Document RjDx346pQX6zR7XkqwVqbEnDB

REMOVAL AND DETECTION OF LIQUID BORNE ASBESTOS FIBERS WITH NUCLEPORE MEMBRANES By Dr. M.C. Porter Vice President of Research and Development and R.G. Berggren Research Chemist Nuclepore Corporation Pleasanton, California 94566 December 24, 1973 TABLE OF CONTENTS Abstract I. Introduction A. Asbestos Structure and Usage B. Use of Asbestos as a Liquid Filtration Media C. Asbestos Health Hazard D. Removal and Detection of Asbestos Fibers with Membrane Filters II. Methodology for Detection and Counting of Asbestos Fibers A. Microscopic Analysis of Membrane Collected Materials B. Transmission Electron Microscopy C. Scanning Electron Microscopy III. Retention of Asbestos Fibers on Nuclepore Membranes A. Experimental Procedure 1. Preparation of Asbestos Test Suspensions 2. Background Counts 3. Challenge Tests 4. Fiber Counting Procedures B. Experimental Results 1. Retention as a Function of Pore Size 2. Retention as a Function of Challenge Level IV. Conclusion References PAGE i 1 1 4 6 7 10 10 13 15 18 18 18 19 21 22 22 22 30 39 41 ABSTRACT The recent publication in the Federal Register with regard to "Asbestos Particles in Food and Drugs" prompted a study of the unique characteristics of Nuclepore membranes for use in detection and removal of liquid-borne asbestos fibers. It was found that the Nuclepore membrane provides an ideal substrate for collection and viewing of asbestos fibers in the transmission or scanning electron microscope. Retention studies indicated that the 0.4pm pore size offers virtually complete retention of all chrysotile asbestos fibers while the 0.2pm and 0 .1pm pore sizes provide absolute retention. Increasing challenge levels resulted in decreased fiber passage presumably due to fiber matting. The effect was most pronounced at challenge levels above 108 fibers/cm2. REMOVAL AND DETECTION OF LIQUID BORNE ASBESTOS FIBERS WITH NUCLEPORE MEMBRANES I. Introduction A. Asbestos Structure and Usage Asbestos is a unique material which has found a variety of applications ranging from asbestos cement products to filtration media. It has been called "the magic mineral", "the mineral of unparalleled properties", "the mineral of a thousand uses".1 The mineral can be divided into six types (see Table I)z. All types occur in long crystals which are so fine as to be fibrous in structure. They vary in chemical composi tion containing magnesium, ferrous or ferric iron, calcium, or sodium cations in various proportions with the silicate anion. All are incombustible and can be separated by mechanical or chemical means into fibers of various lengths and diameters. The six generally recognized varieties include the fibrous form of serpentine, known as chrysotile, and five minerals of the amphibole group. Chrysotile, or white asbestos, is by far the most common variety,accounting for over 90% of the world's usage. Its positive surface charge, softness and flexibility, low refractive index and poor acid resistance, set it apart from the other varieties of asbestos. With individual fiber diameters down to 0.30 microns (300 Angstroms) it represents the smallest commonly-known fiber. The usable surface area,as measured by the nitrogen adsorption method,is from two to five times greater than any other known asbestos. The high surface area and positive surface charge make it ideal for a depth filtration media with exceptional dirt loading capacity. Chrysotile's crystal structure is similar to that of the kaolinite minerals consisting of alternate layers of tetra hedrally coordinated silica and octahedrally coordinated magnesium hydroxide joined in a composite sheet. The outer surface of the fibril is essentially magnesium hydroxide in the brucite form. Figure l3 is a representation of the structure showing the hydroxyl outer surface of the fibril. The resultant cationic charge on the surface enhances its affinity for electronegative particulates (in filtration) and for many resin systems (reinforced plastics). Pauling* , long before the layer structure of chrysotile was recognized, predicted that a magnesium analog of the kaolinite minerals would have to exist as curved sheets because of a mismatch in the repeat distances of the individual layers (8.9 A f o r the silica layers versus 9.4 A for brucite). Pauling's prediction has been verified by x-ray diffraction studies and by electron microscopy showing closed cylinders of 1 Table I Comparative properties of asbestos minerals ( 2 ) P ro p e rty C hryso tlle A n th o p h ylllte C ro cid o lite A ctin o lite T rem o lite A m o site C om po sition C rystal habit M g 3S i20 5( 0 H ) M g , S i B0 22( O H ) 2 N aaF e^S isO js(O H )2 C a 2F e 5 S i8 0 2 2 (O H )2 M g 5C a 2 S i80 22(OH)2 (F e , M g ) , S i * 0 22( 0 H ) 3 Fine, easily P rism atic Long brittle fibers P rism atic crystals P rism atic crystals P rism atic crystals sep arab le crystals and and fibers and fibers and fibers fib ers fib e rs V eln ing C ro ss-, slip-, S lip - & m assive- C ro ss-fib er S lip - & m assive- S lip - & m assive- C ro ss-fib er & m assive- fib er fib er fib er fib er C o lo r W hite, gray, G ray-w h ite B lue G reen W hite, g ray-w h ite, A sh g ray or brow n green green ^ ^ is te r S ilky V itre o u s to S ilky to dull S ilk y S ilky V itre o u s to p e a rly p e a rly Textu re S o ft to harsh, H a rs h S oft to harsh Harsh H a rs h H a rs h also silky F lexib ility Very good Poor Good Poor Poor Good H ardness 2 .5 -4 5 .5 -6 4 6 5.5 5.5 -6 Fiber d iam eter, A T e n s ile stren g th , p s.i. 2 0 0 -4 0 0 30 0,000-800,000 6 0 0 -9 0 0 <4000 6 0 0 -9 0 0 40 0,000-600,000 6 0 0 -9 0 0 <1000 6 0 0 -9 0 0 1 0 0 0 -8 0 0 0 6 0 0 -9 0 0 1 0 0 ,0 0 0 -2 0 0 ,0 0 0 M o d u lu s o f e la s tic ity , p.s.i. 23x10' -- 27x10' -- __ 24x10' S p e c ific h e at, B t u /lb ./ F. 0 .2 6 6 0 .2 1 0 0 .2 0 1 0 .2 1 7 0 .2 1 2 0 .1 9 3 S p ec ific gravity 2 .4 -2 .6 2 .8 -3 .1 3.2 -3 3 3 .0 -3 .2 2 .9 -3 .2 3 .1 -3 .3 R e frac tive In dex (m ean ) 1 .5 1 -1 .5 5 1.61 1.7 1.63 1.61 1 .6 4 pH 9 .5 -1 0 .3 9 .4 9.1 9 .5 9 .5 9.1 E le c tric c h a rg e In aq u eo u s slurry R esistan ce to acids P ositive Poor N eg ative V ery.go od N eg ative Good N eg ative F a ir N eg ative Good N eg ative Good 2 0.95A HYDROXYL OXYGEN MAGNESIUM a SILICON 0.06A FIGURE I: Chrysotile structure (after Whittaker). 