Document M40o625eN2k6oXdRVGy9V5BoM

DownloadSearchable PDFRandom document
OccUup.Sa.tioDnaElPSAafRetyTaMndEHNeTaltOh FAdLmAinBisOtraRtion WASHINGTON, 0 .C 20210 . Office of the Assistant Secretary Mir 3 ri PLAINTIFF'S EXHIBIT CAM-322 2 0 MAY T977 :Mu-u-.-r > v Enclosed is a copy of A Report of the Fiber Content Eighty Industrial Ifelc Samples Obtained from, and Using the Procedures of the Occupational Safety and Health Administration (OSHA). I" - | The analysis discussed in the report was conducted by the National Bureau of Standards at the request.of OSHA. The purpose of this study was to determine the fiber content of industrial talcs and thereby answer parties regulated by OSHA who challenged data of both OSHA and the National Institi for Occupational Safety and Health. The enclosed final report was received! by OSHA on May 10, 1977. I hope you find that this report addresses your concerns and consider.it, as does OSHA, a means to.assure healthful working conditions for employees. Sincerely, .r &u<U. , Eula 8ingham - :- Assistant Secretary Occupational Safety and Healthy ,, Enclosure 1 '* ; * . . . - V- f r - i( s ^ - ^ 0 CAMC-Abeyta-001872 4* .r-<l-v:4.hh^ey' . -rT-^ f?.i-*.To**.-' V'" -. . *': .<-< TALC SAMPLE NUMBER 39 40 *-' . tta tiS X '"-'-- '.......................... ' ; . . . . _ 4 -- 4 :.. s-- i a:***$.isr ^ i v . - '' '" ' '* ... -..- ', /', ' .COMPANY AND DESIGNATION Vanderbilt Talc Company .Nytal 300 Vanderbilt Talc Company Nytal 400 * . r*r ' * Vanderbilt Talc Company 5X - **CC.TOCv ` Vanderbilt Talc Company 325 - - - . Ci": Vanderbilt Talc Company '~- .'Ji*tV-*yi.Atrjl,-1 " : '.?: *v, ' '-- '\ r 4 Vanderbilt .Talc Company >?* -:rv- FT . / : . ,i,.Vanderbilt Talc Company 4i24*fc-'*" _' * ; 3Y . . - .. ` v' : ^ 4 <-'> ' />-T :'* -'v*. . ., : -' r--' " Hitchendc Corporation;, y ' "' -'.-4 -Murphy,.N.-C. .,,vf Produc. t T.a.. l,, c . V- ' - * Xliv- - - *v Windsor Minerals, Gassets "37" : 4 July 29, 1' 9 '75'' . ..' -"'^Windsor Minerals, West Windsor 'Mineral Ray Grind Composite Windsor Minerals, Ludlow "36" * July 29-30, 1976 jftijjlsor Minerals, "Gassetts" ^ * 3 6 , July 28-30, 1975 Vermont Talc Company -- Bagging Bulk - Sample, 7/16/75, 2nd shift - w .- - Vermont Talc Company - Bagging ' r Bulk Sample, vertah XXXI) ,:is- < ; ''r.f- 4 9/17/76 : "1st. sh; ift -..'' ; v _ - CAMC-Abeyta-001873 COMPANY AND SAMPLE DESIGNATION Southern Talc Co. (Cohutta Mill) Chatsworth, Ga. Product (Powder) Sample Mine, MESA I.D. #04-00493 Packer Station #3, Vanderbilt Talc, Mill #1 Gouverneur, N. Y. 1st shift 11/4/75 (FT) . Packer Station #2, Vanderbilt Talc, Mill #1 Gouverneur, N. Y. (Nytal 100), 1st shift 11/4/75 Ceramitalc HDT, Vanderbilt Talc Co. 1st shift 11/4/75 Bagging Station #3, Vanderbilt Talc Co., Mill #1, Gouverneur, N. Y. 11/4/75 (Nytal 200) Hitchcock Corp. #5 Mine MESA ID# 3100222 Hitchcock Corp., #3 Mine MESA ID# 3100683 Wrau Mine 3 mill, Int. Minerals and chemicals MESA ID# 31-0008 Bagging Station, Vanderbilt Talc Co. 2nd shift, Mill #1 Gouverneur, N. Y. (Nytal 300) 11/3/75 Bagging Station, Vanderbilt Talc Co., 2nd shift, 11/3/75 (Nytal 200) Southern Talc Co., Southern Mill Product Talc sample, Chatsworth Ga. Rock Cliff Mine, MESA ID# 0900492, 9/3/74 Southern Talc Co., Southern Mill, Chatsworth, Ga." ore from Rock Cliff MESA ID# 90-00492 Southern Talc Co., Southern Mill Ore from Jude Hole Mine MESA ID# 09-00492 CAMC-Abeyta-001874 TALC SAMPLE NUMBER 55 56 58 59 60 61 62 63 64 - 3- ` COMPANY AND SAMPLE DESIGNATION - Winterborg Mine & Mill, Ore from American Talc Co., Alpine, Ala. Willowreek Mine, MESA ID# 01-00433 9/4/75 Thompson-Weinman Co., Cartersville Plant, Cartersville, Ga., 9/5/74, Ore from Brady Mine, MESA ID# 09--00224 Southern Talc Co., Cohutta Mill Chatsworth, Ga. 9/3/74, Ore from Ernest Mine, MESA ID# 09--00493 American Talc Co., Alpine, Ga. Winterboro Mine & Mill, 9/5/75 MESA ID# 01-00433, Final Prod. Pioneer Talc Co. Inc., Garren Pit. Prod. #6 Allamore, Hudspeth Co., Texas MESA ID# 41-00869 Pioneer talc Co., Inc., Apache Pit Product, #4, Texas MESA ID# 41-00869 ` Pioneer Talc Co. Inc., #4 Texola Pit Product, Texas ' MESA ID# 41-00869 Southern Clay Products, #2 Bottom Product, Allamore, Hudspeth Co., Texas #18 Milwhite Co. Inc., Van Horn, Cumberson Co., Texas MESA ID# 41-01527 Ground Calcined Product Milwhite Co. Inc., #11, Van Horn, Culberson Co. Texas, Eagle Hat Pit Crude MESA ID# 41-01527 Pioneer Talc Co. , Allamore, Hudspeth Co. Texas, #8 Garren Ceramic MESA ID# 41-00869 CAMC-Abeyta-001875 TALC SAMPLE NUMBER 65 , 66 67 68 69 , 70 ( !' 1 71 72 73 74 75 76 78 ' 79 - 4- COMPANY AND RAMPT.K TVERTGNATTON Milwhlte Co. Inc., Van Horn, Texas #14 Tumble Down Pit Product, #14 MESA ID# 41-01527 Eastern Magnesia Talc Emtal 42, - Co. Vermont 8/28/75 Pfizer (Barretts Mill) Montana minus 12 Micron Pfizer (Barretts Mill) Montana minus 325 mesh Pfizer (Barretts Mill) Montana minus 20 microns Cyprus Industrial Minerals (Three Forks Mill) Montana minus 325 mesh Cyprus Industrial Minerals (Three Forks Mill) Montana Mistron Vapor (minus 10 micron) Cyprus Industrial Minerals ^ (Three Forks Mill) Montana Floor Sweepings (Bagging Area) Cyprus Industrial Minerals (Yellowstone Mine) Montana Drill Hole Southern Talc (Southern Mill) Ga. 981 passing 200 mesh Southern Talc (Southern Mill) Ga. 99% passing 325 mesh (#1767) Southern Talc (Cohutta Mill) Ga. 95% passing 325 mesh Milwhite Talc Co. Inc., (Tex.) Finished CM4 product (#2) Westex Talc Co. (TDM) Dark - (Milwhite Inc.) (#1) Tex. CAMC-Abeyta-001876 TALC SAMPLE NUMBER 80 81 82 83 84 86 87 88 89 91 