Document EveGb713mLy008XvEbb640o8n

I V I ) KUIN A. MLIILM AN, EDITOR National Institutes of Health P O. Box 1829 Rockville, Maryland 20850 (301) 496-3155 ASSOCIATE EDITORS Morris E. Cranmer N.iiion.il Center for Toxicological Research food and Drug Administration lellerson, Arkansas 72079 Gary Flanim National Cancer Institute Bethesda, Maryland 20014 Leon Golberg Chemical Industry Institute of Toxicology P.O. Box 12137 Research Triangle Park North Carolina 27709 Robert A. Goyer Health Sciences Center University of Western Ontario London, Ontario N6A5CI Robert G. McConnell Division of Research and Development Searle Laboratories P.O. Box 5110 Chicago, Illinois 60680 David P. Rail National Institute of Environmental Health Sciences P.O. Box 12233 Research Triangle Park North Carolina 27709 Gerald N. Wogan Department of Nutrition and food Science Massachusetts Institute of Technology Cambridge, Massachusetts 02139 oP U B L IS H E R HEMISPHERE PUBLISHING CORPORATION 1025 Vermont Avenue, N.W., Washington, D.C. 20005 E D IT O R IA L S T A F F: Evelyn Walters Pettit, Editorial Director; Mary A. Phillips, Managing Editor, Christine R Flint and Martha M. Mahuran, Editorial Assistants; Ronald L. O'Dwyer, Production Supervisor. Earlene Tarwater, Assistant to the Editor. SUBSCRIPTION R A TES (Six issues per volume, bimonthly): Libraries and Institutions- in U S. and Canada, $58.50; elsewhere, $61.00. Individuals-in U S . and Canada, $29.50;elsewhcre, $32.00. Individual subscriptions must be accompanied by a personal check or money order. SUBSCRIPTION S E R V IC E : Shipments will be made only upon receipt of remittance. All orders, remittances, and communications related to handling subscriptions, including notifications of change of iddrcss, should be sent to the Publisher. Change of address notices should provide the old as well as the tew address, including zip code number. Changes will become effective within approximately 20 days. Missing copies will be replaced only if valid claims arc received within 90 days from date of mailing. Replacements will not be allowed if the subscriber fails to notify the Publisher of a change of address or f a copy is misplaced at destination. Request for refund will be honored only prior to delivery of issue lumber 1 of each volume. SUBMISSION OF M ANUSCRIPTS: Original and three copies of each manuscript in English, which ias not been published elsewhere, should be submitted to the Editor or any of the Associate Editors Printed in the United States of America pyrighi ) 1976 by Hemisphere Publishing Corporation. All rights reserved. No part ol ihis publication m.iy tu firodoted in any form oi by any means without the prior written permission <>i it.- p n l . l i - i > CONSUMER TALCUMS AND POWDERS: M IN ERAL AND CHEMICAL CHARACTERIZATION A . N. Rohl, A . M. Langer, I. J. Selikoff, A . T ordini, R . Klimentidis Environmental Sciences Laboratory, Mount Sinai School ol Medicine of the City University of New York New Y o rk , New York D. R . Bowes, D. L . Skinner Department of Geology, The University of Glasgow, Glasgow, Scotland Representative consum er talcums and powders, including 20 body powders, bub) pow ders, facial talcums, and also one pharmaceutical talc, were analyzed to determine their mineralogical and chem ical com position. Where kn ow n , oil were form ulated prior To 1973. O l the 20 pro du cts 10 con tain ed detectable amounts o f trernolite and anthophyllite, principally asbestiforrn, while some also contained fragmented forms ol these minerals. The amounts ranged from tenths o f a percen t to o ver 14% by weight; tw o contained detectable amounts o f chrysolite asbestos fiber. Light contained quartz, seven ranging fro m 2 to 5%, with one as high as 35%. The analyses show ed that the con su m er p ro d u cts exam ined were rarely the p u re mineral talc, but rather were m ixtures o f various m inerals; some samples consisted o f three to five minerals, only o ne o f which was talc. O th er co m m on mineral phases in clu d ed chlorite, plate Serpentine, py ro ph yllite, mica, and carbonate minerals. Kaolin additive was idenlilied in tw o pro du cts. The single pharm aceutical talc exam ined conloinvtl o n ly a trace amount o f quartz. The chem ical com position o f these products, including both m ajor o xid e and trace elem ent con ten t, co rrela ted with their mineral com ponents, l o u r samples contained substantial concentrations o f nickel, cobalt, and chrom ium , suggesting lattice substitu lio n o r the presence o f rrace mineral phases. Geological provenance o l the talcs may be ascertained on the basis o l chem istry. Possible adverse health eHects Iront interm ittent use o f these pro du cts, especially those that contain axhc'slitorm and fragm ented anthophyllite an d trernolite, chrysotile, quartz, and trace metals, are presen th u n kn ow n and warrant evaluation. INTRODUCTION Consumer l.iltuins and powders arc considered by lire general public In be laic, an impression lirai is conveyed and strengthened by lire produci this study vv.is su p p o n i it in p.m hy grani I S 009 28 Iront liti* National Involute I nviroitm riil.il IKm Ii Ii b u cnccv. One ol uv (AM I ) ic ic iv c tl vuppoii untici l .n e ri N iic n iis i A w a iJ l S MSI 2. National In silim i' ol l u v n iim m o l.il llia t lli S iic iu r s U.-HIUNI-. I . . I i r p i i t i l ' . slt< m ill In- M ilt !> N 0..1.1 I m i . . .in u m o .1 S . . I ............... ' 156 A N ROHL ET AL names and ingredients listed on the container labels. However, knowledge of the geological occurrence and mineralogical character of source materials suggests the nature of talc to be highly variable and complex. Twenty-one consumer talcums and powders (Table 1) were mincralogically and chemically analysed to determine whether these products arc actually talc mineral. Talc is a defined mineral entity, based on specific chemical, crystalline, and physical properties (Ford, 1957). The empirical chemical formula, Mg( Si4O |0(OH)2, is seldom observed in nature as a result of cation substitution. For example, magnesium is frequently replaced by iron, nickel, chromium, or manganese in the crystal structure. Talc is a sheet silicate, with a structural unit consisting of three layers; a sheet of octahedrally coordinated magnesium hydroxide groups is sandwiched between two layers of tctrahedrally linked silica layers. The Van dcr Waals bond between the talc sheets are of low energy accounting for the case with which talc as well as other sheet silicate minerals (micas, clays) cleave T A B L E ). Designation of Contents in Brand Name or la b e l Sample nu. D a te c i formulation Products designated as tales 3 Not available 8 Not available 12 Approxim ately 1972 7 September 1972 17 Not available 6 April 16, 1973 19 May 1973 21 Not available Products designated as powders (talc on label) 4 December 1970 18 Not available 20 December 1970 S February 1973 16 Not available Products designated as powders (or dust) 1 Between |anuary 1968 and July 1970 14 October 3, 1970 IS Between October 1970 and March 1973 9 February or March 19 73 11 July 1969 13 July 16, 1970 ? Not available CONSUMER TALCUMS AND POWDERS 2S7 or break inio plaly fragments. This facile cleavage, with resultant high surface area, and its softness, small particle size, and whiteness confer upon talc its usefulness as a cosmetic material. GEOLOGICAL OCCURRENCE OF TALC Talc rocks (including those commercially worked) arc formed by several complex geological processes reacting upon many possible, chem ically diverse preexisting rock types. Hydrothermal alteration of magnesiaand silica-rich ultramafic rocks, under a range of low-to-modcratc tempera tures and pressures, may produce talc. Thermal metamorphism of silicarich dolomite (C aM g (C O j)j| will produce talc as well. These processes, however, also commonly result in the formation of a number of other coexisting mineral phases, predominantly hydrous magnesium silicates. Some of these, for example, anlhophyllite, tremolitc, and serpentine minerals (including chrysotile), occur as microscopic intergrowths with talc, as macroscopic nodules, or even as discrete zones within or adjacent to talc (Table 2). Talc rock is therefore generally not monominerallic but is often a mixture of minerals that may vary widely with respect to kind and quantity. Phlogopitc, a magnesium mica, and chlorite, a group of minerals related to the micas, are also commonly associated with talc. Some of these associated mineral phases are asbestiform amphiboles and chrysotile (sec discussion of the terms asbestos and asbestiform in Appendix A). Conversely, talc has been described as a common accessory mineral in commercial asbestos deposits (Hurlbut and Williams, 1935). Talc deposits may be zoned, with different mineral assemblages physically changing in occurrence and proportions over extremely variable T A B L E 2. Minerals that Com m only O ccur in Talc De|H>sils Mineral group Phase 1 orm ula C jr bona les A m phiboles Serpentine O th i-is Calcite D olom ite Magnesite Trcm o litc0 A n th o p h y lliic 7 Antigoiitc Chrysotile (uncommon) 1i/ardilc (uncommon) Quart/ Mica, c.g., phlogopitcChlorite-, c g.. penninite Pyiophyllite CaCO, CaM gICO ,), MgCOj U ,M g ,S i.O ,, (O II), (lc M g ),S i.O ,, |O II), M g .Si,0 ,(01 1). M g ,Si,0 ,(01 1), M g ',Si,O ,(O il), S.O , K.lM g .l r |. |S i. A t,l> ,.