3 extreme length-to-diameter ratios up to 1000:1 . (See Figure 2). Recent high resolution electron micro graphs taken in Japan, have shown the average outside diameter of the individual chrysotile "fibril" to be about .03omicron (250A) and its inside diameter to be about 140A.2 In addition to use as a filtration media and as a reinforcement for plastics (including cold-molded, thermoplastic, and thermosetting plastics) chrysotile is used in the manufacture of safety clothing, curtains, lagging cloth, woven brake linings, clutch facings, asbestos cement, millboard, roofing felts, pipe covering, fine quality electrical papers, insulating papers, asbestos-latex flooring felt, vinyl floor tile, paint, roof coating, caulks, sealants, sprayed insulation, asphalt paving and curbing, and many other products. Crocidolite, or blue asbestos, accounts for 3.5% of the total asbestos useage with principle use in the production of asbestos cement sheets and pressure pipes. Amosite represents 2.5% of the total asbestos usage primarily in the insulation field and in asbestos cement. Anthophyllite, because of its low strength (see Table I) finds limited usage in products in which reinforcement is required such as asbestos cement, floor tile, etc. It is,however, used as a filler in the rubber and plastic industries and in various adhesives and cements. Tremolite and actinolite are the least important of the asbestos minerals from an economic standpoint due to insuffi cient fiber strength. They find limited utilization as filter media and as inexpensive fillers. B. Use of Asbestos as a Liquid Filtration Media As mentioned above, chrysotile fibers represent one of the most effective fibrous depth filtration mediums commercially available. It is, of course, recognized that fibers with high length to diameter ratios (L/D) provide a large effective collision radius and a large area for adsorption of small particles thereby enhancing the "dirt loading" capability of the filtra tion media. Chrysotile has an exceptionally high L/D ratio and adsorption area (50-60 m 2/gm-BET nitrogen value)6 . However, glass fibers with similar geometry and kieselguhr with high BET adsorption areas are not as effective as chrysotile fibers.7 4 FIGURE 2 I d e a l l iz e d Packing Con figu ration Ch r y s o t il e F i b r i l s 5 The explanation appears to be in chrysotile's positive zeta-potential. Because of chrysotile's magnesium hydrox ide surface layer, in aqueous solution, the fibril adsorbs positive ions-notably H+. Even in the range of pHlO-12, the surface displays a positive charge.5 Because many particulates are electronegative at neutral pH (organics particularly), chrysotile is an effective "getter". (It is interesting to note that adsorption can also occur with cationic species like meth lene lue dye6under favorable conditions.) The importance of the OH groups for effective filtration is demonstrated by the report that a severe acid treatment,leaving only the silica residue, resulted in a loss of filtration efficiency.7 As a result,chrysotile asbestos pads have found increasing use in the food and pharmaceutical industries. Filtration of beverages (such as beer and wine), liquid sugar, and parenteral solutions are notable. The fibers have been felted into sheets with cellulose and glass to provide pre formed, pre-cut filter media for plate and frame or tank type filters. Although such pads do not provide absolute filtration, they are nevertheless effective in removing yeast along with other microorganisms and particulates. Unfortunately, "sluff-off" from these pads adds some contamination to the filtrate. Asbestos Health Hazard Recently, the Food and Drug Administration was petitioned to publish in the Federal Register8 , a proposed regulation: "Food or drugs that have come into contact with filters made wholly or partially of asbestos may reasonably be expected to become contaminated with asbestos particles which may be injurious to health when ingested. Accordingly, any food or food additive,drug or drug component produced, manufactured, processed or prepared using a fiber made wholly or partially of asbestos shall be deemed to be adulterated in violation of section 402(a) of the Act." The Federal Register pointed out that asbestos inhalation has been recognized for sometime to be an occupational hazard for workers in asbestos mines or factories. The asbestos is inhaled and lodges in lungs causing the develop ment of a fibrotic disease known as "asbestosis". It is reported that there is a 5 to 7 fold increase in lung cancer in asbestos workers and that 7% of deaths of asbestos workers are caused by pleural and peritoneal mesothelioma. 