92 93 94 95 96 - 5- COMPANY AND SAMPLE DESIGNATION - Milwhite Co., Inc. (Tex) Marshel mix - 92-200 mesh (#3) Westex Talc Co. (Milwhite Inc.) (Tex.) White mine 93-200 mesh (#4) Southern Clay Products - Tex. Garren Mine #2 (#5) Texas Talc Co. - TPI (Tex.) Mine (Texas Pacific) (#6) Texas Talc Co. Tex. David Mine (#7) Cyprus Industrial Minerals Co. Loyce Fine Mine (Tex.) #9 Pioneer Talc Co. Inc. (Tex.) Texola Mine (#10) Pioneei Talc Co. Inc. (Tex.) Dees Mine (#11) Milwhite Co. Inc. Bryant Soapstone Mine Arkansas Laws Mill Micro talc from White Eagle Mine Standard Industrial Minerals Laws Mill Desert Talc Product Standard Industrial Minerals Laws Mill Holiday Talc Product Standard Industrial Minerals Keeler T-076 Tecopa Mine (100% Panamint) Cyprus Industrial Minerals, Montana Keller HGO-55 Blend 4 Panamint, 3 Oasis, 3 White Eagle Cyprus Industrial Minerals, Montana Keeler Sierra Cloud Blend 4 Panamint, 1 White Eagle Cyprus Industrial Minerals, Montana CAMC-Abeyta-001877 * \ TALC SAMPLE NDMBER 97 98 99 100 101 102 103 104 105 106 107 C 108 ( 109 BKK/mmc:3-21-77 Retyped/atb/5-20-77 6- - COMPANY AND SAMPLE DESIGNATION . Keeler, MSC Mistron Panamint + 0.52 stearite Cyprus Industrial Minerals Keeler Sierralite Talc City (Frisco) Cyprus Industrial Minerals . Keeler Glacier 200 Talc City (Tailings) Cyprus Industrial Minerals Keeler DSP Oasis Mine (1002) Cyprus Industrial Minerals Keeler White Eagle Screenings Cyprus Industrial Minerals Keeler _ Suggar Mine Cyprus Industrial Minerals Pfizer Inc. (Cal) Talcron CP 44-31 Eclipse, Bonnie, Mongolian Pfizer Inc. (Cal) Cercron CF 96-38 Eclipse, Acme, Mongolian Pfizer Inc. (Cal) Cercron CF 96-36 White Eagle, Mongolian, Apex Cyprus Industrial Minerals (LA) Furnace Creek Tecopa (1002 Panamint) Cyprus Industrial Minerals (LA) Vanderbilt Co., crude Soft LB - American Talc (Alabama) 84 Color Dark 200 Mesh Willow Creek Mine American Talc (Alabama) 92 Color 99% 400 Mesh Willow Creek Mine CAMC-Abeyta-001878 iimi-- nhrtatfe. r r A REPORT ON THE FIBER CONTENT i: OF EIGHTY INDUSTRIAL TALC SAMPLES OBTAINED FROM, AND USING THE PROCEDURES OF, THE i OCCUPATIONAL SAFETY AND HEALTH ADMINISTRATION i r: L c { Prepared for Occupational Safety and Health Administration Department of Labor Washington, D. C. 20212 r L Prepared by the Staff of the Analytical Chemistry Division, P. D. LaFleur, Chief Institute for Materials Research National Bureau of Standards Washington, D. C. 20234 May 1977 ! . I U.S. DEPARTMENT OF COMMERCE, Juanita M. Kreps, Secretary Dr. Betsy Ancker-Johnson, Assistant Secretary for Science and Technology N A TIO N A L B U R E A U OF S T A N D A R D S . Ernest Ambler, Acting Director CAMC-Abeyta-001879 t I. INTRODUCTION A. Purpose of Study This report has been prepared in response to a request received by Dr. John D. Hoffman, Director of the Institute for Materials Research of the National Bureau of Standards (NBS), in a letter dated September 1, 1976, from Dr. Morton Corn, Assistant Secretary of Labor, Occupational Safety and Health Administration (OSHA). In that letter, Dr. Corn stated that analysis of talc samples for their asbestos content was being performed by OSHA and the National Institute of Occupational Safety and Health (NI05H) but that the methodology was being challenged by some of those regulated by OSHA. Dr. Corn indicated that his request to NBS was composed of two tasks: (1) to resolve the variability in the definition of asbestos fibers in talc, and (2) to determine the asbestos content of some 80 talc samples to be provided by OSHA. Copies of the letter containing'this request and the subsequent correspondence between Drs. Hoffman and Corn which provide further background information, are included in this report as Appendix I. It was agreed by both parties that the first of these tasks, i . e resolving the variability of the definition of asbestos fibers, would be a complex, long-term program which would require input from a number of sources both in the private and public sectors. In view of this, it was agreed that the more limited task, that of determining the asbestos content of the OSHA talc samples, would be addressed by NBS first. As Dr. Corn pointed out in his letter of October 8, 1976, this was recognized not to be a research task but would involve an analysis performed according to the procedures given in 29 CFR 1910.1001. 1 CAMC-Abeyta-001880 29 CFR 1910.1001 [39 FR 23502, June 27, 1974, as amended in 41 FR 11505, March 19, 1976] deals exclusively with airborne asbestos and does not describe procedures for, nor make any reference to asbestos in talc. It does, however, include the following two paragraphs which are pertinent to the work discussed in this report: "(a) Definitions. For the purpose of this section, Cl) 'Asbestos * includes chrysotile, amosite, crocidolite, tremolite, anthophyllite, and actinolite. (2) 'Asbestos fibers' means asbestos fibers longer than 5 micrometers. * "(e) Method of measurement. All determinations of airborne concentrations of asbestos fibers shall be made by the membrane filter method at 400-450 X (magnification) (4 millimeter objective) with phase contrast illumination." These regulations do not, however, describe a measurement procedure but rather prescribe a method of measurement (viz., phase contrast microscopy). A procedure contains a detailed listing of the sampling of the material, the specific experimental steps to be performed during an analysis, and, often, descriptions of the mathematical calculations to be performed and of the format for reporting