscope is specific. When electron dif fraction is used in conjunction with electron microscopy a better, but still not conclusive, identification of amphiboles is possible. The electron microprobe is a relatively recent development which is proving of great value for the characterization of small amounts of material. With this instrument it is possible to obtain a complete chemical analysis on a particle as small as one micron. It is also possible to study the same parti cles with both the electron microprobe and the electron micro scope, thus making a more complete characterization possible. Table 6 is a listing of the properties of the asbestiform min erals used for their characterization. SOURCES OF FIBER Hendry(30) at the 1965 Conference on Biologic Effects of As bestos summarized the pertinent aspects of the occurrence, production and commercial applications of asbestos fiber. Table 7 presents the world production of asbestos in 1966 ^3). Production has increased with few changes in the relative standings of the producing countries. The major difference is the great increase in Russian pro duction which now outranks Canada as the major producer. The increased Droduction in the United States over the past few years is due to the recent development of the Coalinga fiber deposits in California. One important change since 1966 is the closing of the Australian crocidolite mines and the elim ination of this source from the market. Thus, South Africa remains as the one significant area producing crocidolite and amosite. Chrysotile accounts for approximately 95 per cent of commer cial asbestos. Chrysotile asbestos from Quebec is available in more than 50 standard or specialized grades to meet spe cific requirements. Some of the larger asbestos mills, such as that of the Jeffrey mine in Asbestos, Quebec, nroduce many of these grades simultaneously by a complex system of conti nuous crushing, screening, and aspiration from the same mill feed material. DPMC-05303 LAM 008813 -26- All fibers from Quebec are classified by a standardized system which, with some modifications, forms the basis for other classification systems in use throughout the world. The long est fibers are Groups 1 and 2, with fiber lengths of over 3/4 in. and from 3/8 to 3/4 in., respectively. These consist of hand-selected cross-vein fiber and are termed "crude" asbestos grades since they are normally given final preparation by the ultimate user. Groups 3 through 7 are classified as "milled" fiber with decreasing fiber length, as measured by the Quebec Standard Screen Test(59)_ Each group is further sub-divided into a number of sub-grades according to their "crudiness", content of grit, bulking characteristics, and absorption pro perties. Subgroup fibers from different sources are not ne cessarily interchangeable for specific end applications. In the asbestos trade, a "crudy" fiber is one which contains a large number of unopened fiber bundles; those fiber subgrades in which the fibers have been well fiberized or subdivided are known as "opened" grades. ' Table 8 gives the approximate distribution by grades of chrysotile fiber produced in Canada and the United States, as com pared to that produced in the U. S. S. R. in 1966. TABLE 8. APPROXIMATE PRODUCTION OF CHRYSOTILE FIBERS BY GRADES IN 1966 Grade #1 Crude #2 Crude 3 4 5 6 7 U. S. S. R. 18,000 130,000 180,000 650,000 390,000 530,000 U. S. and Canada 200 45,000 500,000 200,000 230,000 620,000 DPMC-05304 LAM 008814 .MfcStMw y-i ii^wai m mmamnstm wmmm -27- Applications. The commercial applications of asbestos are so numerous that this review can do no better than refer the reader to the many excellent texts which adequately cover this subject^, 62) ^ The uses range from asbestos-cement products or floor tile which consume hundreds of thousands of tons, to specialty filtration applications which may consume only sev eral tons annually. Asbestos imparts to a great variety of products a combination of properties which cannot be attained by using other materials. Its strong fibrous form reinforces other media, such as plastics or cement, or controls viscosity of many systems; its inorganic nature is important for resis tance to heat and chemical or environmental agents; its fine size contributes filtration efficiency and insulating effi ciency; its abundance and low cost is a significant factor in promoting commercial applications. On the basis of relative abundance alone, chrysotile will be used wherever possible in preference to other forms of as bestos. Where a combination of extreme bulking characteris tics plus low water content and high temperature resistance are desirable, e.g. , thermal insulations, amosite has usually been preferred. Applications requiring resistance to acids usually take advantage of the relatively good acid resistance of crocidolite. Textile products require a soft, silky, long er grade of chrysotile although crocidolite has also been used for acid resistant textile forms. Geographic considerations may exert economic influence to increase the usage of specific amphiboles, such as anthophyllite in Finland, or the crocido- lites in South Africa. Anthophyllite has shown specific ad vantages over other asbestiform minerals in reinforcing poly-, propylene products and is used almost exclusively. In general, specific grades of chrysotile have been developed by the asbestos industry for each market. Although the longer fibers are considered to be of better quality, it would be just as impractical to use relatively long 4-grade fiber in floor tile as it would be to attempt to make satisfactory as bestos paper or asbestos cement products with 7-grade fibers. In fact*, some applications may even require the presence of a considerable amount of the non-fibrous, fine grained particu late serpentine which is contained in some of the 7-grade subgroups. Normally, the consumer selects the least expensive grade of fiber which will meet his needs. DPMC-05305 LAM 008815 I I'jftdHMWHii flilfM -28- Other Sources. Commercial production and applications of as- bestos fibers are usually considered to be the only signifi cant sources of asbestos fibers entering the environment. However, on closer examination it becomes evident that we should consider not only the question of impurities in asbes tos, but also the fact that asbestiform minerals are an ubi quitous impurity in many deposits of commercially valuable nonmetallic minerals, such as mica and talc. Talc particularly is a mineral product with widespread commer cial and cosmetic applications. Figure 24, an electronmicro- graph of a typical beneficiated industrial talc, reveals the presence of considerable fibrous tremolite. Approximately 8,000 tons of talc are used annually as a carrier for pesti cides ('D . Windom, et al^9), investigated the distribution of talc in the atmosphere and in glacier and snow samples to study the migration of pesticides. Their samples covered a world-wide geographic distribution. In practically every sam ple, amphiboles were detected along with the talc, as might be anticipated from the common occurrence of amphiboles in talc. Cralley), et al, have recently investigated twenty-two cos- f ymetic talcum products. All had significant fiber contents \ Y' ranging from 8 to 30 per cent by count of the total talc par- k Hticfnuialat-toecs., and averaging 1199 ppeerr cceenntt., The fibrous talc in- eluded tremolite, anthophyHite, and chry"sotile. They macle special note of"tire tact that COsme-tirc--tatchm,.products--should be included ~as~aT6urce of fibers from which may be derived ferruginous bodies obsexved -in--the., lun.gs_.of. humans .