6 More recently, it has been discovered that pulmonary fibrosis and a variety of neoplasms can occur with much less exposure,such as that experienced by those living in the household of an asbestos worker or merely living within 800 yards of an asbestos plant.9 On the other hand, the evidence for possible hazard from ingestion of asbestos particles is contradictory. There is some evidence of increased incidence of gastrointestinal cancer among asbestos workers.10 A number of workers have shown in animal tests that asbestos fibers can penetrate through the gut into the blood stream and accumulate in various tissues.11 However, there is also evidence that the bloodstream can clear itself of these fibers 11 It has been demonstrated that parenteral administration of asbestos may lead to wide dissemination of the fibers causing local malignant tumors, mesotheliomas of the pleura and peritoneum, similar to those occurring after inhalation of the fibers.9'12 Evidence such as the above, prompted the National Institute of Environmental Health Sciences and the Environmental Protection Agency to sponsor an international conference on the biological effects of ingested asbestos in Durham, North Carolina in November 1973. Attended by 130 scientists and public health officials from 12 countries, the conference emphasized the limited amount of information available on the effects of low-level, prolonged exposure. The draw backs of making inferences from animal models were repeat edly pointed out by speakers at the conference. It should also be realized that a health hazard may exist with other fibers besides asbestos. Tests conducted by Dr. Mearl F. Stanton at the U.S. National Institutes of Health (National Cancer Institute) with conventional asbestos fibers, fine glass and sapphire fibers, compared these fibers with coarser fibers and powders of the same materials.13 Fine fibers (1.5 to 5 microns in diameter and less than 80 microns long) appeared to cause a high incidence of pleural cancers much more readily than the coarser materials. These tests suggest that fiber geometry is more important than chemical composition. D. Removal and Detection of Asbestos Fibers with Membrane Filters Since asbestos is found in many municipal drinking water supplies ( 10 8 fibers/liter in some cases) and is used in virtually all pipes carrying drinking water, it may be impossible to eliminate asbestos by simply prohibiting its use as a filtration media in food and drug plants. Thus, the FDA commissioner has suggested that a "non-asbestoscontaining" or "non-fiber-releasing" filter such as a membrane 7 filter be used as a final filter prior to bottling to reduce the asbestos concentration. The effectiveness of membrane filters in removing asbestos fibers has not been thoroughly investigated. Recently, Manalan and Reti 1^presented a paper on the effectiveness of a Millipore cellulose ester membrane filter in retaining asbestos fibers dispersed in water. A summary of their results is presented in Table ii. Unfortunately, Manalan and Reti were unable to reduce the background count down stream of the membrane to zero. The background count was about 1 0 3 fibers/liter. This necessitated the assumption that fiber passage occurred if, and only if, a concomitant increase in the downstream fiber count resulted when the number of fibers added upstream of the membrane increased. Naturally, passage cannot be disproven on this basis. Nevertheless, two conclusions were drawn from this work: 1) Passage of some fibers can occur in cellulose ester membrane filters having an equivalent pore diameter greater than 0.45ym, 2) Retention on 1.2ym pore size cellulose ester membrane filters is increased by small amounts of beer or wine (from 95.8% to 99.8%,and 99.7%, respectively). Any investigation of membrane retention or passage of fibers is plagued by the sensitivity of the technique used to detect and count fibers downstream of the membrane. With micro organisms, detection and counting is simplified since the microorganism can be cultured and developed into a large colony visible in a low-power optical microscope or even with the naked eye. With non-living, inorganic fibers of the dimensions specified (down to 0.03 microns in diameter with various lengths), one must resort to the electron micro scope. Even with magnifications of 10,000 to 30,000X, the fibers are often hard to detect,depending on the texture of the support. This paper will describe a detection technique using Nuclepore membranes which we believe greatly facilitates detection and counting of asbestos fibers. The technique has been used in determining the retentivity of various pore size Nuclepore membranes (from 0.1 to 1.0 ym) for chrysotile asbestos fiberswhich data are presented at the end of the paper. 8 Pore Size 1 .2ym 1 .2ym 0 .8ym 0 .8ym 0.65ym 0.65ym 0 .45ym 0 .22ym TABLE II Retention of Asbestos Fibers With Millipore Cellulose Ester Membranes11* Filter Type Fiber Challenge Minimum % retained 'RAWP 1 ,000,000 98.28 RAWP 100,000,000 96.56 AAWP 1,000,000 98.62 AAWP 100,000,000 99.93 DAWP 1 ,000,000 99.93 DAWP 10,000,000 99.11 HAWP 100,000,000 99.993 GSWP 100,000,000 99.997 o".O Passed 1.72 3.44 1.38 .07 .07 .89 .007 .003 9 II. Methodology for Detection and Counting of Asbestos Fibers A. Microscopic Analysis of Membrane Collected Materials The key to reliable detection and counting of asbestos fibers is the selection of an appropriate substrate for collection and viewing of the fibers with an electron microscope- In the last five years, a new membrane filter has evolved which makes possible ideal collection, prepara tion, and viewing of particulates--whether from aerosols or liquid media. Membrane filters have pore sizes ranging between 300& to 12 microns and may be broadly categorized into two major types according to pore configuration. Figure 3 is a scanning electron photomicrograph (SEM) of the surface of a cellulose ester membrane filter typical of the first type-- having a homogenous, sponge-like pore network. Figure 4 is an SEM of a Nuclepore membrane filter which is unique because of its sieve-like straight through cylindrical pores. Although the pores in the cellulose ester membrane look considerably larger than those in the Nuclepore filter, their retention is equivalent becaue