results and their associated errors. A method of measurement is defined by a very general statement of the type of measurement to be made, from which a specific procedure is developed. Since 29 CFR 1910.1001 stipulates only that the method of phase contrast microscopy is to be used for the determina tion of asbestos, NBS requested a detailed procedure from OSHA. In response, NBS was provided with a copy of the OSHA document "Asbestos Fiber in Air" - Method No. PCAM 239, 2 CAMC-Abeyta-001881 issued March 30, 1976. NBS was also informed that the determination of asbestos, including asbestos in talc, was performed at the OSHA Salt Lake City, Utah, laboratory (OSHA-SLC), using PCAM 239. Method No. PCAM 239, however, deals only with the determination of asbestos in air using phase contrast microscopy to examine membrane filters. Since the determination of asbestos in talc is not described in the document, NBS was told to contact Mr. Willard C. Dixon at OSHA-SLC to obtain an exact description of the procedure used on talc samples. To become familiar with the asbestos in talc procedure, an NBS scientist then visited the OSHA-SLC laboratory and obtained verbal and written descriptions of the procedure used. He spent approximately one and one-half days at the laboratory observing the procedures and techniques used and discussing them with the OSHA employees. At the conclusion of the visit he wrote a detailed report of what he had observed. This report was submitted to Mr. Dixon for comment. A copy of the report is included as Appendix II to this document. The portions of the report that are enclosed in boxes are the comments added by OSHA-SLC personnel. NBS scientists also discussed the analysis of talc with a number of other persons. Those contacted are listed in Appendix III to this report. Prior to commencing the actual analysis of the 80 talc samples, we were assured by OSHA personnel that 29 CFR 1910.1001, PCAM 239 and the annotated NBS trip report were the only documents appropriate for documenting the analytical procedure to be employed, and that the process followed by NBS in developing the detailed procedure was proper. 3 CAMC-Abeyta-001882 midi / tV: \ The detailed procedure followed at NBS is given in Section II of this report; however, a brief description of the procedure, and some comments, are given below: Phase contrast microscopy, with a mounting medium having an index of refraction of 1.546, is used. Both fibers (a fiber being defined as having a minimum length of 5 pm, a maximum diameter of 5 ym, and a minimum length to diameter ratio of 3:1) and other particles are counted. I I I. r Early in the study we determined that the identifi cation of "asbestos" with the procedure used was extremely difficult for the following reasons: (1) Since phase contrast microscopy is only a ( contrast mechanism, it does not indicate the degree of * difference between the refractive index of the liquid and a particle or fiber. . For example, if a specimen is mounted in a liquid which matches a refractive index of chrysotile, then the particles and fibers counted would include all fibers which do not have that particular refractive index, e.g., talc, anthophyllite, wollastonite, fiber glass, etc. The same would be true for any other liquid used. (2) Most mineral species have three refractive indices, therefore, any mineral fiber not lying in the correct orientation for the selected liquid will not match and would be visible and would be counted. ' ijJ \u (3) The amphiboles are end members in solid solution. As a result there may be a large range of ( refractive indices from one end member to another. One example of such a series would be the tremolite-actinolite solid solution. 4 CAMC-Abeyta-001883 (4) In addition, the most common methods of talc formation are the hydrothermal alteration of ultrabasic rocks such as serpentine and tremolite and the thermal metamorphism of siliceous dolomites. Therefore, during the formation of talc in contact with other minerals, there may be extensive interconversions between talc and the minerals serpentine, tremolite and anthophyllite. These interconversions may give rise to single particles which have a combination of the talc, anthophyllite, and serpentine mineral phases.* (5) There are many materials which may be present in talc which have overlapping refractive indices, 6 9 Mineral Wollastonite Tremolite Quartz Talc Chrysotile Refractive Index 1.63 1.599 - 1.637 1.55 1.539 - 1.589 1.493 - 1.567 (6) Even if the refractive indices were known, positive identification of minerals could not be made since there may be interferences from other materials. * Deer, W. A., Howie, R. A., and Zaussman, J . , Bock Forming Minerals, Vol. 3, Longmans, Green and Co. Ltd, London, England (1967) pp. 126-128. 5 CAMC-Abeyta-001884 In his first letter to NBS, Dr. Corn requested a determination of the "number of fibers per unit weight or unit volume. Also, the asbestos content as percent of total weight...". This was to be based upon the definition of asbestos to be resolved by NBS. Since the definition of asbestos is to be investigated as part of the longer study, reporting results obtained in these analyses in terms of "asbestos" was not possible. The procedure used by OSHA-SLC did not use weight or volume of total sample, but instead used percent fibers as a function of total numbers of particles on a slide; that convention was followed at NBS. Thus, all data given in this report are on the basis of numbers or percent of fibers per total number of entities (particles plus fibers) counted (this is referred to as number per^ftii't). B. Goal of the Study In view of the current limitations both in the definition of asbestos and in the ability to distinguish minerals by the OSHA methods using the existing definition of asbestos, NBS established as the goal for this portion of the study to analyze the 80 talc samples supplied by OSHA to NBS and to report the number percent (and, if possible, the limits of error) of fibers using the phase contrast optical microscopy procedure discussed above and outlined in detail in Section II below. 