|(O II). (M g.A l.l -),,| ( S i .A I ) .0 ,,, 1 (0 11 ),, A l.llS i. ,.. I(O H ). 258 A N K O III. T A L distances, tanging from centimeters to tens of meters. Mineral phases in such deposits may include talc plates and fibers, trcmolitc and anthophyllile fibers, inlcrgrowths of amphibolc and talc, serpentine minerals (which may include chrysolilc), and free silica (quart/) (Ross et al., 1968). The liber inlcrgrowth is often such that even extensive bcncficialion may not yield a pure product. Thus, where fine-grained inlcrgrowths of talc and trcmolitc occur,' the processed product will likely contain residual tremolite. Further details concerning the crystal chemistry, structure, synthesis, and geological occurrence of talc arc found in Appendix B, INDUSTRIAL AND COSMETIC GRADE TALCS It is generally recognized that various commercial grades of talc arc marketed in the United Stales (Appendix C). Hildick-Smilh (1976) has stated that a talc suitable for pharmaceutical purposes, used in cosmetic and toiletry products, contains al least 90% talc mineral and no detectable asbestos. Such stated compositional restrictions are not placed on indus trial grade talcs. One study demonstrated that a number of industrial talcs contained substantial quantities of trcmoMlc, up to 87% by weight of the sample (Schulz and Williams, 1942). In 1968, a study (Cralley el al., 1968) of 22 cosmetic talcum products demonstrated fiber contents ranging flom 8 to 30% by count with an average of 19%. The fibrous material was predominately talc but probably contained minor amounts of trcmolitc, anlhophyllilc and chrysolilc as these arc often present in fibrous talc mineral deposits. With the exception of 4 of the 22 cosmetic talcum products analyzed, the levels of free silica, cobalt, nickel, chromium, and manganese were generally of a low magnitude and within a narrow range. . . . The levels of silica, chromium, and nickel in these four products arc sufficiently high, however, to be of concern in their potential to cause disease. Thus, as late as 1968 some consumer talcum products marketed in the United Slates contained asbestiform minerals, free silica, and trace metals. HUMAN D ISEASE ASSOCIATED WITH TALC EXPOSURE For nearly half a century a number of repot Is have shown that occupational exposure to talc dust is associated with a line diflusc interstitial lung scarring known as lalcosis. Fibrous talcs appeared to he more pathogenic than plaly talcs, producing in addition to lalcosis, CONSUMER TALCUMS ANOPOWDERS 2S9 Messile, 1960; Klcinfcld el al., 1967). Studies concerning the biological consequences associated with talc dust exposure, including cancer, are referred to in Appendix D. O BJECTIVES OF THE PRESENT STUDY Twenty-one samples of consumer talcums and powders, including baby powders, body powders, facial powders, and a pharmaceutical talcum, were obtained al retail stores in the New York City area. These samples were acquired and studied during the period 1971-1975 (Table I). The major purpose of the study was to determine the mincralogical and chemical composition, with particular emphasis on the quantitative determination of trcmolite, anthophyllile, serpentine minerals, and quartz. Another objective was to establish a base line for consumer talcums and powders, based on a sampling of products available during the period 1971-1975. This base line provides an index for evaluating possible changes in subsequent formu lations METHODOLOGY AND RESU LTS OF MINERAL AND CHEMICAL CHARACTERIZATION The analytical techniques employed for mineral idem ifra ti~.i and quantification included optical microscopy, transmission electron micro scopy with selected area electron diffraction, X-ray diffraction, and scanning electron microscopy with X-ray analysis capabilities. Chemical determinations (bulk chemistry and trace metals) were made with a number of standard instruments and geochemical techniques, including spectrophotometry |S iO j, TiO a, A ljO Jp total Fe ( F c 2O j by difference from FcO), P2Os | ; atomic absorption (MnO, MgO, CaO); flame photo metry (Na; 0 , K jO ); wet chemical assay (FeO ); weight loss, volatiles (H jO , C 0 2, organics). X-ray fluorescence was used for all trace metals (Bowes and Langcr, 1974). Optical Microscopy Optical microscopy is a conventional technique for the identification of minerals and for the study of mineral relationships. A microscope equipped with bright field illumination and polarized light optics was used to analyze the cosmetic powders. Approximately 0.5 mg of powdei was placed on a precleaned glass slide and immersed in index oils of known refractive indices. These were checked on a refractometer. The information obtained on particles with this method included most ol the measurable optical properties, including indices of refraction, extinction angles ol libers, gencial morphology, and size characteristics ol mincial j ha-is 1 ig. I I It .U K E I Optical photom icrographs ol Cosmetic talcums and powders grossly contam inated with ashcsiilon n minerals. Photos obtained in plane polarised light ( A - F ) and between crossed polais (<i). Scale in (A ) lor all photos, l ibers shown have optical properties consistent with irem olile (I'r) and aotliopliyllite (A n ). Quart/ grains (Q u) and calcile fragments (Ca) arc also shown. Most ircm nlitc libers tend to possess smaller length lo w id lh ratios than antliopliyllite. Frem olilc is " ' "hi imvd between i rossed pol.os (fi) also dem ons H alts the p icsriite CONSUMER TALCUMS AND POWDERS 261 by morphological and oplical characteristics. One product consisted entirely of cornstarch. In most samples, however, the powders were loo fine grained, with particle dimensions significantly less than 1.0 pm, for the technique to be useful. The limiting factor for determination of optical constants, and hence for identification of particles, is the resolving power of the microscope. The presence of talc fibers, which may have indices of refraction similar to amphibole (tremolite, anthophyllilc) fibers, also confounds analysis. Therefore, although this technique is an excellent diagnostic instrument in some instances these restrictions limited its usefulness. Other investigators have experienced similar difficulties, fo r example, in a study of the asbestos content of talc (Stanley and Norwood, 1973). there was difficulty in applying oplical microscopy to the problem. The authors concluded that oplical microscopy while it works well on pure samples of fairly massive fiber length Irom 3 to 5 microns, our observations by transmission electron microscopy have shown that naturally occurring usbestiform minerals often lie below the working resolution capabilities o f the light microscope and furthermore while massive fiber bundles can often be observed by either light or electron microscopy the observation of individual fibers smaller than 0.5 by 0.2 micrometers often will require the high resolution capability of the transmission electron microscope (emphasis added). It is further slated that 'Tight microscopy was helpful only in screening samples with large particles and high concentrations ol objectionable fibers." By comparing the results of oplical microscopy with those of quantita tive X-ray diffraction and electron microscopy, we observed that large numbers of fibers go undetected. In addition to the restraints of resolution imposed by light microscopy, another major drawback rela'cs to the strong tendency of asbestiforni minerals to cleave or break along planes ol weakness when they are crushed, producing large numbers of small fibers. For example, light microscopic examination of a talc sample (no. 8), which contains over 7% tremolite (sec Table 4), demonstrates the presence of mineral fragments that are primarily not asbestiform (Fig. 1C). How ever, electron microscopic examination of the same sample demonstrates that many of the submicroscopic tremolite particles are fibrous (see Fig 31'). The problem involving the determination of the relative proportions of each ol these morphologic.il phases iii the same sample (mass vs. 262 A N ROML ET AL. X-ray Diffraction The application of X-ray diffraction analysis in step-scan mode for quantitative determination of asbestos in talc has been described in detail (Rolil and Langcr, 1974), including (he selection of talc and asbestos reference materials, the preparation of standard dilutions of asbestos minerals in talc to ensure sensitivity and reproducibility, the selection of characteristic X :ray reflections to be scanned, and instrumental technique. Selection o f talc mineral standard. Screening of various types of talcs for use as reference material was made by X-ray diffraction and trans mission electron microscopic analysis. A continuous scan was first made to identify the major mineral phases present. Talcs (hat showed the presence of any serpentine