of pore constrictions and tortuosity in the depth of these membranes. The Nuclepore membrane has a lower porosity (10% compared to 80% for the cellulosic membrane) but is only one-twelfth as thick, resulting in higher or equivalent flow rates due to less resistance to flow. This is important since air and liquid samples are passed through the membrane to obtain relatively uniform distributions of particulates over the surface. However, the high porosity (and consequent high surface area) of the sponge-like membranes are associated with a higher dirt loading capacity permitting collection of larger quantities of particulates before plugging. Although both membranes are widely used for particulate and aerosol analysis, the Nuclepore membrane has been recognized15-18 to possess some unique advantages for particulate sample collection, preparation, and analysis. 9 We have endeavored to exploit some of these advantages in the collection and detection of liquid-borne asbestos fibers. 1. Sample Collection The uniform pore size distribution across the filter face allows for uniform particle deposition.17 10 Cellulose membrane filter, 0.45 micron pore size FIGURE 3 li Scanning electron micrograph FIGURE H 2 . Sample Preparation a. Surface replication for transmission electron microscopy (TEM) is easy. Silicon monoxide replicas of the surface usually enclose and "extract" the collected particles for subsequent examination. The membrane is readily dissolved in chloroform or methylene chloride. (For an example of a TEM. of asbestos on Nuclepore, see Figure 5). b. Coating with gold for scanning electron microscopy (SEM) is easily done to obtain uniform metal coat ings. 17 '18 Sample Analysis a The membrane is extremely smooth (<0.3i_im peak-tovalley distance), particularly on the shiny side (one side is dull) . Since particles larger than the pore rating are retained on the surface rather than embedded in the pore network, all particulates will be in the same focal plane and will appear with greater contrast against the smooth, lawn-like back ground. 15,17,18 b. The uniform pore diameter affords a coarse internal size calibration. 18 c. Neutron activation analysis, x-ray analysis, trans mission infrared and emission spectroscopy, and atomic absorption spectroscopy analysis of particu lates is facilitated due to the low trace element background in the membrane (ash weight is one-third that of a typical cellulosic membrane.) B . Transmission Electron Microscopy Workers at the National Center for Atmospheric Research have published techniques for preparing replicas of particulates on the Nuclepore membrane surface for transmission electron microscopy. 15 The filter was mounted with cellophane tape on a microscope slide. Silicon monoxide was vacuum evaporated to a thickness of ibout lOOoA at vertical incidence, followed by a thin (<200A) coating of chromium at 15. The membrane was then gently freed from the slide, and a 5mm square excised with clean scissors. The square was then placed SiO side down on an electron microscopy grid which, in turn, was placed upon a small square of polyurethane foam about 5mm thick in a glass petri dish. Chloroform was poured into the dish and the cover was placed on it for 7 minutes (as shown below). REPLICATION TECHNIQUE G 'K l 11 Ch r y s o tile Asbestos Pa r t ic l e s on l.OyM PORE SIZE NuCLEPORE MfMBRANE 20,000 X TEM (20) i /i The wick action of the chloroform through the polyurethane foam and through the pores of the SiO replica of the filter surface dissolved the membrane leaving the replica on the grid with fibers on top of the SiO substrate. Chloroforminsoluble particles were trapped in the replica, while soluble species left their impressions. Figure 6 is an example of a TEM prepared with this technique showing Amosite asbestos fibers on a 3ym pore size Nuclepore membrane.21 Murchio and coworkers22 have also expressed a preference for transmission electron microscopy (TEM) over scanning electron microscopy (SEM) in detecting and counting asbestos fibers due to the superior resolution obtained with TEM. Murchio's technique was similar to that used by Frank, et al15at NCAR. The specimen was, however, left in the petri dish for 4-6 hours, rather than 7 minutes, to insure that all of the membrane dissolved. C . Scanning Electron Microscopy The measurement of asbestos fiber diameters is greatly facilitated by TEM due to its superior resolution. However, for general membrane counts where fiber numbers are more important than dimensions, we believe SEM to be equally reliable and more convenient. The membrane sample is easily mounted on a brass stub with conductive silver paint and coated with a 300& thick gold film in a matter of minutes. The scanning electron microscope facilitates rapid scanning so that a larger area may be counted. Figure 7 is an example of an SEM showing chrysotile asbestos fibers on the surface. It will be noted that individual fibrils (300-500il in diameter) and fibiril bundles (up to 2000A diameter) are easily seen. The 0.2ym pores provide a convenient built-in calibration (2000A) for estimating fiber diameter. For the current study, a relatively pure sample of chrysotile asbestos was obtained. For more contaminated samples, asbestos fibers can be distinguished from other fibers and particulates by looking at the fractured ends. Gross and coworkers23 point out that asbestos fibers may be readily distinguished from non-asbestos fibers by recognizing that asbestos fibers are composed of compactly opposed finer fibrils, whereas non-asbestos fibers are homogeneous units. Because of this difference in structure, the fractured ends of asbestos fibers are irregular and resemble a bundle of fragile sticks which have been broken transversely. (Different