6 CAMC-Abeyta-001885 II. Analysis Procedure 1. Equipment used by NBS a. A Leitz* Ortholux 1 microscope equipped with lOx and 40x phase contrast objectives and a phase contrast condenser. b. Cargille refractive index liquid (mounting medium) 25 n C = 1.546 c . Porton Reticle d. lOx Periplan oculars 2. Specimen Mounting Procedure Two drops of the mounting medium were placed on a clean microscope slide. The end of a new (clean) metal wire (paper clip) was dipped into the mounting medium on the slide and then into the sample which was contained in a bottle. The material adhering to the wire was blended into the mounting medium on the slide. A cover slip was then placed over the preparation. Note: Ho attempt was made to homogenize the contents of the sample bottle prior to drawing the specimen. A vigorous blending or mixing could itself lead to mechanical breakdown of varticles into fragments with fibrous appear ance or the separation of one fiber into many fibers. HBS did not attempt to homogenize * In order to.describe materials and experimental procedures adequately, it is occasionally necessary to identify the sources of commercial products by the manufacturer's name. In no instance does such identification imply endorsement by the National Bureau of Standards, nor does it imply that the particular product is necessarily the best available for that purpose. 7 CAMC-Abeyta-001886 the samples as the OSHA-SLC procedure does not include homogenization. 3. Sample Natation Upon receipt of the samples, a scientist who was not involved in the analysis divided the samples and assigned internal NBS numbers. This prevented the analysts from knowing which of the 80 samples they were dealing with and permitted blind replicate analyses. Note: In the original transfer of the samples from OSEA to NBS a list was inadver tently included which gave the sample identi fications. As soon as this was discovered, the list was sealed in an envelope and placed in the NBS Security Office safe until it was returned to OSHA. The only person at NBS who saw the list did not participate in any any of the analytical work. 4. Microscopy Techniques Phase contrast microscopy, was the method used by NBS, since it is the only technique specified both in 29 CFR 1910.1001 and in OSHA Method No. PCAM 239. The detailed procedure followed by NBS is described below. Any steps which differ from the OSHA-SLC procedure are noted. (1) Before the analysis of a specimen was performed, the microscope illumination was adjusted for Kohler illumin ation, the annular diaphragm and phase-shifting elements were aligned and the Porton reticle was calibrated against a stage micrometer. The configuration of a Porton reticle is shown in figure 1. A detailed description of both the calibration and use of this reticle is given in PCAM 239. 8 CAMC-Abeyta-001887 (2) The counting field was defined as the six rec tangles on the left half of the Porton reticle. (3) Fiber dimensions were determined by comparing the length and width to the diameters of the calibrated circles on the Porton reticle. (4) The unknown samples were mounted as described above in a Cargille liquid with a refractive index of 1.546. Note: OSHA mounted their samples in liquids n r Op with n equal to 1.546 and 1.47. The 1.546 index was used to identify quartz and ohrysotile which appear blue in bright-field observation with their Zeiss phase-contrast microscope, while the background is light brown and most other materials are brown or black. The 1.47 refractive index liquid is carried over from techniques for identifying airborne fibers collected on membrane filters as opposed to the 1.546 liquid used for talc analysis. (5) Counting fields were selected by advancing the mechanical stage in a pattern of traverses of the slide in one axis and steps along the other axis resulting in a rectangular zig-zag pattern. This procedure was followed until the necessary number of fields was counted. (6) NBS used the definition of (asbestos) fibers, discussed above, viz., a fiber is greater than or equal to 5 ym in length, has a length to width ratio of at least 3 to 1, and has a maximum diameter of 5 ym. Any particle meeting these criteria was counted as a fiber. 