------------- In another investigation by the U. S. Public Health Service of the source and identification of respirable fibers, Cral ley (15) noted that there are more than one hundred different natural minerals with some degree of fibrous structure which may occur in respirable sizes. In addition to the asbestos minerals, these included fuller's earth, zeolite, vermiculj.te, calcium carbonate, gypsum, pyrophyllite, talc, kyanite, horn- \blende, mica, magnesite, and many others. Pure serpentine is considered to be composed of non-fibrous antigorite or lizardite based on petrographic and X-ray exami nation. However, electron microscopy reveals the fact that all serpentine rocks contain significant amounts of chryso tile. Figure 25 is an electronmicrograph of a practically translucent "museum grade" serpentine specimen from Warren County, New York, obtained through Wards Natural Science Es tablishment. Despite its apparent content of approximately 20 per cent of fibers, optical microscopy showed no chrysotile whatsoever. DPMC-05306 LAM 008816 i -29- Examination of many other authenticated samples of "pure ser` pentine" has revealed the presence of chrysotile. Serpentine rock deposits are widespread throughout the world. in the U. S., they form the Franciscan serpentine belt along the en tire length of California, just as they form much of the Ap palachian range on the East Coast. They are used as the ba sis for many large-scale applications, such as ballast, road construction, aggregate, building stone, etc. During grinding and preparation for such commercial usage, there could be op portunity for escape of fibrous material. CONCLUSIONS It is important for the medical investigator of the biologic effects of asbestiform minerals to properly understand the wide diversity between the several individual asbestos mine rals, the ubiquitous nature of their occurrence, both in com mercially valuable form and as impurities in other materials, and the widespread existence of many other minerals with fi brous form. It is only by relating experimental biologic evi dence with the variations in physical size and form, in phy sical strength attributes, in physico-chemical surface reac tions, in chemical reactivity, and in associated impurities, that we can ultimately arrive at valid medical conclusions. ACKNOWLEDGEMENT The authors wish to express their gratitude to J. W. Axelson, D. A. Bailey, G. P. Reimschussel, and W. C. Streib of the Johns-Manville Research and Engineering Center for their in valuable assistance during the preparation of this paper. DPMC-05307 LAM 008817 1. Badollet, M. S., U. S. Patent 2,068,219. 2. Badollet, M. S., Can. Min and Met. Bull., April 1951, p 1. 3. Badollet, M. S., Encyc. Chem. Tech., Vol. 2, p 734 (1963). 4. Badollet, M. S. and Streib, W. C., U.S. Patent 2,616,801. 5. Ball, M. C. and Taylor, H. F. W. , Min. Mag. 12, 754 (1961); 33, 467 (1963). 6. Ball, M. C. and Taylor, H. F. W. , J. App. Chem. 13_, 145 (1963). 7. Bates, T. F. , Sand, L. P., and Mink, J. F., Science 111, 512 (1950). 8. Berger, H., Asbestos Fundamentals, Chem. Pub. Co., New York (1963). 9. Berman, H., Amer. Mineral, 7, 313 (1932). 10. Bowen, N. L. and Tuttle, 0. F. , Bull. Geol. Soc. Am. 6() (3) , 439 (1949) . 11. Brindley, G. W. and Hayami, R., Min. Mag. 5, 189 (1965). 12. Brindley, G. W. and Zussman, J., Amer. Mineral. 2, 461 (1957). 13. Burman, D. R., Paper No. 2-8, Oxford Conference on the Phy sics and Chemistry of Asbestos Minerals, 1967. 