fibers of the bundle break at different points.) This irregular profile with multiple step-like projections and depressions 15 FIGURE 6 3.0Amosite Asbestos Fibers on mm pore s i z e Nuclepore Membrane 22,000 X TEM (21) 16 FIGURE 7 Chrysotile Asbestos F ibers collected 0,2on um pore s iz e Nuclepore Membrane 10,000 X SEM  212SEC V S :2900 HS FIGURE 3 X - ray Fluorescence Spectrogram of Ch r y s o tile Asbestos F ibers on a Nuclepore Membrane mounted on a SEMBrass Stub with Gold Coating for is not seen with non-asbestos fibers which are generally homogeneous in character and exhibit uninterrupted linear profiles at the fractured ends. In addition, scanning electron microscopy (surface view) shows the longitudinal, parallel delineation of individual asbestos fibrils, whereas a sample of a non-asbestos fiber is devoid of such longitud inal lines. These characteristics are evident in the asbestos fibers shown in the SEM of Figure 7. Fractured ends are also evident in the TEMs of Figures 5 and 6. Doubtful cases can be resolved with the use of electron diffraction, microprobe techniques, or x-ray fluorescence spectrometry. The latter was used in conjunction with the SEM to look at the fibers of Figure 7; the results are shown in Figure 8 . The spectrogram registers peaks of all elements present having an atomic number greater than 8--Mg and Si from the chrysotile asbestos; Au from the gold coating, Ag from the silver paint; Cu and Zn from the brass stub mount; and Fe from magnetic contamination. Estimates of the total number of fibers collected on a membrane can be made by counting the number of fibers on a statistically appropriate number of fields in the SEM-as discussed below. This in turn can be directly related to the volume of the original liquid suspension and the concentration of fibers in that suspension. III. Retention of Asbestos Fibers on Nuclepore Membranes The following will summarize the experimental procedures and data acquired in a study of the retention of chrysotile asbestos fibers on six different pore size Nuclepore membranes. Chrysotile asbestos was selected for these studies for several reasons: 1) It is by far the most common variety of asbestos-accounting for over 90% of the world's usage. 2) It is the fiber of choice in asbestos filtration pads for use in the food and pharmaceutical industries and therefore represents the type most likely to contaminate these products. 3) With individual fibril diameters down to 300 Angstroms, it represents the smallest commonly known fiber and should therefore, present the membrane with its severest challenge. A . Experimental Procedure Preparation of Asbestos Test Suspensions The chrysotile asbestos used in this study was Union Carbide's RG-144 which is a highly refined, purified fiber containing more than 98% chrysotile fiber and less than 0.7% magnetite. The RG-144 is also unique because it is said to contain a higher concentration of fully liberated asbestos fibrils than any other chrysotile product. Thus, if these claims are correct, it should represent a severe challenge for the Nuclepore membranes. 18 It has been estimated that there are about 1x1014 colloidal fibrils per gram of RG-144 asbestos. Since the "stock" suspension was required to have a concentration of 1-5x1010 fibrils per ml, 0.3 gram was suspended in one liter of ultrapure water using a blender to yield a theoretical concentration of 3x10 10 fibrils per ml. The resulting suspension was allowed to settle overnight and the uppermost 100 ml withdrawn by pipette. This 100 ml was labeled "stock" suspension. Serial dilutions were then made from the "stock" suspension; each dilution was 100 fold. The "stock" suspension and its various dilutions were then filtered through a O.lym pore size Nuclepore membrane to determine the concentrations of the suspen sions. The results were as follows: STOCK A B C D 1.29x10 10/ml 1.29x10 8/ml 1.29x10 6/ml 1.29xl0Vml 1.29x102/ml The stock solution had 43% of the theoretical concentra tion which is not unreasonable when 24 hour settling is taken into account. Figure 9 is a typical SEM field of the membrane surface used to collect the fibers in suspension "A". One ml of the suspension was diluted with ultrapure water and washed through a 47mm 0.1)-im pore size Nuclepore membrane, with a "capture" area of 9.6cm z. One SEM field at 10,000X had an area of 9.71xl0~7cm2. The average number of fibers found in one SEM field was 13. Therefore, the concentra tion of "A" suspension was 99..671x10-7-)= 1.29x10 fibers/ml 2. Background Counts In view of the difficulties encountered by Manalan and Reti (of Millipore) in reducing background counts to zero (clean water showed 103 fibers/liter), exceptional precautions were taken to eliminate background counts in this study. Fortunately, we had access to water from Nuclepore's ultrapure water loop. This is a recirculating loop utilizing reverse osmosis, ion exchange, UV sterilization, and final filtration with an 0.2]-tm pore size Nuclepore pleated cartridge. 19 FIGURE 9 Ch ryso tile Asbestos F ibers from O.lyMSuspension "A " collected on Nuclepore Membrane, 10,000 X SEM FIGURE 10 Bl a n k 0.1yM Nu c l e p o r e Me m b r a n e AFTER PASSING 1 LITER OF ULTRAPURE WATER THROUGH,10,000 X SEM Blanks were run passing 1 liter samples of this ultrapure water through the test hardware described below (washed in ultrapure water), collecting on an O.lym Nuclepore membrane, and scanning with the electron microscope. Because of the exceptional cleanliness of the membrane in these scans (see Figure 10), 860 fields were scanned at 10,000X (each having an area of 10-6cm2) on each SEM mount (a square measuring 0.43cm x 0.43cm). The scans were made in the form of a cross, as shown below. SEM mounts were made from different positions on the membrane filter to check against mal-distribution of contamination. * NO asbestos fibers were discovered in the course of these scans. One organic fiber (lint,perhaps) was detected on one of the "blanks"--probably due to con tamination from the atmosphere while transferring the membrane sample from its holder to the SEM mounting stub. The reliable zero background count as determined above, greatly facilitated interpretation of results. 