9 CAMC-Abeyta-001888 t?r' (7) The samples were continuously viewed over the range of focal planes covering the sample thickness during the counting. (8) A preliminary scan of the slide was performed to attempt to find fibers in the specimen. If a fiber was found the following counting procedure was performed on newly prepared slides. Fibers were counted in as many fields as necessary to yield a total fiber count of 100 with the . following exceptions: (a) The analysts counted at least 20 fields even if they counted more than 100 fibers, and (b) they stopped at 100 fields even if they did not count 100 fibers. (9) For fibers and particles that crossed either one or two sides of the counting field, the following procedure was used to obtain a representative count: The analysts counted all particles or any fibers of the correct dimensions which: 1. Were entirely within, the counting area, or 2. Crossed the left or bottom sides, or 3. Crossed the upper or lower left corner, or 4. Crossed both the top and bottom sides. Any other fibers and particles were not counted. (10) When particle agglomerates covered a large portion of the field of view, the field was rejected and replaced by a new one. (11) The measurements at NBS were performed @ 500x magnification. OSHA measurements were done @ 400x to 450x magnification, but NBS could not duplicate that magnification with the equipment available. Calculations performed indicate that the differences in magnification would not result in any significant changes in the number of fibers and particles identified. 10 CAMC-Abeyta-001889 (12) The numerical aperture, NA, of the NBS objective was 0.65. OSHA specifies a NA range of 0.65 to 0.75. With' an objective having NA = 0.65, and at a total system ma g nification of 500x, the smallest diameter of a suspected fiber which could be seen was approximately 0.4 pm. Hence, any fiber 5 pm long but having a diameter less than 0.4 pm, could never be found unambiguously. The class of particles defined as fibers and visible by the procedure is depicted as the cross hatched portion shown in figure 2. Note: The OSHA-SLC laboratory t -in addition to phase-contrast microscopy, sometimes uses polarized light microscopy with a retardation plate to distinguish asbestos fibers from talc plates oriented on end and from other non-asbestos fibers such as glass or organic fibers which may have been present in the samples. Since these techniques are not part of the phase-contrast method they were not used at NBS. NBS used two analysts working independently and following the procedure given above in obtaining the results on the OSHA-supplied talc samples. While the analysts did not have, prior to this study, an extended period of experi ence with the specific methodology employed, they have had extensive experience in a wide range of particle characteri zation techniques, including general optical microscopy. 11 CAMC-Abeyta-001890 III. Pr eli mi nar y -Investigations of Samples A. Macrscopic Description of Samples ' A qualitative description of the 80 samples supplied to NBS by OSHA is given in table 1. The sample numbers are those a s s i g n ^ by OSHA. The number after the description refers to the number of the. color chip in the Inter-Society Color Council'NBS Color Name Chart (Supplement to NBS Circular 553). B. Sc&mning of Samples . Specimens of the 80 talc samples were mounted on slides by eacita analyst independently, using the "dip" method described in Section II.2 above. These specimens, each containing V L & 0 0 particles, were -then scanned rapidly, i.e., only tBre presence or absence of fibers was noted and n b e r count* were not made at this time. As a result of these scans* specimens taken from 31 samples were identified by both analysts as containing at least one fiber, specimens taken froa in aadditional 21 samples were identified as containing at. least one fiber by one analyst but no fiber was seen fry tthe other analyst, and no fiber was seen by either aalrstt for specimens taken from the remaining 28 samples. IV. Courd..^^ and Analysis of Samples A. Sampnles Noted by Both Analysts as Containing Fibers As noted above, 31 samples were identified by both analysts s containing fibers during scanning. Data for these sar. >les are given in table 2. The observed counts for each spec '.en by each of the two analysts, listed as F/(P+F) (t.e., fi: rs divided by non-fibrous particles + fibers) are shown. I- addition the estimated percent of fibers (p ) , and the lower (L) and upper (U) limits of a 95 percent confidence interval or the -true" value of p for the slide on which the 12 CAMC-Abeyta-001891 count was made are given.* It should be noted that the - confidence interval refers only to the portion of fibers on the slide on which the count was made, and not necessarily to the true average of fibers in the entire sample. (Estima tions of the latter would require that the slide constitute a random sample of a homogeneous material under examination.) It is seen that for eight of these 31 materials, the confidence intervals for replicate slides failed to overlap, indicating a lack of statistical compatibility between these replicate determinations. To study in more detail the consistency (or lack thereof) of the results from replicate slides, a more exact statistical analysis was performed by considering the actual fiber and particle counts for each material as the elements of a contingency * The calculation of confidencelimits, and other statiscal analyses in this report, are valid under the assumptions (i) that the fibers are randomly located among the particles on a slide, and (ii) that the stopping rule of Section 11.4(8) above introduces negligible bias. The confidence interval is based on the relation: Prob [at most c occurrences of event E in N trials] ** Prob where P is the probability of an event E, F is the F statistic with n^ and TI2 degrees of freedom, and n^ = 2(c+l) TI2 = 2(N-c) This relation is cited in Fisher, R. A., and Yates, F . , Statistical Tables, Hafner Publishing Company, N. Y. (1963) p. 3. 