14. Cralley, L. J., et. al. To be presented at Amer. Ind. Hyg. Conference, St. Louis, Mo., flay 13, 196 8. 15. Cralley, L. J., et. al, To be published in Amer. Ind. Hyg. J. (1968). 16. Clark, F. W. and Schneider, E. A., Am. J. Sci. 3rd Ser. 40, pp 303, 405, 452 (1890). 17. Deer, W. A., Howie, R. A., and Zussman, J. "Rock Forming Minerals" Vol. 2 and 3, John Wiley & Sons, 1962. 18. Della Faille, M., Ph. D. Thesis, University of Louvain (1965) . 19. Faessler, C., and Badollet, M. S., Can. Min. Jour. 68, No. 3, 157 ( 1947) . DPMC-05308 LAM 008818 ,:A\| mmm 20. Faust, G. T. and Fahey, J. J., U.S. Geol. Survey Prof. Paper 384-A (1962) . 21. Faust, G. T. and Nagy, B. , Am. Mineral. A_L, 817 (1956). * 22. Faust, G. T. and Nagy, B., U.S. Geol. Survey Prof. Paper 384-B (1967). 23. Frankel, J. J., Mining Mag. 89_, 73 , 143 (1953). 24. Frazier, S. E., Bedford, J. A., Hower, J., and Kenny, M. E., Inorganic Chem. , 1693 (1967). 25. Gaze, R., Annals N.Y. Acad. Sci. 132, pp 23-30 (1965). 26. Gregorjeva, R. F., et. al. Paper No. 3-6, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 27. Hall, A. L. , Union S.A. Geol. Survey Mem. 12, 2nd Ed. p 324 (1930) . 28. Ilarington, J. S., Annals New York Acad. Sci. 132, 31-47 (1965) . 29. Harris, M. R. , Chem. and Ind., Feb. 6, 1965, p. 268. 30. Hendry, N. W., Annals New York Acad. Sci 132, Art. 1, pp 12-22 (1965). 31. Holt, P. F. and Clark, S. G., Nature, London 185, 237 (1960) . 32. Hodgson, A. A., Min. Mag. 3^, 291 (1965). V 33. Hodgson, A. A., "Fibrous Silicates" Royal Inst, of Chem. Lecture Series No. 4, 1965. 34. Ipatiev, W., and Mouromstev, B. A., Compt. rend. 185, 647 (1927). 35. Jaunarajs, K. L. , Unpublished information, Johns-Manvi1le Research & Engineering Center, Manville, New Jersey. 36. Kjpling, J. J. and Tester, D. A., J. Chem Soc. 1952, 4 123 . 37. Langer, A. and Kerr, Paper No. 2-2, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 38. Lynch, J. R., Paper submitted to J. Air Poll. Cont. Assn. 39. Makarova, T. A., et. al, Pager No. 3-7, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. DPMC-05309 LAM 008819 40 . Martinez, E. , Amer. Mineral., 46 , 901 (1961). 41. Martinez, E. , Trans. Can. Min. & Metal Bull. 63, 1305 (1966) 42. Martinez, E. , and Zucker, G L* r J. Phys. Chem. 64, 924 (1960). 43. May, T. C., Min. Yearbook, U. S. Dept, of Interior, 1966, p 426. 44. Maser, M., Rice, R. V., and Klug, H. P., Amer. Mineral. 45, 680 (1960). 45. Mumpton, F. A., Jaffe, H. W., and Thompson, C. S., Amer. Mineral. 50^, 1893 (1965). 46. Nagy, B., Econ. Geology. 8, 591 (1953). 47. Nagy, B. , and Bates, T. F., Am. Mineral. 3_7, 1055 (1952). 48. Nauman, A. W., and Dresher, W. H., Amer. Mineral., 51, 711, (1966). 49. Nickel, E. H., Canadian Mineral. , 307 (1959). 50. Noll, W. , and Kercher, H., Naturwiss. 31_, 540 (1950). 51. Novak, 1. J., U. S. Patent 1,967,062. 52. Patterson, J. H., and Thompson, R. L., Paper No. 2-5, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 53. Pundsack, F. L. , J. Phys. Chem. 59, 892 (1955). 54. Pundsack, F. L. , J. Phys. Chem. , 5 , 30 (1961). 55. Pundsack, F. L. , J. Phys. Chem. G0_, 361 (1956). 56. Pundsack , F. L., and Reimschussel, G. P- , J. Phys. Chem 1218 (1956). 57. Pundsack, F. L., and Reimschusse1, G. P. , U. S. Patent 3,304,197. 58. Pundsack, F. L., and Reimschussel, G. P. , U . S. Patent 3,173,831. 59. Testing Procedure for Chrysotile Asbestos Fiber, Quebec, Asb. Mining Assn., 2nd Edition (1966). 60. Reimschussel, G. P., Unpublished information, Johns-Manvi1le Research & Engineering Center, Manville, New Jersey. DPMC-05310 LAM 008820 `. V 61. 62. Reimschussel, G. P., and Wegrzyn, S. mation, Johns-Manville Research & Manville, New Jersey. w. , Unpublished inforEngineering Center, `fc Rosato, D. V. , Asbestos, Its. Industrial Applications, Reinhold, New York (1959). < 63. Roy, R. and Roy, D. M., Amer. Mineral. 