mineral, tremolile, or anthophyllitc were eliminated as reference materials. The possibility of false negatives for these minerals was checked by step scanning the diagnostic reflections (Table 3). Further verification of the absence of asbestiform minerals was made by trans mission electron microscopy. These techniques permitted the selection of a matrix talc that was completely free of asbestiform minerals. Chlorite minerals arc hydrous iron-magnesium silicates, frequently associated with talcs. Their presence may interfere with the detection of serpentine minerals, both platy (antigoritc) and fibrous (chrysolilc). This is particularly true if they arc present in equal or larger amounts than these latter minerals. Two intense basal reflections at 14.2 A (001) and 7.1 A (002) are characteristic of chlorite minerals. The latter reflection occurs dose to the (002) reflection of serpentine minerals (7.3 A). The 3.66 A (004) reflection of serpentine was selected as diagnostic (Table 3), since the lower intensity (004) reflection of chlorite (3.S3 A) in this region was found not to cause interference. Selection ol the asbestos standard. Reference samples of pure asbestos minerals were obtained from various mineral collections and from the International Association for Research on Cancer (IA R C ). These were screened for purity and particularly for the presence of interfering contaminants according to the procedures previously described for talc. Two different specimens of chrysolilc were used as reference materials: a triple air-jet milled sample from the Jeffrey Mine, Quebec (provided by the (ohns-Manvillc Corp.) and a specimen from Coalinga, California (provided by the Calidria Division, Union Carbide Corp ). Sample preparation. Among the variables that strongly influence the precision and accuracy of quantitative X-ray diffractomctry are particle sice, preferred orientation, and surface flatness. Variation due to particle st/e can be minimized by crushing and screening the asbestos and talc standards to ensure a uniform sice distribution, with an effective crystallite dimension on the order of 5 pm or less. The effect of preferred orientation is more difficult to control. The tendency for preferred orientation is largely the function ol mineral cleavage properties llotli the CONSUMER TALCUMS AND POWDERS 263 T A B L E 3. Calibration Curve Data (or the Determ ination ot Asbesliform M inerals jn d Q u a il/ in T alc by X-ray Analysis47 Mineral phase Parameter Am hophyllitc Chrysolite Qoart/ Miller index of diagnostic reflection Corresponding //-spacing (A) Relalivc intensity Step-scan interval (2 theta) (2101 8.26 55 10.0-11.0 (004) 3.66 80 23.5-25.0 (211)'' 1.54 15 59.5-60.5 Percent mineral in talc and corresponding area of reflection Trem o litc 1110) 8 38 100 10.0-1 1.0 Am hophyllitc % in.' r so 10 0 15.0 20.0 25 0 30.0 35.0 0.92 2.35 3.99 4.65 5.12 7.22 9.92 Chrysolite % in.1 0.25 OS 1.0 2.0 5.0 7.0 8.5 8.9 0.01 0.03 0.08 0.22 0.94 1.52 2.21 2.34 Quart/ % in. 5.0 10.0 20 0 25.0 30 0 40.0 2.75 3.75 7.11 9.30 13.20 16 90 ficm oliic % in.* 0.1 0.5 2.2 27 43 5.0 7.0 10 0 0.13 0.30 0.4 3 0 67 1.25 1.65 2 08 2.80 D etection limit Wt % A n ttiophyfliic = Area (2 1 0 ) - 0 .2 7 (% A n) Wt % Ch ry so tilc = Area (004) - 0 27 (%Ch) W i% Q u .til/ = Area ( 2 1 1) - 0.43 (% Q) W i% Irem olile = Area (110) = 0 28 (% Tr) p.52 R ' - 0.95 0 .5 0 R = 0 .98 0.60 R* = 0 96 0.04 R 1 - 0.98 2.0% 0.7% 14% 0.1% ^Instrumental settings: target/filter, Cu/N i, 45 LV /2 0 m A ; scintillation counter, 1450 V1H . m onochromator, graphite; pulse-tiinc analyzer, 20 V. 5 V (w idth, level); conti iu o u s scan, I 2 P/rnin; step-scan, 0.02 2 0 at 2,000 counts fixed. ^Khom bohcdral index. ` Repealed m easurements ol areas under cuives with a polar planim ctcr in dicated aveiage deviation of t 0.02-0.05 in .1. fibrous in lire ease of Iremolile, anlliophyllile, and clirysolile (see Kolrl and Lunger, 1974, Fig. 2). In attempting to reduce or eliminate the eflects of preferred orientation in X-ray analysis, a number of sample preparation and instrumental techniques have been developed (Bragg, 1967; Biirulley and Kurlossy, 1961; C u llily, 1956, King and Alexander, 1954) hr the present study these techniques were tested, but none was found to provide adequate precision (reproducibility). Accordingly, a sample preparation technique was developed that was successfully used, in conjunction with X-ray rlilliaclion itr the step-scan mode, to delect diagnostic rcl led urns ol I I .1 t r \ i `is.- rsl r r r i l i i ' l t l l f I |('|t` 0 264 A N. R O H l ET AL. The reproducibility of reflection intensities was also greater than other preparation techniques tested. Binary dilution standards of chrysotile, anlhophyllitc, and tremolile in talc were prepared gravimetrically. Asbestos fiber concentrations were prepared initially at 5.0, 4.0, 2.0, 1.0, 0.5, 0.2, and 0.1%. Fifty milligrams of the talc-asbestos mixtures were homogenized in 10 ml filtered water utilizing ultrasonic energy. This slurry was poured into a 30 cc syringe and filtered through a 0.22 pm pore size membrane filler. To prevent stratification due to differential particle size and density effects, the syringe is held in a horizontal position, rotated, and shaken during filtration. The residue forms a flat cake of about 0.5 mm uniform thickness on the membrane filter. When dried, the sample is affixed to a glass slide for X-ray diffraction analysis (Rohl and Langer, 1974). Selection o f X-ray reflections. Because of crystal structure similarities in the minerals being studied (i.e., tremolile and anlhophyllitc), consider able overlapping and interference of X-ray reflections occur. The low symmetry and consequent complex X-ray diffractograms of such minerals as talc, chlorite, and mica, as well as possible interferences from admixed phases such as kaolinile, make it necessary to select a reflection or set of reflections for each mineral component (hat could be used as an index of the amount of that mineral in a mixture. Such diagnostic reflections were selected by referring to standard X-ray powder diffraction data. These diagnostic reflections were step-scanned at 0.02 2 theta in a fixed count mode (2 X I0 J counts). Precise positions and profiles of the diagnostic reflections were determined. In the fixed count mode, each of the angular intervals selected arc scanned with equal accuracy. Thus weak reflections can be determined with equal precision as high intensity reflections. The statistical accuracy depends only on (he total number of counts recorded, and the counting rale selected gives a percentage probable error of about 2%. Profiles of the diagnostic reflections, plotted as a function of number of counts vs 2 theta, are measured with a compensating polar planimclcr. The intensity of a reflection is proportional to, but not necessarily a linear function of, its concentration. Other factors that may influence reflection intensities include instrumental conditions, particle size, degree of pre ferred orientation, sample thickness and flatness, and absorption character istics (Klug and Alexander, 1954; Rohl and Langer, 1974). Figure 2 shows calibration curves obtained for chrysotile, anlhophyl litc, tremolile, and quartz using the step-scan technique. Measured areas of diagnostic reflections arc plotted against percent dilution in talc As indicated in Table 3 tremolile may be determined at levels as low as 0.10% by weight, chrysotile from 0.25 to 0.50%, and anlhophyllitc, as low as 2% in talc. It is important to note that the limits of detection given in Table 3 are higher and based on a best fit regression analysis For example, regression analysis indicates that the detection limit lor chryso lite is (I 7"' .. I,..... .. I ....... n sr . CONSUMER TALCUM S AND POWDERS V> . u s w & 3 0 b i .:`- i sarf Trem olue / / // y / / / Chrysolite /r // Quarw / 0/ A / s '0 Am hophylhie 10 %/ 123456789 10 Percent asbestiform m inerals in talc 5 10 15 20 25 30 35 40 45 Percent asbestiform am phibole/quari/ in talc F I G U R E 2. Calibration of asbestiform am ph iboks and quaria in talc. IPfelli| pIpffltM:e fell; m-myi i n * t i depending on panicle size, degree of cry stall inil y, clc. The changes in slope al ihc lower end of the curves are not reflected so that axial intercepts are exaggerated on the high end of the abscissa. By using X-ray diffraction in the step-scan mode, Stanley and Nor wood (1973) were able lo detect a minimum of 0.25% trcmolite in tal and a minimum of 0.5% chrysolile and the other asbestiform minerals However, such low levels of chrysolite were not delected when chlorite was present Step scanning. The contents of the containers were thoroughly mixed with a sample splitter lo avoid stratification effects. Aliquots ol each, weighing 50 mg, were prepared using the identical methodology described for preparation of the dilution standards. The filler-mounted samples weie then step scanned over Ihc gnniomclric intervals diagnostic lor the standard asbestiform minerals and quartz. Instrument operating conditions were identical with those used for analyzing the dilution standards. Profiles plotted for the diagnostic intervals and reflection areas altei peak stripping were measured by polar planimetry The weight percents ol anthophyllite, trcmolite, and quartz contents were estimated by iclening lo the appropriate regression curve. The quantities ol asbestiform inalciials in the 21 tab lints and powcleis found by litis iccbniigic ate shown in 5 Q5 a5 o5 o. 5 Q- High K ; 0 and N a ,0 suggests that this phase is a mica (m uscovite/biotite). 