13 CAMC-Abeyta-001892 table* and calculating the corresponding chi-square value*. A significant value for chi-square, at the 5 percent level of significance, was taken to indicate statistical incompati bility between the replicate slides. The samples for which significance at the 5 percent level was thus determined are identified by an asterisk in table 2. Of the 31 materials, 17 showed incompatibility between replicate slides at the 5 percent level of significance. The fact that over 50 percent of the results of these 31 samples were incompatible suggests problems either of inhomogeneity, or of difficulty in the determination of fiber morphology, or both. If the sample is inhomogeneous, the reason for the statistical incompatibility is obvious. The problem of determining fiber or particle morphology is very difficult as there is often a great deal of subjectivity in deciding whether or not a particle meets the criteria to be classed as a fiber. Photomicrographic examples of (1) a specimen containing unambiguous fibers, (2) a specimen in which no fiber can be discerned, and (3) a specimen which required a great deal of judgment are shown in figure 3, a, b, and c respectively. In figure 3 six different fields of view are shown for each specimen, three using an optical microscope and three using a scanning electron microscope (SEM). (The SEM fields do not correspond to the optical fields, but were chosen randomly.) In the case of talc samples, one must also judge which of the apparent fibers are talc platelets seen on edge. It is possible to "roll over" some of these platelets by moving the cover plate slightly, but while doing so, one also may "roll" another platelet into a position where it would subsequently be viewed from the edge. * Snedecor, G. W . , and Cochran, W. G . , Statistical Methods, 6th Ed., The Iowa State University Press, Ames, Iowa (1967). E q n . 8.10.3, p. 217. 14 CAMC-Abeyta-001893 B. Samples Noted Initially by Both Analysts as not Containing Fibers Because of our concern over the incompatibility of results in the 31 samples in which both analysts agreed there were fibers, we decided to reexamine a portion of the group of 28 samples originally classed by both analysts as not containing fibers. Ten samples were selected at random from the pool of 28 samples and were scanned a second time. (Before discussing the results of additional scanning and counting in this group of samples, it is useful to calculate the probability of seeing no fibers in a count of N objects (particles or fibers). The probability is a function of the true fiber content of the material from which the slide is prepared. The calculation is based on the assumption that the specimen (i.e., the portion of the material on the slide) is a random sample from the material being measured. According to the binominal distribution,* the proDability (zero fibers in count of N) equals (1-P) where P=j q F and p is the number percent fiber content of the material. Table 3 lists values of this probability for various values of p, for N= 200, 500, 1000, and 2000. It is evident that the probability of seeing no fibers in two slides of 1000 counts each (N=2000) is extremely small unless the true fiber content of the material is less than 0.25 percent. (Even for N=200, the probability of seeing no fibers is very small unless the fiber content of the material is less than 2.5 percent). One would therefore expect the materials in this group to have fiber contents not greater than 0.25 pe rcent. It is also interesting to note that for materials with very low fiber content, say 0.01 percent, there is only an 18 percent probability of finding at least one fiber in a count of 2000. In order to reach a probability of 95 * Ibid.j Chap ter 8. 15 CAMC-Abeyta-001894 percent of finding at least one fiber in such a material, one would have to examine at least 30,000 particles. For a material with a 0.1 percent fiber content, the number of particles required to find at least one fiber with 95 percent probability is 3000.) As a result of the second scan on newly prepared slides, four samples were still identified as fiber-free, but fibers were observed in the remaining six samples. The results are given in table 4. Four samples showed fiber contents larger than 0.25 percent by both analysts. Furthermore, for two of these samples, the two new slides gave incompatible results (as shown by the chi-square test). These are indicated by an asterisk in table 4. The results obtained from this reexamination of the materials in the group of initial negative scans thus support the interpretations given by the analysis of the samples in the first group: inhomogeneity and/or subjectivity in deciding fiber morphology. C. Samples Noted as Containing Fibers by Only one Analyst During the Initial Scan The remaining pool of 21 samples, which were noted as containing fibers by only one of the analysts during the initial scan, was then examined. In all of these cases the percent of fibers was estimated to be low and the arguments given above about the probability of finding fibers as a function of total number of particles are equally valid. In this case, eight samples were selected for fiber counts. Of the eight samples selected, seven provided incompatible replicates, given further support to the interpretation made on the basis of the other two groups of samples. These data are given in table 5. Again, incompatible results are indicated with an asterix. 