3. Challenge Tests The determination of retention of asbestos fibers by the six pore sizes of Nuclepore membranes tested was accomp lished by a two stage filtration: 1) the challenge suspension was passed through the test membrane in a 47mm Nuclepore Swin-Lok polycarbonate holder; 2) the filtrate from this holder (which was potentially contaminated with fibers which slipped through the pores of the test filter) was then passed through a 47mm, or 13mm, O.lym Nuclepore membrane held in a Hydrosol (funnel type) stainless steel filter holder. The collection filter (0.1ym) was removed from its holder and one or more sections cut out for mounting on a brass SEM stub,as described previously. For samples having low counts of fibers on the membrane surface, three cuts were made (center, mid-radius and edge) to check for mal distribution of particles in the Hydrosol type of holder. 21 Two standard challenges were selected for use in testing all pore sizes from 0.1 urn to l.Oum - 1.29x1010 and 1.29x1011 fibers. This was a compromise between the desire to insure that every pore in the membrane was challenged and the fear of significant fiber "matting" which would increase the apparent retention of fibers by the membrane. Of course, if fibers are capable of passing through the pores of a specific membrane, a certain number will pass through before a matte can form. In other words, a membrane which shows absolute retention for asbestos cannot be said to do so because of fiber matting. 4. Fiber Counting Procedures . Fiber counting was usually done at 10,000X. Three count ing procedures were used depending on the number of fibers found on the O.li-im collection filter. 1) If the number of fibers on the membrane ranged between 50-150 fibers per field (10-6cm2), two fields were selected at random and counted. This amounted to 2x10" 5% of the total membrane surface. 2) If the number of fibers ranged between 1-50 fibers per field, 10 fields were selected at random and counted. This amounted to roughly 10-1*% of the membrane surface. 3) If the number of fibers per field was less than one, a large number of scans were made across the entire SEM mount in "cross" fashion as discussed above (860 fields were scanned per SEM mount). The results from 2-6 SEM mounts were then averaged. This amounted to roughly .02 to .05% of the membrane surface when 47mm discs were used. Smaller area membranes (13mm) were also used to concentrate the fibers in a smaller area (0 .6cm2 as opposed to 9.6cm2). Only one SEM mount could be obtained from these 13mm membranes. In this case, a total of 860 fields were scanned, representing 0.14% of the total collection surface. B . Experimental Results 1. Retention as a Function of Pore Size Table III is a summary of the retention data obtained on Nuclepore membranes with pore sizes ranging between 0.1 and l.Oym. The total number of fibers passing the membrane and the percentage of the challenge which passed are calculated. 22 Membrane Pore Size (Microns) 0.10 0.10 0.20 0.20 0.40 0.60 0.60 0.80 0.80 1.00 1.00 TABLE III RETENTION OF CHRYSOTILE ASBESTOS FIBERS ON NUCLEPORE MEMBRANES AS A FUNCTION OF PORE SIZE Challenge (Fibers) 1.29x1010 1.29x1011 1.29x1010 1.29x10 11 1.29x10 10 1.29x1011 1.29x1010 1.29x1011 1.29x1010 1.29x1011 1.29x1010 1.29x1011 No. of Fields Scanned 2580 2580 1720 1720 5160 3440 10 9 2 2 2 2 Total Fibers Passed1 0 0 0 0 2.42x10" 2.79x103 4.13x10 7 4.4 3x10 7 7.42x108 6.92x10s 1.44x109 1.24xl09 Maximum Fibers Passed2 3.53x10" 7.38x10 3 6.14x10 7 5.85x10 7 10.54x10 10.05x10 1.78x10 1.55x10 Percent Passed (%) 0 0 0 0 1 .88x 10" " 2.16xl0-6 0.320 0.0343 5.75 0.536 11.16 0.961 Maximum Perce Passed2 (%) - 2.74x10" " 5.72x10" 5 0.476 0.0453 8.17 0.779 13.80 1.20 o o 23 1. Total calculated from area factor, total fibers passed 2. 95% confidence range calculated from one-sided t test. fibers counted (maermebarasnceannaerdea ) There is, of course, some uncertaintly associated with missing some fibers which should have been counted. Therefore, the question was asked, "what value represents the maximum number which could have passed the membrane with only a 5% chance that the number is higher still?" This 95% confidence limit was calculated from a "one-sided t test". N max,95% -- N + t s(N) where N max,95% = the maximum number of fibers passing ' the membrane within 95% confidence _ limits N = the average number of fibers passing the membrane t = the difference between the mean of a sample and the true mean of the popula tion from which the sample was drawn, divided by the estimated standard deviation of the mean - determined from _ standard statistical tables s(N) = the estimated standard deviation of the mean. This is estimated by dividing the standard deviation of the counts by the square root of the number of counts The results of applying the ttest with a 95% confidence level are shown in Table III as the "maximum fibers passed" and the "maximum percent of challenge passed". Figures 11 through 16 are photomicrographs of the O.lym collection membranes used to collect asbestos fibers in the filtrates from 1.0, 0.8, 0.6, and 0.4ym pore size membranes at the challenge levels of Table III. It will be noted that stray contamination with a diameter of 0.5ym is evident from the filtrates of the l.Oym filter (see