5 q. a . a. 5 a . i q5 o- i J ^ 'll 5 a- I 5 a- I < 5h c~o----6ojCQ. " o* -aoc - -Ua-UOO.^>-<rua. <* CONSUMER TALCUMS AND POWDERS 267 minerals, ranging in amounts from a few tenths of a percent to over 14%. Trcmolite was the most commonly found (9 of the 21), and anthophyllite occurred with trcmolite in 6 of the 21 A serpentine mineral phase was indicated in two samples, in amounts at or near the lower limits of detection. Verification of the serpentine phase as chrysotilc in the two samples in amounts corresponding to the observed concentrations was made by electron microscopy. Continuous sconning. In order to study the presence of all mineralg ica! (and possibly other crystalline) components the samples were scanned from 5 to 70 2 theta at a scanning rate of 1 2 theta per minute. This technique, as expected, proved satisfactory for the identification of major components, but it was generally found to be incapable of detecting trcmolite, anthophyllite, or serpentine minerals except in cases of gross contamination. The high noise level (low peak-to-background ratio) often prevents the detection of quantities on the order of 4-6% or less and also excludes this technique for quantitative analysis. As a result of high noise level three false positives, as indicated by continuous scanning, were subsequently shown to be negative for amphibolc by step scanning and electron microscopy. In addition, continuous scanning may not generally detect serpentine minerals in the presence of chlorite or kaolinite. Electron Microscopy and Electron Diffraction The transmission electron microscope has been shown to possess the sensitivity required for fiber identification and for determination of particle size distribution of submicroscopic particles (Langcr and Pooley, 1973; Langer ct al., 1973). Accordingly, aliquots of talcum samples were prepared for election microscopic analysis by a technique that disperses particles in a drop oi nitrocellulose solution on a glass slide. A second glass slide is placed on the first and the two arc drawn lightly apart, leaving a film. This technique is intended to minimize (he alteration of particle size distribu tion. The film is mounted on electron microscope giids and scanned at magnifications of X20.000. Morphologically, amphibolc minerals are generally quite dissimilai Iron) other silicate minerals. Both anthophyllite and trcmolite arc rectilinear, often with amphibolc-lypc step cleavage or, infrequently, with prismatic terminations. Trcmolite fibers tend to be electron dense and shorter than anthophyllite (Fig. 3), while the latter has a tendency to be election translucent and to show diffraction contrast figures (l.anger and Pooley. 1973). Sheet silicate minerals (talc, chlorite, micas) lend to be equidimensional in shape, often with pscudohexagonal outlines (Tig 4). ( nr led talc plates or talc fibers on edge may superficially resemble asbcsliloim minerals, but selected area election diffraction patterns easily distinguish 1 F IG U R E 3 Iu n sm issio n election micrographs showing range in m orphological characteristics ol jv h tslilo r m Ircm olitc and anihophyllitc in talc. The entire range ol m orphological variations observed lor these minerals is observed in the asbestos standards: rectilinear libers with parallel ends and edges ( A ) ; slc p - c k jv jg c ends (B ); unit fibrils protruding from lib er bod y, ( C ) . curvilinear liber with amphibolc cleavage end (D ); high length to w idih ratio libers ( t ) ; libers piolruding from interiors of talc plates ( F ) . All ol these m orphological variations and forms ( A - L ) have been described in a n ih o p h y llitc jn d trc-molilc ascbcsios sjm p lcs. The am phibolc structure was confiim cd in all cases hy selected area electron d iffia clio n characterization. Seale as m arked. Micrographs ohi.iiin'd on a 11 O l. |LM 120 U with an accelerating voltage at 120 k V . CONSUMER TALCUMS AND POWDERS 269 F I G U R E 4 Transmission electron micrographs of cosmetic talc samples com posed prim arily of plates with an occasional large talc fiber ( B ,C ) ; talc with many talc fibers (D ). Stair* is the same in (A ) and (B ) and in (C ) and (D ). All selected area electron diffraction patterns obtained on libers yielded those consistent for talc plates (see Fig. S, A and B) In all tale-containing samples exam ined by election m icroscopy talc grains tended to range from 2$ to 0.2 pm in greatest dimension. Micrograph obtained on a J L O L |CM 120 U with an acceleration voltage at 120 k V . w.is used to verify lire presence of amphibole in the 10 samples shown to he positive by X ray diffraction. Electron microscopy, while not quantitative, also showed that amphibole fibers were present in relative amounts that corresponded to their percentages .is shown by X-ray diffraction. The presence of traces of chrysolite, rather Ilian pl.ns' serpentine, in samples 12 and IS was verified by election microscopy (tig. f>). Hy F IG U R E 5. Transmission electron micrographs and accom panying selected area electron detraction (SA C D) patterns lor cosmetic talc samples. I he talc plate (T P ) in (A ) shows typical polygonal cleavage and d illra ciio n contrast contours lor the mineral species. The accom panying S A I D pattern displass the characteristic reciprocal ab* plane pseudohexagonal sym m etry loi talc. Talc a* and b* directions indicated on (A ). Measurement oi pattern indicates a $.3 a repeal along a* and a 9.1 A repeat lor b (measured at 11 101). Talc liber ( I F ) in (B ) displays irregular ends and mmrec (linear edges. The SAL'IJ pattern is also pseudohexagonal. but some reflection intensities |c .g ., the (0t0). (0 .1 2 .0 )| arc more pronounced. This may be due to both orientation and siiu ctu ral e lle cls. lh c curled laic (dale (C T ) in (C) displays an incipient Debye-Scherrer ring pattern (the e lic it s ol both lolding over ul talc and small associated grains), lh c ain phi bole lihct ( 13) was d illra ctcd only u ii one ol the protruding unit lib rils. lh c c* axis is show n, with repeal measured al S I A . A n a s " hcie ditlr.iclion patterns were obtained are indicated hy location ciicle s. p .n liih -s wen- photo gr.iplicd .it Hie SA I I) iiiagnilic.tiioii / . `ii.SltO Scale is it s.mn- m ,\ ( ih m |i n. n l . i l F I G U R E 6. Transm ission electron micrographs of asbcstilorm m inerals in m s me I a la ic, other than am phibolc. (A and 0 ) Tw o chrysolite libers with m orphological characteristics induced b\ election beam damage No d illra ciio n p a licrn was obtained on either liber. Arow m arkers (A and ID indicate areas where these beam-damaged features arc most prominent Doth lib er bundles appear to rest on la k plate substrates. Free chrysolite libers and fibrils (C ) weie loui.d in a sample found negative for asbestos by all other technii|tics. Scale as m arked. M icrogiaphs obtained on a |l O l 11 M 120 U with art accelerating voltage at 120 k V . A N ROIIL ET AL comparison with known dilution levels of chrysolite in talc observed by electron microscopy, the levels of contamination of chrysotilc in the two samples correspond to about 0.25-0.5% chrysotilc, which was suggested by the X-ray diffraction results. The chrysolite fibers were all shorter than 2 pm and the diameters less than 0.2 pm, explaining why they were not visible by optical microscopy. Chemistry of Consumer Talcums and Powders The bulk chemistry (Table 5) and mineral contents (Table 4) of the talcums and powders complement each other in that one data set implies limits for the other. For example, analysis of sample 1 shows the presence of FeO and CaO (Table 5). Recalculation of these oxides into values for the empirical formulas for tremolitc and anthophyllile indicates that sufficient quantities are present to account for the presence of these minerals (Table 4). An appreciable decrease in S i0 2, which should normally occur, was not observed, because of the occurrence of over 5% quartz content. The A l20 3l Na20 , and K 20 arc reflected by the presence of chlorite (probably the penninile phase). Variations in oxides were also observed in the other samples that contain amphibole minerals (6-8, I I , 12, 15-17, and 21). In these samples, a number of other factors account for the wide variations in oxide percentages: in 6, the presence of chlorite T A B L E 5. M j|ur O xide Content of 21 Consumer Talcum and Powders0 Sample no. Major oxide 1 2b 3` 4 5 6 7 8 9 10 1 1 SiO , iiO : A I, O , le .O , IcO MnO MgO C aU N.i () K .U *.