39, 957 (1954). 64. Ryznar, J. W., Greon, J., and Winterstein, M. G., Ind. Eng. Chem, 38, 1057 (1946). 65. Sinclair, D., Unpublished information, Johns-Manville Research & Engineering Center, Manville New Jersey. 66. Speil, S., et. al, U. S. Bureau Mines, Tech. Paper 664 (1945). 67. Terreil, A., Acad. Sci. (Paris) Comptes Rendus 100, 251 (1885). 68. Thompson, R. L., Paper 1-8, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 69. Thompson, R. L. , Paper 1-9, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 70. Turkevich, J., and Hillier, J. , Anal. Chem. 21, 475 (1949). 71. U.S. Dept, of Agriculture, Agriculture and Conservation Service, The Pesticide Review, 1966. 72. Warren, B. E., and Bragg, W. L., Z. Krystallog. 76, 201 (1930) . 73. Warren, B. E., and Herring, K. W., Phys. Revs. 59, 925 (1941). 74. Weeks, T. J., Unpublished information, Johns-Manville Research and Engineering Center, Manville, New Jersey. 75 . Weeks, T. J., and Leineweber, J. P., Paper No. 2-7, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 76 . Whittaker, E. J. W., Acta Cryst. 6, 747 (1953); 8, 571 (1955); 9, 855, 862, 865 (1956); 10, 149 (1957). 77. Whittaker, E. J. W., Paper No. 1-1, Oxford Conference on the Physics and Chemistry of Asbestos Minerals, 1967. 78. Whittaker, E. J. W., and Zussman, J., Min. Mag. 3^, 107 (1956). DPMC-05311 LAM 008821 r*Zg? m ---- i -m-r-y -r-r^Tiitirr ^ 79. Windom. H., et. al v>fE$viron. Sci. & Tech. 1_, 923 (1967). 80. Woodroofe, H. M. , TrAns . Can. Inst. Min. & Met. 59 , 363 (1956). r v 81. Yada, K. , Acta Cryst.. 3, 704 (1967). 82. Yang, J. C. , J. Amer.' Ceram. Soc. , 3, 542 (1960). 83. Yang, J. C., Unpublished information, Johns-Manvilie Research and Engineering Center, Manville, New Jersey. 84. Young. G. J., and Healv, F. H., J. Phys. Chem. 5, 881 (1954). 85. Zettlemoyer, A. C., et. al, J. Phys. Chem. 7, 649 (1953). 86. Zukowski, R., and Gaze, R., Nature, London 183 , 35 (1959). DPMC-05312 LAM 008822 TABLE 1 . CHEMICAL COMPOSITION OF CHRYSOTILES a 4 TABLE 2. IONIC RADII FOR ELEMENTS ASSOCIATED WITH CHRYSOTILE SCi ++ + + Mg++ Al+++ Fe++ Fe+++ O.Ul Angstroms 0.65 0.50 0.76 0.6U Ni++ Co++ Ge++++ Mn++ Cu++ 0.78 Angstroms 0.78 0.53 0.80 0.69 TABLE 3. CHEMICAL COMPOSITION OF ASBESTIFORM AMPHIBOLES Typical Ranges in Per Cent Si02 MgO FeO Fe^-0-3 A]23 CaO K')0 Na20 h2 Crocidolite ^9-53 0-3 13-20 17-20 0-0.2 0.3-2.7 0-0 L.0-8.5 2.5-1' 5 Amosite **9-53 1-7 3U-lll4 -- -- -- 0-0. J* tr 2.5-**. 5 Anthophyllite 56-58 28- 3** 3-1? -- 0.5-1.5 -- -- -- 1.0-6-0 Actinolite 51-56 15-20 5-15 -0-3 1.5-3 10-12 0-0. 5 0.5-1.5 1.5-2.5 Tremolite 55-60 21-26 o-fc 0-0.5 C-2.5 11-13 0-0.6 C-1.5 0.5-2.5 DPMC-05314 LAM 008824 rntmUmMg &%* i l --TABLE 1*. SOLUBILITY OF ASBESTOS MINERALS IN 25? ACID OR CAUSTIC % Per Cent Loss in Weight, Re-fluxing Two Hours Chrysotile Crocidolite Amosite Anthophyllite Actinolite Tremolite HC1 55-69 L. 38 12.8L 2.66 20.31 l*-77 CH3COOH 23.1*2 O.91 2.63 0.60 12.28 1.99 H3PO4 55.18 1*.37 11.67 3.16 20.19 *.99 H2S0i, 55.75 3.69 11.35 2.73 20.38 L.58 NaOH 0.99 1.35 6.97 1.22 9-25 1.80 Reproduced with permission from Canadian Mining and Metallurgical Bulletin, April 1951- DPMC-05315 LAM 008825 TABLE 5- PHYSICAL PROPERTIES OF ASBESTIFORM MINERALS Ore Samples - Tensile Strength xlO^, kg/cm2 Youngs Modulus xl0 kg/cm2 Average Cross- Sectional Area of Fibers Tested x'lO"5 cm2 Chrysotile, Arizona, U.S.A. Chrysotile, Thetford, Canada Crocidolite, Koegas Cape Province Crocidolite, Koegas Cape Province Crocidolite, Pomfret Cape Province Crocidolite, Pomfret Cape Province Crocidolite, Cochabambo Bolivia Amosite, Penge Transvaal Amosite, Penge Transvaal Anthophyllite, Paakilla Finland 38.5 37.1 29-0 31-5 1*7.5 36.2 lL .7 26.3 20.2 25-0 1.1*8 1.1*9 1.50 1.5l* 1.72 1.78 1.73 1.1*6 1.1*6 1.59 2.07 2.13 1 * I.65 1.33 1.6U 1.31* 2.1*3 2.71 1.83 0.95 DPMC-05316 LAM 008826 'miH* C.