Figure 11). There is a progression towards shorter fiber lengths with the smaller pore size filtrates. The lengths are however, often 2 to 3 times larger than pore size through which they passed. The number of fibers appearing in a given field decreases with decreasing pore size; it markedly decreases as the pore size decreases from 0.6 to 0.4Mm (compare Figures 15 and 16). With the 0.4ym pore size, no field on the collection filter contained more than one fiber and most had zero. Figure 16 (0.4ym filtrate) shows a contaminant particle (not asbestos) having an 0 .2ym diameter; this would not have been counted. Figures 17 and 18 plot the percentage of the challenge passing the Nuclepore membrane as a function of pore size on semilog and linear graph paper. Figure 18 shows that the 0.4ym membrane removes virtually all of the asbestos. However, Table III and Figure 17 show that only the 0.2 and O.lym pore sizes show truly absolute retention. (This was further checked by running an additional filtrate through a 13mm membrane and scanning 860 fields. Again, no fibers could be found.) FIGURE 11 Ch ryso tile Asbestos F ibers in F il t r a t e from l.OyM Nuclepore Membrane- challenged with 1.29xl010 fib e r s , 10,000 X (45"angle) SEM (Note 0 , 5 mm stray c o n ta m in a tio n ) FIGURE 12 Ch r y s o tile Asbestos F ibers in F il t r a t e from l.OyM Nuclepore Membane- challenged with 1.29x1011 f ib e r s , 10,000 X SEF1 FIGURE 13 Ch r y s o tile Asbestos F ibers in I Fil t r a t e from 0 , 8 uM Nuclepore Membrane- challenged with 1.29x1010 f ib e r s , 10,000 X SEM FIGURE 14 Ch r y s o t i l e Asbestos F ibers in F il t r a t e from 0 . 8 viM Nuclepore Membrane- challenged wi th 1.29x10" fibers, 10,000 X SEM FIGURE 15 Chrysoti le Asbestos F ibers in F ilt ra te from 0 . 6 yM Nuclepore Membrane- challenged wi th 1.29x10" fibers, 10,000 X SEF1 FIGURE 16 Ch r y s o t i l e Asbestos F i bers in i Fi l t r a t e from 0 , % m Nuclepore | Membrane- challenged with | 1.29xl010 fibers, 10,000 X SEM (Note 0 , 2 fm stray contaminant) FIGURE 17 n G C % 28 C8o')) G E N E R A L E L E C T R IC C O M PA N Y , S C H E N E C T A D Y , N. Y.. U.S.A. 14'1'H) IncL L!vis;on*> POPF, SIZE (MICRONS) sO The difference in percent retention (or passage) for the two different challenge levels is believed to reflect the phenomenon of fiber matting which is increased at the higher challenge resulting in reduced passage of fibers. Figure 19 is a plot of the equivalent data from the work of Manalan and Reti^with Millipore membranes. The reason for the cross-over of the two challenge curves is unknown. It has been suggested that the non-zero background counts may explain the scatter. However, a finite background (1000 fibers/liter) should only lend uncertainty to low fiber count data whereas the most difficult data point(on Figure 19)to explain is the higher percentage passage with a higher challenge level at the 1.2ym pore size where large fiber counts were obtained. Photomicrographs were also made of the test membrane surface to better ascertain the degree of matting. Figures 20-27 are a collection of SEMs taken at 10,000 and 3,00OX showing the total matte build-up on the surface. There is considerable mal-distribution of fibers on the membrane surface with free areas for passage in some cases, and dense entanglement in others. As expected, the fibers in these photomicrographs are consid erably longer than those detected in the filtrates (Figures 11-16) . 2. Retention as a Function of Challenge Level Since the effect of fiber matting at the challenge levels of 1010 and 1011 fibers was shown to be significant, a study of the effect of lower challenge levels was run on the 0.4ym pore size membrane. The 0.4ym pore size was selected since it represented the cut-off point between absolute retention and detectable fiber passage. Table IV indicates the total fibers passed and the percentage passed as a function of challenge level down to the 1.29xl06 fiber level. Again, the one-sidedt test was used to calculate maximum fiber passage at the 95% confidence level. Figure 28 plots the percent of challenge passing an 0.4ym Nuclepore membrane as a function of challenge level. As expected, the retention is enhanced by higher challenge levels. The percentage passed at the 106 challenge level is not plotted since only one fiber found in 1720 fields makes the difference between 0 and 0.4% passed. 30 o PORE SIZE (MICRONS) Challenge (Fibers) 1.29x10 6 1.29x10 7 1.29x10s 1.29x10 1.29x1010 1.29x1011 TABLE IV RETENTION OF CHRYSOTILE ASBESTOS FIBERS ON 0.4ym NUCLEPORE AS A FUNCTION OF CHALLENGE LEVEL No. of Fields Scanned 1720 1720 1720 1720 5160 3440 Total Fibers Passed1 0 2.24x10 4 4.46x10** 1.11x10 ** 2.42x10** 2.79x10 3 Maximum Fibers Passed2 - 4.07x10** 7.07x10** 2.41x104 3.53x10** 7.38x10 Percent Passed (%) Maximum Percent Passed2 (%) 0- 0.174 0.316 0.0346 0.0548 8.6x10"** 1.87x10" 1.88x10"** 2.74x10"** 2.16xl0-6 5.72x10"6 1. Total calculated from area factor, total fibers passed 2. 95% confidence range calculated from one-sided t test. fibers counted .membrane area, 'area canned FIGURE 20 Chrysotilf. Asbestos F ibers Re t a i n e d on 1.0mm IJuclepore Membrane ( 1,29 x 1010Fibers Di s t r i b u t e d 9 . 6 cm2) 3,000 X SEM on FIGURE 21 Ch r y so t il e Asbestos F i bers Re ta i ne d on l.OyM Nuclepore Membrane ( 1, 29 x 1 0 10 F ibers Dis t r i b u t e d on 9 , 6 cm2) 10,000 X SEM FIGURE 22 Chrysoti le Asbestos F ibers Ret ai ne d on 0 , 8yM Nuclepore Membrane (1,29 x 1010 F ibers Dis t r ib u t e d on 9 ,6 cm2) 3,000 X SEM Figure 23 Ch r y s o t i l e Asbestos F ibers Re tai ned ON O.oyM NuCLEPORE MEMBRANE (1.29 x 10io Fibers Distributed on 9 . 