(), Vula nils' ln l.il 61.99 0.10 0.82 0.00 1.51 0.00 29 60 0 10 007 0.02 0 01 5 36 9*J 88 000 0.00 1.30 0.00 0.00 0.00 0.00 0.00 000 0.00 0.00 98.70 62.68 0.14 0.37 0.08 0.32 0 00 29.83 0.04 0.03 0.02 0.03 4 66 58.72 0.04 0.24 0.04 3 03 0.00 28.25 0.10 0 03 0 02 0.03 5 38 61.94 0.03 0.45 0.03 0.51 0.01 30.19 0.04 005 0.02 0.01 5.51 100.00 98.20 95.88 9 8.7 9 52.95 0.17 1.16 002 0.86 0.00 29.02 3.46 0.00 0.05 0 13 IU.32 59.67 1.40 0.87 0.18 0.84 0.00 29.23 0.90 0.05 0.05 0 14 5.64 98.14 98.97 71.93 0.19 15.73 0 10 0.34 0.00 2.95 2.37 0.48 1.37 0 05 5.25 58.68 0.08 0.58 0.12 3.12 0.00 29.23 0.17 0.00 0.00 o.uo 5.34 51.65 0.71 1.45 0.09 0.34 0 00 27.90 4.75 0 03 0.00 0.05 9.72 49.48 0.22 2.32 1.23 0 07 0.00 25.67 1.61 0.42 0.19 0.01 1U.41 100.76 97.32 96.69 91.63 "iv.-iehi percent. recalculated as o tides, following standard p elro ilicm ic.il procedures. H it hulk analysis ol Iht pu ssd iis r c llc ils llu- m m liin id m in iral content a fici additives were e xtra iled using w jie i. Jilu ii' IK I, . m l one, ben/ene, and cih cr. Allah sis nf 2. s la lih i organics jn d volatiles; 1.1% A I , ( ) , (aluminum i lilorhydi ale ') ' Si.inti,ltd i.ih used as in ain x lm lilu t stand.m l dilutions. CONSUMER TALCUMS AND POWDERS 271 sharply reduces Ihe S i0 2 conlenl and elevaies ihe A l20 3 content. The presence of ihe carbonate mineral calcile increases ihe expected CaO and volatile contents (the latter includes C 0 2); in 7, the high TiO j is reflected by the presence of the mineral rutile (T iO ,); in 1), the high CaO, moderately high A l20 3, and low S i0 2 contents reflect the presence of calcitc, chlorite, and tremolile. Several of the above samples (numbers 8, 12, and 15) require special evaluation. Sample 8 is extremely high in S i0 2, A l20 3, Na20 , and K 20 and extremely low in MgO content. The chemistry indicates that this material was not derived from a talc rock, but rather from one rich in alumina and silica. The mineralogy reflects this, as do the trace metals (see trace metal section). The amphibole minerals in sample 8 are associated with both pyrophyllitc | AI4SiB0 22(OH)a ) and quartz, both present in substantial quantities. The mica phase is not plilogopitc, but muscovite, accounting for the presence of substantial quantities of K 20 and Na20 . A plagioclase feldspar was also detected in the mineral phase.Sample 12 is very high in A l20 3 but extremely low in S i0 2, apparently Ihe result of high chlorite content as well as substantial amo'in'.s of pyrophyllitc. Sample 15 is low in S i0 2 and extremely high ir. volatile content, reflecting the presence of both carbonate phases am1 organic additives. The trace element analyses (Table 6) show distributions that are in accordance with the known behavior of trace elements in minerals. With T A B L E 5 [co n iin u fd ] Major O xide Co n ie m of 21 Consumer Talcum and Powders1' Sample no. 12 l 3 14 IS 16 17 18 19 20 21 A*' u* 47.32 0.18 9.34 0 05 1.22 0.00 29.83 069 0.03 0 05 0.21 10.26 99.1 8 53.83 O il 1.74 0.03 0.70 0.00 27.14 5.49 0 07 0.00 0 14 10.81 57.47 0.1 1 1 65 0 02 0 65 0.00 26.98 5.52 0.05 002 0.13 10.41 44 83 0.07 069 0.00 0.62 0 00 23 88 3.67 0.35 0.07 0.10 21.34 1 0 0 0 6 98 01 9 5.6 2 62.26 0.06 0.45 0.18 1.04 0 .0 0 30.00 0 13 0.09 0 05 0.01 5.92 54.45 0.10 4.26 0.04 1.41 0 .0 0 30.79 1.13 0.09 0.02 0.04 7.54 59.93 0.10 0.79 0 .0 0 0 84 0.00 30.40 0.50 0.09 0.02 0.13 5 65 58.54 0.18 111 0.00 1.38 0.01 29.83 M3 0.00 0.00 0.00 5.94 62.19 0.08 0.69 0.00 0 81 0 00 30 19 0 43 0 07 0.00 0 13 5 14 56 14 0.12 1.15 0.02 1.39 0.00 27.90 1.53 0 07 0 00 0 02 9 99 57.34 0.21 2.30 0.11 1 05 0.00 27.**.4 1.69 0.10 0.10 0 7C 8 03 100.19 99.87 9 8.4 5 98.21' 99 7 3 98 7 1 98 20 61.49 0 01 1 20 0 38 107 0 .0 0 30.5 1 0. I(. 5.00 100.1 1 ^A vei.ijy: ut JO u k sam ples. ' Aver.i|*c il ft f.ili .m.lysis in D eri i l .tl. (|9(>2). 1V *>l.iIili s .n r litsl o il l^ n il in n ( l n l .i l II . ( ) , I ( > . , nt|>.iiiit.Sl .ml n l h n v o l.illie s ) - \. lues in p a n s per m illion. The sym bol < indicates the lim n of detection of the analytical m ethod. The symbol > indicates concentrations of Zn are greater .in 1.000 ppm. O OO m' -- -- to V ' V V-- i O ^ v-> X o ooi V V' i sO f*l O O' O' " I i O vvi mO 5 oo V V: o o O'TOr-r'iOcoor-m -- r-< V V: ooi VV --O O' - O' vV>1 o o v'rO> Nom- r'O--vnOvi>vO"i O O wom O -- y r- -- V V mV V oo< VV' -- tt O-- r* co-i -> --O- oo > OO' VV' r- u-i i oo( VV' oo< ---- VV- o o i mo vO^ f-- i Vw Vr , ~ v - V 7- oo CTn -- 1 1 cv>j *vs\ ----o r- m^ O' V 1 O O 1 00 O vr>--O O 1 I wi/i( O--v mN O- VV: ~V V O' O O '--J ^V O^' I om/>yO-->\>orj oO--' O O O r-> -- -- TJ V ' v1 --v ro, o\ "oVw>> V V- ooo m ---- V TM V V <n O--O--0v0i rvi v r _ cn y V V V-- O-- O--mv.0w0i OO1' V V -- 'O O-- O-- ro-'i mW 00 cn V VV- ---- ooo V V "* O I- O COv CONSUMER TALCUM S AND POWDERS the crystal lattice of a mineral acting as a sorting mechanism for cations, the cations can enter a crystal structure providing they have appropriate size and charge. These phenomena apply to major as well as minor elements. Thus, barium is present in large amounts in sample 8, which has a high K jO content (1.37%). Since barium and potassium have similar ionic radii, barium is easily admitted into potassium minerals, such as micas and feldspars (both found in 8). Rubidium and strontium are also enriched in sample 8, since these metals also easily substitute for potassium. Gallium is found in large amounts in samples 8, 12, and 17. These samples arc also very high in A l20 3. Gallium has the same ionic charge and radius as aluminum and, in fact, is found only in aluminum bearing minerals. In four samples (1, 4, 9, and 16) there arc significantly higher concentrations of cobalt, chromium, and nickel than found in the other samples. These four samples also have high contents of FeO (Table 5). The association of these four transition metals has been observed before in certain geochemical environments, particularly in ullramafic rocks. Since talcs derived from the metamorphism of serpentines and peridot it es (ullramafic rocks) arc considerably enriched in FcG (Deer cl at., 1962), it is likely that the divalent cations are substituting for iron in the brucile layer of the talc. .t DISCUSSION AND CONCLUSIONS Talc used in the United States represents a wide range of mincralogica' substances. Industrial grade tales arc obtained from different rock types of highly variable mineral composition with the result that the mineral talc may actually be a minor constituent. However, it has been slated that consumer talcum products should contain at least 90% of the mineral of the same name and no asbestos fiber (Hildick-Smilh, 1976). Review of the literature suggests that at least until 1968, materials that were marketed as cosmetic talcum products did not necessarily conform to these criteria. Talc mineral may occur in a plaly form or in a librous form, la ic fiber may occur as a small proportion of the mineral desposil oi as a major constituent. Intergrowths of talc with other mineral phases aie common. These phases may be simply macroscopic /ones adjoining talc mineral or may occur as microscopic intergrowths within the talc OI the many minerals that may coexist with talc, a number ol ashcsiifoim phases commonly occur: tremolile, anlhophyllile, and chrysolite have been identified in these deposits. In addition, Iree silica (quart/) is a Irequem constituent. The trace metal content may include elevated levels ol nickel, chiomium, and cohall. I here is general agreement between the mineral romposilion and the . i .. ........................... . i it........................ . i ........... i o> f v y d :V fill & y .w ? stt 276 A. N. ROHL ET AC basis of mineral and chemical contents, the type of geological provenance may be ascertained. Methodology has been developed for quantitative X-ray diffraction determination of anthophyllitc, tremolite, serpentine, and quartz in con sumer talcums and powders. Important factors in the calibration standard development include selection of talc and reference minerals and the selection of diagnostic X-ray reflections. The sample preparation technique is sensitive and reproducible. Dilution standards arc step scanned over diagnostic reflection areas, peak areas arc measured, and a set of standard calibration curves is developed by regression analysis. Samples of consumer talcums and powders are prepared and analyzed under identical conditions and compared with the calibration curves, permitting quantitative analysis of these minerals. X-ray diffraction alone cannot distinguish between asbestiform and fragmented forms of anthophyllile and tremolite nor between asbestiform and platy serpentine varieties. Electron microscopic analysis was used to distinguish between these forms. Mineralogical characterization of 21 consumer talcums and powders showed that 10 contained measurable concentrations of asbestiform tremo- lite and anthophyllile, and some also contained fragmented forms of these minerals. Two samples contained trace quantities of chrysolite (0.25-0.5%). These observations were confirmed by transmission electron microscopy. The amphibole phases present in these talcum products ranged in amounts from several tenths of a percent to over 14% by weight. Quartz was present in eight consumer talcs in amounts ranging from 1.6 to 35.1% by weight. Consumer talcum products are for the most part complex mineral assemblages, which confer X-ray sorbing and fluorescing effects that arc not equivalent to, and are usually greater than, those of the binary systems used in preparing the dilution standards. In consumer talcum pro ducts minerals such as talc, micas, chlorite, calcite, dolomite, and others tend to diminish reflection intensities of asbestiform minerals by sorbing X-rays or by contributing to background noise. Also, repeal runs on some selected specimens have demonstrated greater peak areas due to slight modifications in instrumental settings (eg., increase in receiving slit width). Therefore, the values for weight percent concentrations given in this report arc conservative. Examination of the same consumer talcum products by both optical and transmission electron microscopy indicates that not all of the mater ials fall within the definition of fiber or asbestiform. For example, one consumer talcum product that contained more than 7% tremolite was observed to contain both fragmented tremolite grains by optical micro scopy and asbestiform fiber with 3:1 or greater lenglh-to-width ratio by transmission electron microscopy. Optical microscopy may provide useful ............... ......... 