Document JN9npGwVD00JKV7Jk1MDExyVB

IMPROVED SAMPLING, -MEASURING, IDENTIFICATION AND ANALYZING TECHNIQUES OF AIRBORN ASBESTOS PARTICLES L. Bartosiewicz Scientific Research Staff, Ford Motor Company Dearborn, Michigan 48121 June 21, 1972 HOT1 1 iSI IJPSlgT Jafci 4 m 1 |i>wwyy^ ' 8005 1405 PRODUCED BY FORD SCF-FORD-0832 ft # V - Mil ft A / ABSTRACT* Improved Sampling, Measuring, Identification and Analyzing Techniques of Airborn Asbestos Particles by . L. Bartoslevicz Collection, identification and data analyzing techniques are presented for the airborn mineral fibers, The limitations of the techniques applications are discussed. It is shcrvm that the adaption of the 1-3/1*" long plenum for sample collection is important since it leads to a much more uniform particle distribution. Separation of the asbestos particles from other solid airborn pollutants vith the refractive index liquid method is accurate and convenient and can have an important effect on the results as shovn by sample D. For analysis of this type the IAM is a very valuable tool. 8005 1406 PRODUCED BY FORD i > UmODUCTIDN 4 ' . J' ! The need to evaluate the hazards from asbestos Inhalation vas I recognized as early as 106 but it vas more than twenty years before a ( * - detailed description of asbestos is vas developed* Since then, it has 'TV* i k- been documented by numerous studies that inhalation of high concentrations of asbestos or any prolonged exposure to moderate asbestos dust level can in many Instances lead to the development of asbestos connected 1 i diseases of the lung. ii . y In general, the available data and information are only pertinent to the high concentrations caused by occupational exposure, and not to the low asbestos fiber content found as background in air. The health ii problem as posed by the biologically active asbestos fibers has led to p the establishment of occupational safety standards. But, little has been done to develop reliable sample collection, data generation, and identification techniques. This paper deals vitb these three aspects of analyzing for airborne asbestos particles. Sampling Sample collection is the most important step in any analysis for air pollution. Therefore one must be sure that the collection technique used gives an unbiased representation of those particles which are present. For this study samples were collected by the Ford Motor Industrial Hygiene Unit. Two sample collection methods were used, the approved NXOSH (National Institute for Occupational Safety and Health) and a modified NIOSH technique. All samples were simultaneously obtained at a manufacturing facility of the ` Ford Motor Company. . * 8005 1407 PRODUCED BY FORD Approved 17I0SH Technique ' ' ' The first group of samples were taken In the mixing room of the plant using the technique approved by H10SR which is to use a short open luce filter holder, with a 37tam, type AA, Millipore filter supported on a hase pad. Two samples were collected with this device using a 2 liter/ minute air sampling flow rate for 48 minutes per sample (a total of 96 liters of air was sampled) Sample A was collected in the casting Yoom and sample B in the mixing room. Micrographs showing the particle distri bution of samples A and B are shown in Figures 1 and 2. Various areas of samples A and B were analyzed and the data is presented in Table 1. Samples A and B 6howed that dividing the entire filters into three equal parts and analyzing each segment separately yielded three different results on each filter. The fiber count for fibers equal to or greater than 5 p tos 5*2 fibers/mX of air sampled in one area of sample A, whale the fiber count was 2.5 and 4.1 fibers/ml in the other randomly selected areas. * The values obtained for particles equal to or greater than 5 for sample B were 1.1 fibers/mH in the first segment, 35 fibers/m in the second, and 6.8 fibers/ml in the third section. It can be seen from Figures 1 and 2 and the data in Table 1 that the particles are not uniformily distributed on the Millipore filters and therefore these samples cannot be assumed to be representative of the sampled working area. For this reason it is almost impossible to avoid a biased area selection for the particle counting. ' 8005 1406 PRODUCED BY FORD TABIE I FUrticle Size Analysis of Samples A and B Sample Where Obtained Plenum Length . Technique Used Total Air Volume Sampled Analyzed Filt. Segments Fiber Count/mi. 5 pm or > Segment number 1 Fiber Count/mi 5 pm or > Segment number 2 Fiber Count/mi 5 (jjo or > Segment number 3 A Casting Room Short Open Face NIOSH Standard 96 liters . 