16 CAMC-Abeyta-001895 V. Conclusions In this report the results of the determination of the fiber content of 80 OSHA-supplied samples of talc are presented, along with a statistical interpretation of these results. Of the 45 samples on which fiber counts were made, the results for replicate analyses on 26 of the samples were statistically incompatible. This incompatibility was not confined to any particular concentration range of fibers. As a result, NBS deemed it inadvisable to report uncertainty limits for the fiber content determinations obtained on the samples measured. Fiber counts could be made using the present procedure on the 35 samples which were not counted, but it is doubtful that any useful additional information would be obtained. The variability of these results raises several questions regarding the OSHA procedure, particularly sampling technique, sample homogeneity, and determining fiber morphology. It is the opinion at NBS that, even under favorable circumstances (e.g.3 homogeneous samples, easily identified fibers, etc.), the existing OSHA procedure is useful only for determining "fiber'V conte.n-t and not "asbestos1* content. Although careful manipulation of the mounting medium might make it possible to identify some of the fibers as "asbestos" , the problem of the definition of "asbestos" still remains. NBS believes that the resolution of the measurement problem, including the definition and identification of asbestos, will be accomplished only by significant changes in the procedure and probably the method as well. In order to complete the tasks requested by Dr. Corn in his letter of September 1, 1976, it will be necessary to arrive at an acceptable definition of asbestos and to develop the necessary measurement techniques and standards. Once that has been achieved, a more meaningful analysis of the 80 OSHA-supplied talc samples can be accomplished. 17 CAMC-Abeyta-001896 Table 1 Description of bulk material Description 1 fine white powder (263) 2 fine white powder (263) 3 fine white powder (263) 4 fine white powder (263) 5 fine white powder (263) 6 fine white powder (263) 7 fine white powder (263) 34 fine white powder (263) 35 coarse grains (sand-like) (92) inhomogeneous 36 fine off-white powder (93) 37 beige fine-grained non homogeneous some dark material, some light (93) 38 beige fine-grained non homogeneous some dark material, some light material (93) 39 off-white powder (92) 40 fine very light grey powder (92) 41 fine light grey powder (264) 42 fine white powder (263) 43 fine white powder (263) 44 fine white powder (263) 18 CAMC-Abeyta-001897 Table 1 (continued) Description fine white powder (263) fine white powder (263) fine white powder (263) light brown powder (264) off-white powder clumps together in small aggregates (263) off-white powder clumps together in small aggregates (263) light grey powder (154) grey granular material, non homogeneous (155) light grey powder (fine) (154) fine white powder (263) fine off-white powder (92) grey granular material, non homogeneous (154) fine white powder, some aggregation (263) fine off-white powder (263) off-white powder clumps together in small aggregates (92) off-white powder clumps together in small aggregates (265) dark grey granular, some small fragments (265) fine off-white powder (263) fine off-white powder (9) 19 CAMC-Abeyta-001898 Sample 64 65 66 67 68 69 70 71 72 73 74 75 76 7S 79 80 81 82 Table 1 (continued) Description fine brown-grey powder (264) fine grey powder (265) fine off-white powder (264) off-white powder, some aggregation (263) fine white powder (263) off-white powder, some aggre gation (263) fine off-white powder (92) off-white powder, some aggregation (263) off-white powder, some aggregation (92) small light grey pebbles and powder, inhomogeneous (264) fine light grey powder (264) fine off-white powder (264) fine grey powder (154) ' lignt grey to brown powder (264) dark grey rock, some large fragments but mostly small grains (265) fine grey powder (264) off-white powder (fine)(9) dark grey rock, mostly large fragments with some granular material (265) 20 CAMC-Abeyta-001899 F" w- Si & 6 f f s I I l' ch f if jl4 Ini 1 1I C! Sample ^ . 83 ' 84 86 87 88 ; 89 91 92 93 94 95 96 97 98 99 100 101 Table 1 (continued) Description dark grey rock, mostly large fragments with some granular and powdered material (265) dark grey rock almost all large fragments, very little granular or powdered material (266) dark grey granular material with some large fragments (265) white and pink rocks, very little granular or powdered material (9) coarse granular light brown several large fragments with some powder (93) grey-brown granular, inhomo geneous (93) fine off-white powder (263) white rock with powder (263) grey and brown rocks (93) fine white powder (263) off-white powder, some aggre gation (263) fine white powder (263) fine white powder (263) fine off-white powder (92) fine beige powder (153) fine white powder (263) white to light grey rocks with a fairly large amount of powder (92) 21 CAMC-Abeyta-001900 Sample 102 103 104 105 106 107 108 109 Table 1 (continued) Description large grey rocks with white powder @ the bottom (153) fine white powder (263) off-white powder (263) fine white powder (263) fine white powder (263) large grey and light brown rocks, some powder in the bottom, very inhomogeneous fine white powder (263) white powder, extensive aggregation (263) 22 CAMC-Abeyta-001901 Table 2 Results of the 31 Samples Identified as Containing Fibers by Both Analysts on the Initial Scan Sample No. 1 2* '3 4 5 6 7* 34* ^ 38* 41* .. 