1.......... .. " t ' - ri ' i i i n n i .in h e o b t a i n e d h v CONSUMER TALCUMS AND POWDERS microscopy, for example, several samples of consumer lalcum products exhibited both free amphibole fiber, discrete from talc grains, and, in addition, numerous small amphibole fibers were visible, apparently interlayered between talc or chlorite plates (see Fig. 6A). Preliminary examination of the asbestiform amphiboles by an electron microprobe technique has demonstrated that individual fiber chemistry is identical to those fibers encountered in the IARC Asbestos Standards (Timbrell and Rendall, 1971). On the basis of the mineralogical and chemical characterization of these products, all formulated prior to June 1973, we conclude that cosmetic grade talc was not used exclusively. The presence in these products of asbestiform anlhophyllile and tremolite, chrysolite, and quartz indicates the need for a regulatory standard for cosmetic talc. This standard should be cognizant of talc complexities, mineralogical and chemical in nature, and should provide for adequate analytical protocols to ensure monitoring. We also recommend that evaluation be made to determine possible health hazards associated with the use ol these products. tm i f !; i w m if t l r ' APPENDIX A: DEFINITIONS OF TERMS USED IN T E X T Asbestos "A name applied to a group of naturally fibrous minerals" (chrysotile, amositc, crocidolite, tremolite, anlhophyllile cited by name) (Bureau of Mines, 1968). The term asbestos has also been applied to commercially exploited fibrous clays, including attapulgite and palygorskile (Whittaker, 1968). Asbestos implies current or possible exploitation, based on the presence of special physical and chemical properties, determined on the bulk sample level. For example, high liber tensile strength, flexibility, low heal conductivity, high electrical resistance, and chem ical inertness are properties of asbestos. Noncommercial varieties of the same mineral may not possess the same qualities on the bulk level. For example, amositc has been considered to be the economically exploited variety of grunerilc (Deer et al., 1962). If so, and even this is contested among mineralogists today, large macrocrystals are signili cantly different physically and structurally. Grunerilc fiber is rigid, amositc fiber is flexible; grunerilc yields well-defined single X-ray reflections with nonrptational film techniques, amositc yields multiple reflections as if rotated in the X-ray beam; grunerilc appears to be a single crystal, amositc splays as if composed of strands. However, when both substances are pulverized, the resultant powder yields submicrostopic fibers, many of which are virtually indistinguishable on the basis o l morphology, s i m e lure (determined hy selected area election ||l , dv'.'e' ' fe te ' m A. N RO H L ET AL technique). Amosite may be considered as an aggregate of unorienlcd, discrete, gruneritc crystals with only the c axis in common alignment. Comminution of such aggregates produces fibers with character istics identical to those of single crystals of gruneritc that have been similarly pulverized. Some workers have suggested that mechanical size reduction of amosite yields fibers with crystal growth surfaces rather than cleavage surfaces. Since amphibole cleavage tends to parallel prominent crystal face planes, such distinctions on the submicroscopic level may disappear. This appears to be the case for tremolite and anlhophyllilc as well. However, because no methods exist to distin guish between possible differences in fiber surface, we do not refer to anlhophyllilc and tremolite fibers in these talcums as asbestos. Instead they are referred to as asbestiform. It should be stressed, however, that evidence docs not exist that would indicate that fibers with crystal growth surfaces or cleavage surfaces possess lesser or greater biological potential than fibers from commercial asbestos deposits. Asbestiform "Formed like or resembling asbestos; fibrous;. . . . (Bureau of Mines, 1968). The term is used herein for amphiboles (anlhophyllilc and tremolite) seen on both light and submicroscopic examination, which resemble comminuted asbestos varieties, on the basis of morphology. Essentially, when these fibers are derived from commer cial deposits we term them "asbestos" and when analytically identical fibers are found as noncommercial intrusions with the mineral talc, we term them "asbestiform." The use of two terms docs not imply differences that can be analytically determined. Fiber " The smallest single strand of asbestos or other fibrous materials" (Bureau of Mines, 1968). We use this term in a broader sense. For example, chrysotilc fibers are called fibrils, possessing unit diameters of about 200-400 . Coherent bundles of fibrils arc also called fibers. Fiber in the present text is used to denote any elongated single mineral unit visible on the light or electron microscopic level. The Occupational Safety and Health Administration has applied a 3:1 length-lo-width ratio to distinguish fiber from mineral fragment. APPENDIX B: C R YSTA L CHEM ISTRY, CRYSTA L STRU C TU RE, AND GEO LO GICAL OCCURRENCE OF TALC Chemistry of Talc The empirical chemical formula of talc is M g,Si4O lu(OI I ) , , but CONSUM ER TALCUM S AND POWDERS y i? \ \v weight. Titanium and aluminum appear to substitute for silicon, whereas iron, nickel, and manganese substitute for magnesium. Alkali metals arc b h i V not readily accommodated in the structure and evidently occur as inter m layer ions or as components of mineral impurities. For example, excess calcium may reflect the presence of the interlayer mineral phase tremolite (Deer el al., 1962; Stemple and Brindley, 1960). One major talc deposit in vMk ?the eastern United States contains substantial amounts of nickel, as much Ksfci-as 0.2%. Notwithstanding these minor components, talc is essentially (by weight) 32% MgO, 63% S i0 2, and 5% structurally bound water. Talc Structure and Crystal Habit The three-layered crystal structure comprises a sheet of octahedra'ly coordinated Mg(OH)2 groups (the brucite layer) sandwiched between two planes of tetrahedrally linked S i0 4 groups (silica layers). Apical oxygens of the silica sheets are directed toward the brucite layer and in part re placed by hydroxyl groups, which form a portion of the inner structural unit. Valence balance is accomplished within the structure, so that there is a net zero charge on juxtaposed unit layers at the silica base interfaces. The basic unit of the talc structure was determined over 40 yr ago (Gruner, 1934; Hendricks, 1938), yet the repeated cell geometry and space group were only recently resolved. X-ray single crystal patterns now |pEjindicate talc to be triclinic (Rayner and Brown, 1966; Ross et al., 1968). 0 if fIn addition to chemical and structural complexities, talc occurs with both plate and fiber habits (Ford, 1957). The development of the fibrous cyrstal form, with an elongated crystallographic o-axis, may be a mani ,V i f t - s ' 1 festation of ionic substitution since its refractive index is higher than plaly f e t e talc (Fleischer and Osborn, 1957; Gruner, 1944). Talc that contains V i $ . $ A substantial amounts of these elongated forms is referred to in tit, mincralogical literature as fibrous talc. Similar observations with regard "> the mineral brucite have been reported (Liebling and Langcr, 1972) in which high iron content in the normally plaly brucite is associated with the development of a fibrous habit. It is of interest to note that minnesotaitc, considered by some to be an iron-rich form of talc, always occurs with a fibrous, or even a needle-like, habit (Gruner, 1944). Formation of Talc in the Laboratory of Nature In closely controlled experimental systems, talc has been synthesized (Bowen and Tuttle, 1949; Yoder, 1952). Bulk chemistry, water lugg aacciiilyy. temperature, and pressure parameters arc defined within extremely rest l i i i 1 cd limits and ranges to produce relatively pure crystallization p'"i..d...u...e..l..s.. v i w ' i a v However, in laboratory synthesis, just as in nature, coexisting minci.il V", VifJ-VV J phases arc produced if slight variations in any ol the paiamctcis aie ii introduced. These phases include anthophyllite, serpentine, and in some i - V * >5 iim rs iii-molitc. 