-O3 w .'.M25 o .M s .ww<m tt, o aow <toC MCC W (OH < u PS Q WCO p CO Mw PS wa. oCCUC C>HO XCl, VO Wt-3 % DPMC-05317 LAM 008827 TABLE 7. PRODUCTION OF ASBESTOS IN 1966 SHORT TONS North America: Canada (sales) United States (shipments) South America: Argentina Bolivia (exports) Brazil Europe: Austria Bulgaria Finland France Greece Italy Portugal U. S. S. R. Yugoslavia Africa: Botswana Kenya Mozambique Rhodesia, Southern South Africa, Republic of Swaziland United Arab Republic (Egypt) Asia: China Cyprus India J apan Korea, South Phillipines Taiwan Turkey Oceania: Australia New Zealand World Totale 1,1*79,281 125,928 2bOe h 1,820 -- 1,1) 30e 13,250 7,720e 85e 90,1)61) 10 1,872,000(1) 8,1)11 880e 73 -- 175,000e 276,597 36,11)2 2,057 ll)0,000e 21), Ul)9 7,61)6 17,067 687 -- 721 1,258 13,1)72 -- 1) ,297,000 eEstimate (-^Unpublished data considered to be reliable (not from MINERALS YEARBOOK) DPMC-05318 LAM 008828 mad. JK` 4 % \1 inch FIGURE 1 CROSS VEIN FIBER CRKYSOTTLE DPMC-05319 LAM 008829 S'.,. .? 1 inch V FIGURE 2 SLIP FIBER CHKYSOTILE DPMC-05320 LAM 008830 1 Inch y FIGURE 3 NODULES OF COALIMJA CHRYSOTILE DPMC-05321 LAM 008831 1 inch j FIGURE 4 CROCIDOLITE IN BANDED IRONSTONE DPMC-05322 LAM 008832 rreaims* WTTl'iBTnTMWiffm.i. jsmtmm F ig u re 5 CD C r~ 0 1 c ~0 o ~n co X m m CO H O a > m H > r~ 31 OD 5 r~ co LAM 008833 TTr' \\ Vfflfi iWMWIB \ i.. t------------------------------------------------------------ ------------------------ 0.5 micron FIGURE 6 ELECTRON MICROGRAPH OF CHKISOTILE CROSS SECTION 195,OOOX REPRINTED WITH PERMISSION FROM AMERICAN MINERALOGIST 4, 680 (i960) DPMC-05324 LAM 008834 w ,n * \\ CO amm '< ig u re 5 u_ DPMC-05325 LAM 008835 FIGURE T FIGURE 8 HIGH RESOLUTION ELECTRON MICROGRAPHS OF CHRYSOTILE DPMC-05326 REPRINTED WITH PERMISSION FROM ACTA CKYST. 2^_, '{Oh (1967) LAM 008836 .f.: ' -K' i V% '*;1 Hy t' /. . - ^ 21 i nvaaii. <-i.g.ii itiir rnrWi.**-- >->n : iHMuatatimSim in- AMPHIBOLE STRUCTURE FIGURE < DPMC-05327 LAM 008837 <>d y0) ID 5La) <0 T <0 H t/) o DPMC-05328 LAM 008838 S o rp tio n o f o rg a n ic liq u id s o n c h ry s o t i le CL) l_ D 01 DPMC-05329 LAM 008839 rnm .*V: V^r \ ,t > '' mmrn Fi gu re 12 LAM 008840 uojjjsoduuoosp /Q (tim e - m in ) F ig u re t3 DPMC-05331 LAM 008841 T itra tio n of c h ry s o t i le ML. 0.5150N HCI F -tg u re 14 ------------ '"*? > r=: CXI v o OJ CO OJ C30 O<T>00f^-COtO^trO0J Hd o DPMC-05332 LAM 008842 '.j i 1 FIGURE 15 ELECTRON MICROGRAPH OF COALINGA C1IRYSOTILE 8000X EMS 58QA DPMC-05333 LAM 008843 ELECTRON MICROGRAPH OF JEFFREY CHKYSOTILE 8000X EMS 485A DPMC-05334 . LAM 008844 FIGURE 17 ELECTRON MICROGRAPH OF CROCIDOLITE 8000X DPMC-05335 LAM 008845 I------------------1 100 microns FIGURE 18 SOFT CHKYSOTILE 200X DPMC-05336 E 984 LAM 008846 I------------------1 100 microns FIGURE 19 HARSH CHRISOTILE 200X DPMC-05337 LAM 008847 Endothermic Exothermic <-----------------------At ---------- -> LAM 008848 100 2 0 0 300 4 0 0 500 6 0 0 7 0 0 8 0 0 9 0 0 1000 TEMPERATURE (C> fig u r e 21 Mb ENDOTHERMIC ------------ 0------------ - EXOTHERMIC DPMC-05339 LAM 008849 > - I 1 micron FIGURE 22 ELECTRON MICROGRAPH OF SPEX MILLED CHRISOTILE 15,OOOX EK3 k90A DPMC-05340 LAM 008850 1 micron FIGURE 23 ELECTRON MICROGRAPH OF BRAKE LINING IX/ST 15000X EMS U70C DPMC-05341 LAM 008851 *4 1 100 microns ' FIGURE 24 ELECTRON MICROGRAPH OF COMMERCIAL TALC 3000X EK3 440B DPMC-05342 LAM 008852 F # -*. .. .. >h FIGURE 25 ELECTRON MICROGRAPH OF WARREN COUNTY, NEW YORK SERPENTINE 8000X BUS 705A DPMC-05343 * LAM 008853