3 5.2 2.5 4.1 B Mixing Room Short Open Face NIOSH Standard 96 liters 3 1.1 3.5 6.8 From these results, it is concluded that the samples collected vith an open face filter holder do not present a reliable or reproducible base for the analysis. Other aspects of reliability of thi6 method vill be discussed later. Modified N10SH Technique The collection method for the second-group of samples was modified so that the filter holders were fitted with a one and three-quarter inch high plenum over the base of the filters. The air inlet on the top of the plenum was the standard size 0.150" opening and all other parameters were kept the same.'. ' .i * -\ -4- 8005 H09 w' I'-V-M PRODUCED BY FORD ' ' Preliminary calculations predicted that the 1-3/**" long plenum vould ' ` create a condition that vould satisfy the desired air flow properties of this small system. Not only does the plenum lengthen the particle path hut it creates a low grade air turbulanee which helps to uniformly distribute particles over the entire filter surface. Experience has shovn that a sampling rate of 2 liters per minute for either 1*8 minutes a6 was the case for sample C, or the 38 minutes collection time used for sample D vill produce an adequate, moderately dense and very uniform particle distribution, aB shovn in Figures 3 and 1* and in Sable II. An analysis corresponding to that used for sample A and B vas made on these specimens. It vas found that the analysis of these samples vith the evaluation method as outlined by the NIOSH, yielded results vith much less scatter between the filter segnents. An evaluation of sample C taken in the casting room, revealed that there vero 5.6 fibers/mA present for fibers 5 pm or greater in length in the first segment of the filter while 1.9 fibers/rai were .found in the second and 5*8 fibers/raA were counted in the third segment. Sample D, taken in the mixing room, shows an equally good distribution between the throe segments of the filter. In the first filter segment 10.5 fibers/m were counted for fibers 5 pin or longer in length, while 11.2 fibers/m were found in the second, and 9.8 fibers/ml were found in the third segment. 11 r 4 56005 U10 PRODUCED BY FORD ' TABUS tl Particle Size Analysis of Samples C and 1) Sample Where Obtained Plenum Length Technique Used Total Air Volume Sampled Analyzed Filter Segments Fiber Count/mX 5 pm or > Segment number 1 Fiber Count/mi 5 pm or > Segment number 2 Fiber Count/mi 5 pjn or > Segment number 3 C` Casting Room 1-3A" long Modified 96 liters 3 5.6 ' I4.9 5.8 D Mixing Room 1-3/V' long Modified 76 liters 3 10.5 11.2 9.8 Comparing the results of the analysis of sample C with sample A, the superiority of the long plenum collection method is apparent. The difference between the high and the low fiber count of sample C is only 9 fiber, while sample A shows a three times higher discrepancy or 2.7 fibers. The relationship between sample D and B is even worse. Sample D shows only a 1.1* fiber count difference between the individual filter segments, while a 5*7 fiber count difference exists between the three filter segments of sample D. The fiber count discrepancy of sample B is slightly more than four times that of sample D. The previously described length of the plenum (1-3/V) and the standard size air inlet were found acceptable for all other flow rates a6 specified by the NIOSH. Since the recomnended flow rate of air is .' '' ,, 8005 1411 PRODUCED BY FORD  *. V* s *V 'v* *. - 4< u. ^ " .C( '' ' /*' v- . .* .. . ' _ . ,, ' " . . . . -, . .* ` between l'and 2.5 liters per minute and the calculations for the '' - . .?' lengthened plenum were made for 2 liters per minute of air flow, no ' * ' '' .' ".`difference is expected at either end of the flow rate*range. It ' " * * . > : 6hould he pointed out that the uniform distribution of heavier particles t " . JM . ` as'seen here (l.e. > 5 pm) insures the uniform distribution of smaller H ** . . particles (l.e. < 5 jim). ` '* Sample Preparation for Asbestos Identification Using a Refractive Index Liquid ' Since the index of refraction (ND) of most asbestos fibers is near to the refractive index of the chemically cleared membrane filter, it becomes necessary to eliminate the presence of the filter. This can best be accomplished using a low temperature asher. This method uses a very slow oxidising process, that will neither harm nor disturb even the finest particles of asbestos or other non-oxidizable material. This instrument allows the oxidation reaction to occur at a low temperature by converting approximately 20$ of the molecular oxygen to atomic oxygen3 vith the use of a radio frequency field. The atomic oxygen reacts with the oxidisable materials in the specimen chamber. A mechanical vacuum pump then removes the volatile oxidation products. This method also eliminates many other oxidizable airborne particles commonly found in industrial operations. Approximately four hours are required to "ash" the filter. Samples C and D were prepared by this technique. The pressure of tic specimen chamber was maintained at .6mm pressure of mercury and 30 vatt6 of power was used. The flow rate of the oxygen was 150 cc/minute. V - * * , f 7 8005 1412 PRODUCED BY FORD - - *!U ' k, ; ~0: ' " ' , .