280 A. N. RO H L E T AL. Nature of Talc Plates Electron microscopic examination of talc minerals demonstrates that single talc grains consist of contiguous single crystals, mosaics of dis oriented crystallites, and intergrowlhs with other mineral phases, particu larly tremolite (Kleinfeld et al., 1973; Stemple and Brindley, i960; Wright, I960). Selected area electron diffraction patterns obtained on these objects display, in order, single crystal arrays, Debye-Scherrcr rings, and superimposed complex patterns characteristic of intergrown single crystal phases. Talc and Mineral Intergrowths Tremolite is one common intergrowth in talc, and it requires relatively little energy thermodynamically to occur. Replacement of magnesium by calcium in the brucilc layer may lead to structural as well as chemical modification (Bragg and Claringbull, 1965). Rotation of unit tetrahedra in talc forms double chains from sheets, readily accomplished by substitution of Mg(OH)2 by Ca(OH)2. The bulk chemistry is thereby changed from Mg6SiBO j0)(OH)4 to Ca2Mg5Si80 22(0 H )2. The final structural array is remarkably similar in both materials; the crystallographic a axis of talc is approximately 5.26 A, which corresponds with the c-crystallographic axis of tremolite (approximately 5.24 A); the 6-axis for talc is approximately 9.10 A, which is equal to the b/2-axis of tremolite; the c-axis of talc, approximately 18.8 A, is about equal to twice the o-dimension of tremolite (approximately 18.2 A). The monoclinic stacking angles, the beta-angle, are within a few degrees of each other. Intergrowlhs may form in which amphibolc formation is not complete so that a mixed phase exists, referred to mineralogically as "talcbolcs." These arc not as rare as once believed and may even be common in the more complex talc deposits. APPENDIX C: IN DUSTRIAL AND COSMETIC GRADE TALCS Studies have demonstrated that industrial talc often consists of a variety of minerals, the utilization of which is based on physical properties rather than mineral composition (Hogue and Mallettc, 1949; Schulz and Williams, 1942; Thompson, 1974; Wells, 1965). Early analyses of cosmetic talc also showed a wide range in mineral composition. Of six such products examined in one study, only 6-47% by weight of the inorganic material that constituted the product was the mineral talc; 14-51%, serpentine minerals; 5-77% carbonate minerals; n 0-trare tremolite; 3-12%. other minerals (Schulz and CONSUMER TALCUMS AND POWDERS 281 APPENDIX D: BIO LO G IC A L HAZARDS ASSO CIATED WITH TA LC EXPOSURE A fine, diffuse, bilateral, progressive fibrosis was observed among miners and millers of Ircmolite talc in Georgia (Dreessen, 1933; Drccsscn and Dalla Valle, 1935). Siegal et al. (1943) studied a population of workers mining and milling tremolitc and anthophyllite-bcaring talc deposits in New York state. In addition to the bilateral fibrosis, pleural plaques, similar to those encountered in asbestos workers, were observed. Review of postmortem material in this study indicated that asbestos bodies were present in lung tissue. These findings were also reported in cases of severe pneumoconiosis in tremolite millers by Daymon (1946), and by Porro and Levine (1946). Millman (1947) reported that exposure to cosmetic-grade talc pro duccd nodular fibrosis in workers. No quartz was delected in the dust. The author concluded that talc itself was capable of producing scarring. The observation was supported in studies by Reichman (1944) and by Wycrs (1949) and in a study of talc miners and millers in Italy where exposure to pure talc produced a 10% incidence of pneumoconiosis in workers (Parmcggiani, 1948). Excess deaths attributed to pneumoconiosis have been reported among workers in northern Italy mining talc con sidered to be free of asbestiform fibers (Rubino et al., 1976). Some investigators have held that fibrous talcs (not differentiated as talc or asbestos fiber) are biologically more hazardous than plaly talcs For example, in a review of the literature by Porro et al. (1942), Gloync and Gardner are referred to as considering lhat the clinical, radiological, and pathological disease states of asbcslosis and lalcosis arc very similar. There are several reports of the occurrence of asbestos bodies in the lung tissue of workers exposed to talc (Daymon, 1946; Hobbs, 1950; Klcinfcld et al., 1973; McLaughlin et al., 1949; Porro et al., 1942) Several studies suggest that fibrous talcs are more dangerous as a rcsull of the included asbestos fiber. For example, McLaughlin et al. (1949) compared fibers in talc with the proportion of fibers recovered from the lung tissue of an exposed worker. A larger concentration of fibers w j . found in the tissues as compared with the raw talc. Talc pneumoconiosis was reaffirmed by Klcinfcld and Mcssilc (I960) in their study of the New York state talc workers. In a study by Klcinfcld et al. (1967) it was demonstrated that talc pneumoconiosis accounted for almost 30% of excess deaths among the talc miners and millers. Most of these were due to the complication of pneumoconiosis, cor pulmonale. However, 21% of the 91 deaths recorded were due to malignant tumors: lung carcinoma, pleural fibrosarcoma, and stomach, colon, rectum, and pancreatic cancers. A peritoneal meso thelioma was reported as well. In addition to these tumors, relroperilon:,.l 11 1 ' * e t c i ki t f n t m d S i . i l M i t . i l i`V.ilu.il i on ol 282 A . N. R 0 H L ET A L . ihcsc data indicaled that a 3- to 4-fold excess of cancers existed in this gioup, as compared to a matched control population. The biological activity of both tremolile and anlhophyllile fibers has been, known for some time, and both have been cited as asbestos minerals by Merewcther (1930) and Noro (1946). Asbestos disease among workers (and others exposed to anthophyllite and tremolile) has been reported (Burilkov and Badajov, 1970; Kiviluolo, I960; Meurman, 1968; Meurman et al., 1974; Schepers, 1965; Wcgclius, 1947; Weiss and Boetlner, 1967). Recent experimental data also indicate that tremolile fibers are biologically active (Graham and Graham, 1967). Some investigators have suggested that inorganic fiber fibrogenicity and carcinogenicity is limited only by its ability to reach the alveolar space (Holt ct al., 1965; Poll and Friedrichs, 1972; Pott ct al., 1974; Robock and Klostcrkdtter, 1976; Stanton and Wrench, 1972). Wagner el al. (1975) reported lung scarring in Wislar rats with pure talc, exposed by inhalation. The severity and extent of the lung scarring was comparable to that produced by chrysolile asbestos under identical experimental conditions. In addition to lung scarring, ingestion of talc was reported to be associated with leiomyosarcoma of the stomach as well as one adenoma and several sarcomas of the uterus. However, the exposure levels were high and the numbers of observed tumors small, so that statistical validation of the carcinogenic potential of pure talc and its relevance to human exposures were not achieved. There arc also extensive data concerning hazards associated with exposure to silica or trace metals, particularly nickel and chromium (National Research Council, 1975). Analytical data are presented here that suggest possible disease potential and the need for investigation in these areas. REFERENCES Bowen, N. L. and T uttle, O . F . 1949. The system M g O - S iO ,- H , O: Bull- Ceof. S u i. Am. 6 0 :4 3 9 -4 6 0 . Bowes, D. R. and Langcr, A. M. 1974. Petrochem istry o l the Manhattan Form ation. Krisidfinikurn 10:39-52. Btagg, R. II. 1967. Quantitative analysis by powder d illra ctio n . In H andbook o f X-rays. New York.: Ml Gr aw-Hill, Bragg. L. and Clatingbull, G. F . 19 6 S . l h t crystal Structure o l minerals. Lon d on: Ucll and Sons. Brindley, G . W. and K urtossy, S. S. 1961. Quantitative determination o l kaolinitc by x i.iy dillraction. Am. Mineral. 4 6 :1205-1215. Bureau of Mines. 1968. Dictionary of mining, mineral and related term s, cd. P. W. Thrush. Washington, D .C .: U S. Government Printing O ffice. Burilkov, f . and Badajov, L . 1970. t in Bcitrag rum Plcuravcrkdlkungcn. Prax. Pnntm al. 2 4 :4 3 3 -4 3 8 endcmischcn Auflrclen doppclsciligcr 1 M M G r.,ih h I I . I dinhart. W S and lig n R M 1968. I i h i . n i s j i u l i n i m i . i l *^*M jp a ; * CONSUM ER TALCUM S AND POWDERS 281 Deer. W. A ., Ilo w ie, R. A. and Zussman, J. 1962. Rock-form ing minerals, vol. 3, S h e d silitates, pp. 2 0 3 -3 7 4 . New Y o rk : Wiley. Drcesscn, W. C . 1933. Effects o l certain silicate dusts in the lungs. /. Indust. Uyg. 15:6 6-78 . Dreessen. W. C . and Dalla V alle. |. M. 1935. The effects of exposure to dust in tw o Georgia talc mills and mines. Publ. Health Repts. 