6 cm2) 10,000 X SEFI OJ Ln FIGURE 24 Ch r y so t il e Asbestos Fibers Re ta i n e d on 0,6pm Nuclepore Membrane (1.29 x 1011 Fibers Dis t r ib u t e d on 9 .6 cm2) 10,000 X SEM u<y>\ FIGURE 25 Chrysotile Asbestos F ibers Re ta i n e d on O.Amm Nuclepore Membrane (1.29 x 1010 Fibers Distributed on 9 . 6 cm2) 3^000 x SEM FIGURE 26 Ch ry so t il e Asbestos F ibers Re tai ned on O.AyM Nuclepore Membrane (1,29 x 101Q Fibers Distributed on 9 . 6 cm2) 10,000 X SEM J FIGURE 27 Ch ry so t il e Asbestos Fibers Re ta i ne d ON 0 2yM [`JUCLEPORE MEMBRANE (1.29 x 1010 Fibers Dis tributed on 9 .6 cm2) 10,900 X SEF1 FIGURE 28 i o 7 i o 8 i o 9 i o 10 i o 11 38 Figure 29 plots the number of fibers detected down stream as a function of challenge level. The number of fibers increases with increasing challenge until levels of 108- 109 are attained and thereafter declines. Figure 29 can also be looked upon as a plot of the number of fibers passing as a function of throughput volume. Every ml passing through the filter adds to the membrane's challenge. As the volumetric through put increases, the fiber build-up on the membrane increases. The plot indicates that when enough volume has passed through the membrane to lay down 108- 109 fibers, significant matting begins to form and further passage is retarded. Thus, the number of fibers detected downstream would be expected to Level off at a constantmaximum value. The fact that the number declines indicates that at the higher challenge levels, the concentration is high enough to result in some entanglement of fibers prior to deposition on the membrane surface. This may explain the mal-distribution of fibers seen in Figures 20-27. IV. Conclusion A. The Nuclepore membrane greatly facilitates detection and counting of asbestos fibers by virtue of its extremely smooth, flat surface against which the asbestos fibers are easily seen and counted. B. Scanning electron microscopy appears to offer some convenience over transmission electron microscopy in sample preparation, asbestos fiber identification, and in rapid counting scans. C. The O.lym and 0.2ym pore size Nuclepore membranes appear to offer absolute retention of chrysotile asbestos fibers. D. The 0.4ym pore size Nuclepore membrane appears to offer virtual retention of all chrysotile asbestos fibers with maximum pene tration rated at 0.3%. E. Increasing challenge levels result in decreased fiber passage presumably due to fiber matting. F. Significant matting appeared to occur at challenge levels above 108 fibers/cm2. 39 i 7 n SEMI-LOGARITHMIC 46 6213 r V rJ ^ 5 7 - Z L :.Z < "C B ' / ' S i O N S h i ;- n u s i C c- fT> ^ X IT- Tlr>J CO i 6 5 UMBER OF FIBERS (104) AS SING -MEMBRANE - 2 1 0 REFERENCES 1. HENDRY, N.W., "The Geology, Occurences and Major Uses of Asbestos", Annals New York Academy of Sciences, Section 1, Asbestos Materials in Modern Technology, pp. 12-22. 2. BYRNE, R.E., Jr., "Asbestos Fibers", Modern Plastics Encyclopedia, McGraw-Hill, N.Y., 1969-1970. 3. WHITTAKER, E.J.W., Acta Cryst. 9, 855-862 (1956), 10, 149-156 (1957) . 4. PAULING, L., Proc. Nat. Acad. Sci. U.S., 16, 578-582 (1930) . 5. PUNDSACK, F.L., J. Phys. Chem. 59, 892-895 (1955). 6. NAUMANN, A.W., "Properties of Asbestos Suitable for Use in Cellulosic Paper", Union Carbide Corp., Tuxedo, New York. 7. WYLLIE, D.M., "Asbestos Filter Sheets", Filtration & Separation, March-April, 1973, p. 175ff. 8. "Asbestos Particles in Food and Drugs", Federal Register vol. 36, No. 188, Friday, Sept. 28, 1973. 9. NICHOLSON, W.J., MAGGIORE, C.J., and SELIKOFF, I.J., "Asbestos Contamination of Parenteral Drugs", Science, vol. 177, July 14, 1972, pp. 172-173. 10. SELIKOFF, I.J., HAMMOND, E.C., and CHURG, J., American Medical Association 204, 106 (1968). 11. PONTEFRACT, R.D. and CUNNINGHAM, H.M., "Penetration of Asbestos Through the Digestive Tract of Rats". Nature, vol. 243, June 8, 1973, pp. 352-353. 12. HARINGTON, J.S. and ROE, F.J.C., "Studies of Carcino genesis of Asbestos Fibers and Their Natural Fibers", Annals of the N.Y. Academy of Sciences, 132: 439 450, 1965. 13. Concentrates, Chemical and Engineering News, Jan. 8, 1973, p . 13. 14. MANALAN, D.A. and RETI, A.R., "Detection and Removal of Asbestos Fibers From Liquids", Paper No. 53a, Symposium No. 53-Membranes and Ultrafiltration, 66th Annual Meeting of A.I.Ch.E., Nov. 11-15, 1973, Philadelphia, Pennsylvania. 4i 15. FRANK, E.R., SPURNY, K.R., SHEESLEY, D.C. and LODGE, J.P.Jr., "The Use of Nuclepore Filters in Light and Electron Microscopy of Aerosols", Extrait du Journal De Micrscopie, vol. 9, No. 6 (1970), pp. 735-740. 16. SPURNY, K.R., LODGE, J.P.Jr., FRANK, E.R., and SHEESLEY, D.C., "Aerosol Filtration by Means of Nuclepore Filters-Aerosol Sample and Measurement", Environmental Science and Technology, vol. 3, No. 5, May 1969, pp. 464-468. 17. DENEE, P.B., and STEIN, R.L., "An Evaluation of Dust Sampling Membrane Filters for Use in the Scanning Electron Microscope", Powder Technol., 5 (1971/ 1972), pp. 201-204. 18 . TODD, R.L. and KERR, T.J., "Scanning Electron Microscopy of Microbial Cells on Membrane Filters", Applied Microbiology, June, 1972, pp. 1160-1162. 19. PORTER, M.C., "Particulate Analysis with Membrane Filters", Industrial Research, Jan. 1974--in press. 2 0 . Photomicrograph taken by D.M. COULSON and I.C. NIEMEYER of Stanford Research Institute, Menlo Park, CA. 2 1 . Photo supplied by Dr. KUELOSLOV SPURNY of the Institut fur Aerobiologie, Sauerland, Germany and prepared by EVELYN ACKERMAN of the National Center for Atmospheric Research, Boulder, Colorado. 22 . MURCHIO, J.C.; COOPER, W.C.; and DE LEON, A., "Asbestos Fibers in Ambient Air of California:, Final Report to the California Air Resources Board (EHS #73-2), March 1, 1973. 23. GROSS, P; deTREVILLE, R.T.P; and HALLER, N., "Asbestos Versus Non-asbestos Fibers", Arch Environ Health, vol 20, May, 1970, pp 571-578. 42