* * v.*T` .* * . * 4. '' *, . " *' f - . ? .- *.<.... ... *' * * . T* , * . * * . *' * Identification and Measurements of Asbestos Fibers Vltb an Automated K`* * < ' Image Analyzing Microscope. k* N. * . ' An extremely convenient and easy method of measuring particles 1b with an Image Analyzing Microscope (JAM). With the aid of an IAM it is not difficult to measure 50 to 100 fields in a relatively short * time. The particles can be bracketed in various size groups, and the measuring of a previously mapped out area can be accomplished. Data collection from 5 fields requires approximately 15 minutes. < . In order to determine the number and size of asbestos particles present on a filter, using the IAM, a multi-process procedure is used. The specimen is first ashed as outlined above to eliminate the filter and any oxidizable particles such as lint. The area (or areas) of interest are mapped out, measured, and tie particles present counted. This vill give the total number of particleB (asbestos and non-asbestos) present. A second measurement of the exact same area is made after a refractive index liquid having the same refractive index as ehrysotile4 is placed on the sample. This vill have the effect of making all the asbestos particles invisible and allow us to determine the number of non-asbestos particles present in the sample. By this means, an accurate measure of only the asbestos particles present can be made. Borne care must be exercised when using the refractive index liquid, 6ince the threshold sensitivity of the instrument's detection system for proper functioning requires approximately 10$ difference in contrast. Minute shade variations due to the close refractive index similarities cannot be detected. Consequently the absolute matching of the various . refractive indexes among the different asbestos types is not necessary. ' 4>W r4 8' - - ' B005 1413 PRODUCED BY FORD o. I , 'X.*- X ( For this experiment a refractive index liquid (ND 1.5^6) vas used that very veil coincideB vitb the index of refraction of chrysotile (the most W videly used fons of asbestos for industrial application) and also acceptable for serpentine since the contrast produced is still a shade below the \ detection limit of the Instrument. On the average a refractive index different of ND .010 is acceptable to make proper measurements. It is also recognized that substances other than asbestos with refractive indexes within this range will be measured as asbestos. This source I of error cannot be totally eliminated, but knowing the total material involvement of the locality to be tested, and the type of asbestos used, the proper choice of the refractive index liquid will minimize this problem. Correlation of the data is very simple. Since the measurements accumulated in the first part of the analysis encompass all particles present, the calculated values do not reflect accurately the presence of the asbestos phase alone. Separating the asbestos particles from all other detectable features by suppressing their visibility with the refractive index liquid in the second portion of the analysis enables the- data accumulation of the non-asbestos content of the sample. Sample C and D were analyzed with the IAM. Calculations for the asbestos fiber content of the sample were carried out using the formulae published-in the appendix of the Criteria Document.2 i. I -98005 1414 PRODUCED BY FORD Filter area ' Field area * (1) Average net count x K Air volume sampled fibers/m (2) Measurements of sample C indicated that only ,2f> of the particles vere non-asbestos, therefore the sample vas viewed as being lOOf, asbestos. The particles vere measured in six different size groups that ranged from 2 pjn to 15 pm. Only 163 particles or 12.75S vere counted as being 5 pm or larger in the total particle population of 1280. Most particles were found in tic 2 pm to 3 pm range and the calculated average particle size was 2.9 pm. The total fiber count of all sizes vas 5.6 fibers per milliliter, and 12.