42 43 45 46 47 49* ' 50* Analyst 1 95% Confidence Intervals Count [F/(P+F)J P m LU (%) (%) 102/760 13.4 11.1 16.0 135/809 16.7 14.2 19.4 100/480 20.8 17.3 24.7 105/660 15.9 13.2 18.9 100/711 14.1 11.6 16.8 158/1606 9.8 8.4 11.4 102/852 12.0 9.9 14.3 101/696 14.5 12.0 17.3 45/652 6.9 5.1 9.1 100/991 10.1 8.3 12.1 105/593 17.7 14.7 21.0 103/885 11.6 9.6 13.9 106/682 15.5 12.9 18.5 104/1163 8.9 7.3 10.7 104/543 19.2 15.9 22.7 142/1263 11. 2 9.6 13.1 124/659 100/848 18.8 15.9 22.0 11.8 9.7 14.2 Analyst 2 95% Confidence Intervals Count P LU [F/(P+F)] (%) (*) (%) 125/833 15.0 12.7 17.6 142/1138 12.5 10.6 14.5 103/445 23.1 19.3 27.3 109/670 16.3 13.6 19.3 109/720 15.1 12.6 18.0 83/714 11.6 9.4 14.2 102/586 17.4 14.4 20.7 105/1028 10.2 8.4 12.2 17/451 3.8 2.2 6.0 70/501 14.0 11.1 17.3 106/492 21.5 18.0 25.4 102/829 12.3 10.2 14.7 86/627 13.7 11.1 16.6 103/1155 8.9 7.3 10.7 136/817 16.6 14.2 19.4 102/393 26.0 21.7 30.6 127/600 76/866 21.2 18.0 24.6 8.8 7.0 10.9 * Incompatible replicates. 23 CAMC-Abeyta-001902 Sample No. 52* 56 58* 60* 63 75* 76* 87 4- 92 94* 95* 96* 104* 109* Table 2 (continued) tL Analyst 1 95% Confidence Intervals Count [F/(P+F)j 100/919 r m 10.9 ~XT_j (%) " 8.9 U C%) 13.1 72/900 8.0 6.3 10.0 48/1863 2.6 1.9 3.4 100/1285 7.8 6.4 9.4 67/1901 3.5 2.7 4.4 87/2237 3.9 3.1 4.8 111/1474 7.5 6.2 9.0 105/1245 8.4 7.0 10.1 100/759 13.2 10.8 15.8 48/471 101/329 35/3276 10.2 30.7 1.1 7.6- 13.3 25.8 36.0 0.7 1.5 ' ` 43/1947 2.2 1.6 3.0 55/1571 5.5 2.6 4.5 69/959 7.2 5.6 9.0 111/2674 4.2 3.4 5.0 Analyst 2 Count [F/(P+F)] 54/767 54/581 29/503 60/1190 79/1685 95% Confidence Intervals P LU (%) (%) (%) 7.0 9.3 5.8 5.0 4.7 5.3 9.1 7.1 11.9 3.9 8.2 3.9 6.4 3.7 5.8 109/686 15.9 *13.2 18.8 100/522 19.2 15.9 22.8 55/527 10.4 8.0 13.4 111/314 35.4 30.1 40.9 105/906 11.6- 9.6 13 8 72/1674 4.3 3.4 5.4 37/2820 1.3 0.9 1.8 36/625 5.8 4.1 7.9 101/773 13.1 10.8 15.6 104/1890 5.5 4.5 6.6 B t II t ^' &i f' c; V- } ffv*. M Ik / r i. -.'I . i Incompatible replicates. f S M S'T M S 'S S '<01n . Confidence limits were ralntiitoj 24 CAMC-Abeyta-001903 Table 3 Probability of Finding No Fibers in Counts of Size N True Fiber Content, in Percent 0.01 0.05 0.10 0.15 0.20 0.25 0.30 0.40 ' 0.50 0.60 0.80 1.00 1.50 2.00 2.50 N=200 0.98 0.90 0.82 0.74 0.67 0.60 0.55 0.45 0.37 0.30 0.20 0.13 0.05 0.02 <0.01 Size of Count N=500 N=1000 0.95 0.90 0.78 0.61 0.61 0.37 0.47 0.22 0.37 0.14 0.29 0.08 0.22 0.05 0.13 0.02 0.08 0.01 0.05 <0.01 ' 0.02 0.01 <0.01 N=2000 0.82 0.37 0.14 0.05 0.02 0.01 <0.01 25 CAMC-Abeyta-001904 T a b le 4 R e s u lt s o f th e 28 S a m p le s I d e n t i f i e d b y B o th A n a ly s t s a s N ot C o n ta in in g F ib e r s on th e I n i t i a l Scan sample No. 36 37 1 54* 55 61 + 62 + 63 + 65 + 66 70 71 72 73 78 80 81 Analyst 1 95% Confidence Intervals Count p [F/ (P+F) ] (%) LU C%) (%) Analyst 2 951 Confidence Intervals Cotint [F/(P+F)] PL (%) .(%) U (%) not counted not counted 22/828 2.7 1.7 4.0 5/899 0.6 0.2 1.3 not counted not counted 4/1395 0.3 0.1 0.7 0/1410 0.0 0.0 0.2 no fibers observed on second scan no fibers observed on second scan 18/1545 1.2 0.7 1.8 10/1598 0.6 0.3 1.1 not counted not counted not counted not counted not counted not counted not counted Incompatible replicates, t Samples selected for second scan. 26 CAMC-Abeyta-001905 ,V. Table 4 (continued). kt Analyst 1___________ ___ Analyst 2 95% Confidence 95% Confidence Intervals Intervals ?y- ' Sample Count PLU No. [F/(P+F)] m m (%) Count P LU [F/(P+F)] (%) (%) (%) +82 no fibers observed on second scan 84 not counted 86 not counted a + 88 no fibers observed on second scan 89 not counted 91 not counted 93 +98 not counted 18/746 2.4 1.4 3.8 9/522 1.7 0.8 3.2 +101 2/200 1.0 0.1 3.6 6/230 2.6 1.0 5.6 * 102 not counted 1 107 not counted ! + 108* 56/833 6.7 5.1 8.6 18/1165 1.5 0.9 2.4 * Incompatible replicates. + Samples selected for second scan. 27 ! CAMC-Abeyta-001906 t . Table 5 Results on the 21 Samples Identified as Containing Fibers by Only One Analyst on the Initial Scan Sample No. 35 .39 40 44* 48 51* 53* 57 59 64 67 69 74 79 83* 9? * 99* Analyst 1___________ 95% Confidence Intervals Count P LU [F/(P+F)] (%) (%) (%) __________ Analyst 2___________ 95% Confidence Intervals <s<>G Count P LU [F/(P+F)] (%) (*) (%) not counted 29/813 3.6 not counted 100/905 11.0 not counted 45/833 5.4 50/1091 4.6 not couhted not counted not couhted not counted not counted 1 not counted not counted 18/430 4.2 36/1861 1.9 30/1212 2.5 2.4 5.1 9.1 13.3 4.0 7.2 3.4 6.0 2.5 6.5 1.4 2.7 1.7 3.5 . 9/550 1.6 0.7 3.1 81/418 19.4 15.7 23.5 39/1354 2.9 2.0 3.9 42/1771 2.4 1.7 3.2 3/630 0.5 0.1 1.4 18/1693 1.1 0.6 1.7 1/739 0.1 0.0 0.8 * Incompatible replicates. 28 CAMC-Abeyta-001907 Table 5 (continued) Sample No. __________ Analyst 1___________ 95% Confidence Intervals Count p L U [F/CP+F)] (%) (%) (%) 100 103 105* 106 not counted not counted 86/1498 5.7 4.6 7.0 not counted __________ Analyst 2 __ 95% Confidence ' Intervals Count [F/ (P+F) ] pL (%) (%) U (%) 27/777 3.5 2.3 5.0 * Incompatible replicates. 29 CAMC-Abeyta-001908 o o n1 2 3 4 5 OO O O 6 7 8 9 D=v5^ Figure 1. Porton Reticle f 30 t l CAMC-Abeyta-001909 % \ MAXIMUM LENGTH ** N O T SPECIFIED u U. iJiiV* . i j A a i PA R TICLE WIDTH, nm L Figure 2. Fiber dimensions that would be observed as per OSHA definitions and resolution limit of microscope. 31 CAMC-Abeyta-001910 S*** Sample 92 Scanning Electron Microscope 1 r Figure 3a CAMC-Abeyta-001911 .>? b F; Sample 81 8 z <1 i Light Microscope tl X* Scanning Electron Microscope V (4' )' s 5 t y L; (- 7 3, ? $>* V ? 1 tf I ! i I Figure 3b CAMC-Abeyta-001912 I-r-*?>* ;* < :k iS. ;-i i' i ( t*;0) Light Microscope Sample 108 Scanning Electron Microscope i 'TO** *t1l (4 v t f Figur Qe CAMC-Abeyta-001913