5 0 :1 4 0 5 -1 4 1 5 . Fleischer, S. S. and Osborn, E . F . 1957. Studies of the system iron oxid c-silica -w a tcr at low oxygen partial pressures. E co n . Geol. 5 2 :9 2 3 -9 4 3. Ford, W. E . 1957. Dana's textbo ok o f mineralogy. New Y o r k : Wiley. G raham , |. and Graham , R. 1967. Ovarian cancer and asbestos. Environ. Res. 1 .1 1 5 -1 2 8 Grner, |. W. 1934. The crystal structure of talc and p y ro p h y llile . Zeit. Krist. 8 8 :4 1 2 -4 1 9 Grner, |. W. 1944. The com position and structure of m innesolaite, a com m on iron silicate in iron formations. Am. Mineral. 29:363-372. H endricks, S. 8 . 1938. O n< he crystal structure of talc and p y ro p h y llile . Zeit. Krist. 9 9 :2 6 4 -2 7 4 . H ildick-Sm ith, G. 1976. T alc: Review of epidemiologic studies. /Vor. Br. O ccu p. Health So.. Edinburgh, Sept. 19 75. In press. Hobbs, A. A. 1950. A type of pneum oconiosis. Am . /. Roentgenol. Radiol. Therap. 5 8 :4 8 8 -4 9 7 Hogue, W. L. and M allelle, F . S. 1949. A study of workers exposed to talc and other dusting compounds in the rubber industry. /. Indust. Uyg. Toxicol. 31:359-364. H olt, P. F ., Mills, J. and Young, D. K . 1965. Experim ental asbestosis with four types of fibers: Importance o l small particles. Ann. N .Y. Acad. S e i 132:87-98. H urlbut, C . S., |r. and Williams, O. R. 1935. The mineralogy of asbestos dust. /. Indust. Uyg. 17:289-293. Kivituoto, R. I9 6 0 . Pleural calcification as a roentgenologic sign of non occupational endemic anthophyllite asbestosis: Acta Rad. Scand. 194:1-67. K leinfcld , M. and Messite, J. I9 6 0 . Problem areas in pneum oconiosis. A tch . Environ. Health 5:428-437. K lcin fcld , M., Messite, J.. K ooym an, O . and Z aki, M. H. 1967. M ortality among talc m iners and millers in New Y o rk Slate. Arch. Environ. Health 1 4 :6 6 3 -6 6 7. K leinfeld, M., Messite, |. and Langer, A. M. 1973. A study o f w orkers exposed to asbcstilorm minerals in com m ercial talc m anufacture. Environ. Res. 6 :1 3 2 -1 4 3 . Klug, H. P. and Alexander, L . E . 1954. X-ray ditfraction procedures. New Y o rk : Wiley. Langer, A. M. and Poolcy, F. D. 1973. Identification of single asbestos libers in hum an tissues. In Protecdinys on the biological effects o f asbestos, ed. Bogovsky cl al.. pp. 119-125. Lyon : 1A R C . Langer, A M .. et al. 1973. Identification of asbestos in hum an tissues. /. O ttu p . Med. 15(3): 287-295. Licbling, R. S. and Langer, A. M. 1972. Optical properties ol fibrous brucite from Asbestos. Quebec. Am. Mineral. 57:8 57 -8 64 . M cLaughlin, A ., Rogers. E . jn d D unham , K, C . 1949. la ic pneum oconiosis. Br /. Indnst. Med. 6 :1 8 4 -1 9 4 . Mcrewether. E . R . A. 1930. The occurrence of pulmonary fibrosis and other pulmonary alfccitons in asbestos workers. / Ind. Uyg. 1 2:1 9 8 -2 2 2 , 2 39 -257. M curm an. L. O. 1968. Pleural fibrocaicific plaques and asbestos exposure. / nviron. Rc\. 2 :3 0 -1 6 . Meutrnan, L. <>.. Kiviluoto. R. and llak am a, M. 1974. M ortality and m orbidity among working populations o l anthophyllite asbestos miners in Finland. Br. /. Indust. Med. 3 1:105-1 12. M illm an, N. 1974. Pneum oconiosis due to talc in the cosm etic industry. O n iip . Med. 4 :3 9 1 -1 9 1 . National Research Co un cil. 197 5. Nickel. Washington, D .C .: National Academ y of Sciences. N oro. L. 1946. O n the history of asbcsiosis. At ta Pathol. M icrobiol. Stand. 2 1 :5 3 -5 9 . Panncggiani, L . 1948. I.c pncum oconiosi del m inatori c dei mugnai del lalcn ncl Pinciolese. R,i\s. Med. Inti 17:16-17. P o rro , I . W and I r v in e , N M 19-16. Patho lo g y o l t a li p n e u m m o n io sis w ill* rep ort ol an a u to p sy. N , ,, i h .V S S t .It , M e d . / 1 2 1-25 284 A N. R O H L E T A L Porro, F. W.. Pal ion, }. R. and Hobbs, A. A. 1942. Pneumoconiosis in the u lc industry. Am . /. Roentgenol. 47:507-524. Pott. F. and Friedrichs, K . H. 1972. Tum oren dcr Raite nach i.p. Injektion faserform igcr Slaube. Nuturwissenschafieri 59:318. Po ll, F., H ulh, F. and Friedrichs, K. H. 1974. Tumorigenic effects o f fibrous dust in experimental animals. In viron . Health Persp. 9:31 3-31 5. Rayncr, |. H and Brown, G. 1966. Triclinic form of talc. Nature 212:1 352-1353. Reichman, V. 1944. (jber Talkumstaublunge. Arch. Gewerbepathol. Gewerbehyg. 12:319-322. Robock, K . and K losterkotter, W. 1976. The biological effect of dusts of asbestos and asbestos lem ent products. Proc. Br. Occup. Health Soc., Edinburgh, Sept. /975. In press. Rohl, A. N. and (.anger, A. M. 1974. Identification and quantitation of asbestos in talc. Environ. Health Perip. 9 :9 5 -1 0 9 . Russ, M , Sm ith, W. L. and Ashton, W. H. 1968. Triclinic talc and associated am phiboles from ' Gouvcrneur Mining D istrict, New Y o rk. Am. Mineral. 5 3 :7 5 1 -7 6 9 . Rubino, G. F ., Scansetti, G .. Piolatto, G . and Rom ano, C . A . 1976. M ortality study of talc miners and millers. /. O ccup. Med. 1 8:1 86 -1 93 . Schepeis, G. W. H. 1965. Discussion. Epidemiology of mesothelial tum ors in the Lon d on area. Ann. N .Y. Acad. Sci. 132:579-602. Schulz. R. 1. and Williams, C . R. 1942. Com m ercial talc, animal and mineral studies. /. Ind. Ilyg. 24:75-82. Sicgal, W., Sm ith, A. R. and Grccnburg, L . 1943. The dust hazard in trem olite talc mining, including roentgenological findings in talc workers. Am . /. Roentgenol. 4 :1 1 -2 9 . Stanley, H. D and Norw ood, R. E . 1973. The detection and identification of asbestos and asbesliform materials in talc. Unpublished report for Pfizer, Inc. Stanton, M. F. and Wrench, C . 1972. Mechanisms of mesothelioma induction with asbestos and fibrous glass. /. Natl. Cancer /ns/. 4 8 :7 9 7 -8 2 1 . Stemple. I. S. and Brin d ley, G . W. I9 6 0 . Structural study of talc and la lc-ircm o litc relations. /. Am. Ceramil Soc. 4 3:34-42. Thompson, C . S. 1974. Discussion of the mineralogy of industrial talcs. U.S. Bur. M in a Circ. 1C 863. 22-44. T im rcll, V. and Rendall, R . E. G. 1971. Preparation of the U IC C ( I A R C ) standard reference samples o l asbestos. Powder Tcchnol. 5 :27 9-28 7. Wagner, |. C.. Berry, G ., C o oke, T . |., H ill. R. |., Pooley, F . D. and Skidm ore, J. W. 1975 Anim al experiments with talc. Proc. Br. O n up. Health S oc.. Edinburgh, Sept. I9 7 S . In press. Wegtlius. C 194 7. Changes in the lungs in 126 cases of asbestosis observed in Finland. A cta Radiol. 28 139-152. Weiss. B and B o clin e r, E . 1967. Com m ercial talc and talcosis. A rch . En viron . Health 14 3 04 -3 08 . Wells, |. R. 1965. T alc, soapstone and pyrophyllile. In Mineral facts and problem s. Washington, D .C .: Government Printing O ffice. Whittaker, L . J. W. 1968. The crystal chem istry ol the amphiboles. At la Crystal. 13-291-298. Wiight, II. D. 1960. Optical study o l talc'ircm olite relations. /. Am. Cecamit S o t. 4 3:4 2-4 3. Wn c is , II. 1949. Ashcstos. Postgrad. Med. / . 6 31 -638. Yoder. II. S. 1952. lh c M g O -A l} O , - S i O , - H , O system and related metarnorphic facies. Am . /. R o uen Mem. VoI. 569-627. > t o t it 'C,. r> 7 t. aaif RESEARCH DATA IN TEG R ITY: A RESULT OF AN IN TEG RA TED INFORMATION SYSTEM M. F. Cranmer, L. R. Lawrence, A. K. Konvicka, D. W. Taylor National Center for Toxicological Research, jeffcrson, Arkansas S. S. Herrick Computer Linguistics Inc., Albany, New York The toxicologic problems o f today frequently require long-term, multidisciplinary experim entation involving large number* o f animals. In order to provide the extensive safety evaluation necessary to pro du ce data that can be reasonably extrapolated I humans, automated research support systems have transcended the position o f u sn u l tools and hove becom e an Integral part o f the total design o f experimental protocols. For an automated information system to fully represent the reality o f the experiment, it m ust be able to assure integrity, os well as provide lor the storage, cultulalion, and retrieval o f data values o f the quality an d qu an tity necessary for fulfilling protot ot requirements. Guarantees ogoinst error and loss o f data, in addition to flexibility and easy act ess, must be an inherent part o f the system if the acceptance and t o n fid en , e o f the investigator are to be obtained. This paper discusses the criteria, philosophies, and benefits o f integrated data systems that ensure integrity o f toxicologn rescarth support. INTRODUCTION The National Center for Toxicological Research has as its primary mission the development of methods by which the technical and scientific problems of toxicologic research can be met and the dispersal of this information to the scientific community. One of the major problems facing the toxicologist is handling the vast amounts of data generated by the quantity and variety of experiments in the current toxicologic armamentarium. Whatever the individual problems in the respective toxicology laboratory (i.c., data accumulation, storage, or analysis), the data system can determine the value of the results Regard less of the sophistication of the data systems, if they do not provide a r 1 iniei'.i il v , they arc of no substantive value. I he %'