% of this or .6 fiber/ml vere 5 pm or larger. The analysis of sample D vas carried out identically to sample C. Micrographs of sample 1) are shown in Figures 5 and 6 before and after application of the refractive index liquid. It vas found that sample I) contained 31.9& of non-asbestos particles. There vas a total of 53*8 particles counted per milliliter, but the 31*9$ or 16.8 particles of non asbestos origin reduced this count to 37 fibers/mH. The size distribution measurements indicated that 3*3^ of the particles were 5 pm or larger. Consequently, for these sizes the fiber count per milliliter was 1.1, The largest number of particles vere again found to be in the 2 pm to 3 pm group, and the average particle size measured 2.3 pm. These data are presented in Table III. ' -10 8005 1415 *1 PRODUCED BY FORD TABI2 HI Data From IAM Analysis Sample $ of Non-asbestos of Xferticles 5 pm or > Fiber count/mi. 5 pm or > Average Particle Size C 0 12.7 .6 2.9 pm t D 31.9 3.3 i.i 2.3 pjn DISCUSSION The present use of the "total fiber count" system as recommended by the NI0S1I for overall evaluation of asbestos concentrate ons coupled vith the "open f&ce" filter collection technique can be a source of error in the measurement of exposure to asbestos. The low number of total particle count and an approximately 1/9000 analyzed area of the total filter surface (as specified by NIOSH) along vith the non-uniform particle distribution cannot produce a reliable analysis. The presence of asbestos resembling fibers also adds to the problems. The accuracy of the IAM method relies on two important facts, the large number of particles counted, and the large number of areas covered. It is also important that within these areas all particles are counted. These factors will minimize the uncontrollable errors. There 1b no correlation between the results obtained by IAM and the microscope counted samples C and D (Tables II and III). This is due to the fact that sample D is composed to almost one-third of non-asbeBtos particles. 1* / 11- 8005 1416 PRODUCED BY FORD ACKW0WIED5MENTS The author is grateful to Dr. E. Eicben for the critical review of the manuscript, and to Dr. C. D. Melvin for the fruitful discussions. 1 also wish to thank Mr. J. V. Sprys and Mr. R. C. McCune for their help with the sample preparation using the low temperature asher, and Mr. E. B. Lick of the Industrial Hygiene Unitfbr collecting the samples. -12- 8005 1417 PRODUCED BY FORD -' .. * J* REFERENCES . , i. . f. : ^1* < G* ilson, J. C., "Man and Asbestos" Ann. N. Y. Acad. Sci. 132:9 . " h -j 4( * * (December 1965) t, * .s 2. Criteria Document: Recommendations for an Occupational Exposure Standard for Asbestos, U. 5. Department of Health, Education, and Welfare. Rockville, Maryland. . I 3* Hollaban, F. R.: Topics in Chemical Instrumentation. XXVII. Analytical Applications of Electrodelessly Discharged Gases, J. Chem. Ed. i+3> A401 (1966). 1*. McCrone, W. C., R. G. Drafty, and G. J. Delly: The Particle Atlas, Ann Arbor Science Publishers Inc., Ann Arbor, Michigan, p. 15I-165 (1967). 8005 1418 PRODUCED BY FORD 1* 4 *\ Figure 1. Non-uniform particle distribution of asbestos fibers and dust, collected vith an open face filter holder and viewed at 500X, / phase contrast. (Sample A) Figure 2. Phase contrast micrograph at 500X, showing the unevenly distributed asbestos dust particles as collected with an open ihce filter.holder. (Sample B) Figure 3 Uniform distribution of asbestos fibers and dust particles I as collected vith the modified filter holder (l-3/V' long y plenum) and viewed at 500K, phase contrast. (Sample C) Figure 1*. Even distribution of asbestos and non-asbestos particles as collected vith the modified filter holder (l-3/V long plenum) and viewed at phase contrast. (Sample D) Figure 5. Asbestos and non-asbestos particles before the application of refractive index liquid, 5OCX. (Sample P) Figure 6. Exact same area as shown in Figure 5 after application of the refractive index liquid, 5OCX* (Sample D) i: fr-r . ` . 8005 1419 PRODUCED BY FORD I ( \ *'/ V ' - ... f- ' * a.`* I V^i' ;' vs <_. }- . . . **/-S'...4 * >*. *- v. . *o * .* i ** r.. . V* '.1 Figure 1. Non-uniform particle distribution of asbestos fibers and dust, collected vith an open face filter holder and vieved at 500X, phase contrast. (Sample A) .1. -r # 1* * * > . i; 9 ' /* V .> H 'l H !/'. t `V- ' A.; V.-iv" rv^r.,y -v/ ; f'j*j*1?* > ./ vv\1 y\ **-.-*. 1*/?`vy.^ * * VS .'v/tV.,*'^;*, v, 5 '* ', ' . ir ! ` ,.yt> t .*m, ,r' *?5* * .* ;.'Fk **, JiJ Figure 2. Phase contrast micrograph at 500Xj shoving the unevenly distributed asbestos dust particles as collected vith an open face filter holder. (Sample B) WNimi BOOS 1420 PRODUCED BY FORD  Figure 3. Uniform distribution of asbestos fibers and dust particles as collected vith the modified filter holder (1-3/**" long plenum) and viewed at JOCK, phase contrast. (Sample C) ' .< i < % 1 p* o > 9 ./ Figure I4 Even distribution of asbestos and non-asbestos particleB as collected vith the modified filter holder (1-3A" long plenum) and viewed at 50QX, phase contrast. (Sample D) 8005 1421 PRODUCED BY FORD Figure 5. Asbestos and non-a6bestos particles before the application of refractive index liquid, 50QX. (Sample D) II i -- r~ * m r in iiia iiA ii mi Figure 6. Exact same area as shovn in Figure 5 after application of the refractive index liquid, 50GK. (Sample D) 8005 1422 A PRODUCED BY FORD