U.S. DEPARTMENT Of COMMENCE
                         National Technical Information Service
                         PB80-152879
Interim  Method  for
Determining  Asbestos in
Water
(U.S.) kiivironmentnl Research  Lab,. Athens, GA
Jan 80

-------
United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30605
EPA 600 
-------
                RESEARCH REPORTING SERIES

Research r^-eons of me Office 01 Research anc: Development U S Environmental
Protection Agency na>.o baen grouped into nine series  These nine broad cate-
gories were estaolisned to faci.itate further development and application of er-
viror.mental tecnnoiogy   Elimination of traditional grouping  was consc'Cusiy
piannea ;o tos'.er techno'ogy uar.ster and a maximum interface in reiaied fields
The nine series are

      1   Environmental  Health Eftecis Research
      2.  Environmental  Protection Tecnnoiogy
      2   Ecological Research
      4   Environmental  Monitoring
      5   Socioeconomic Environmental  Studies
      6   Scientific ana Tecnmcal Assessment Reports (STAR)
      7   Interagency Energy-Environment Ressarch and Development
      3   'Special' Reoorts
      9   Miscellaneous  Reports

This report has been ass.gned 'o the ENVIRONMENTAL MONITORING series
This series describes research conducted to develop new or improved meihods
and instrumentation for  the identification ard quantification of  environmental
pollutants at the lowest conceivably  significant concentrations It also u,eludes
studies to determine the ambient concentrations of pollutants in the environment
and.'o; the variance of pollutants as a function of time or meteorological ractors.
This document is availab'e to the public through the National Technical Informa-
tion Service. Sonngfield. Virginia 22101.

-------
                              TECHNICAL REPORT DATA
                        /Please read Insirui'tiuiis on the /vrmr beJJre completing)
 1. Rcr'ORT NO.

  EPA-600/4-80-OOS
                                                   3. RECIPIENT'
                                                                    O.
 j. TITLE A/MO SUBTITLE
  Interim Method for Determining Asbestos
  in Water
            5. REPORT DATE
            January. 1980 issuing date
            6. PERFORMING ORGANIZATION CODE
 •>. AUTHOR(S)
                                                   8. PERFORMING ORGANIZATION REPORT NO.
  Charles H. Anderson and J.  MacArthur Long
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
  Environmental  Research  Laboratory
  U.S.  Environmental Protection  Agency
  Athens, Georgia  30605
            10. PROGRAM ELEMENT NO.

              A37B1D
            11. CONTRACT/GRANT NO.
 12. SPONSORING AGSNCY NAME AND ADDRESS
  Environmental Research Laboratory-Athens,  GA
  Office of Research and Development
  U.S.  Environmental Protection Agency
  Athens, Georgia  30605
            13. TYPE OF REPOR-r AND PERIOD COVEHF.O
             Intarim, 7/76-12/78
            14. SPONSORING AGENCY CODE
             EPA/600/01
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
       This manual  describes  an  interim electron microscope  (EM)
  procedure for measuring the concentration  of asbestos in water samples.
  The  main features of the method include  filtering the sample through a
  sub-micron polycarbonate membrane filter,  examining an EM  specimen grid
  in a transmission electron  microscope  (TEM) ,  and verify.ing fiber
  identity by selected area electro'n diffraction (SAED) .

       This interim method is a  revision of  the procedure issued in
  1976 and reflects the improvements that  have been made in  asbestos
  analytical methodology since that time.
                           KEY WORDS AND DOCUMENT ANALYSIS
               DESCRIPTORS
 Water  pollution
 Asbestos
 Serpentine
 Amphiboles
 Electron Microscopy
 Electron Diffraction
                                       bJC/ENTIFIERS/OPEN ENDED TERMS
 Asbestos Measurement
                       c. COSATI Field/Group
   07D
   14B
   68D
  013 rmguTiON STATEMENT

 Release to Public
19. SECURITY CLASS (Thi\ Reran/
Unclassified
21. NO. OF PAGES
    44
                                       JO. SECURI TY CLASS I'l'llis paRf)

                                        Unclassified
                                                              22. PRICE
EPA Form 2220-1 (9-7?)
                                                          • US' 
-------
                                       EP^-600/4-80-005
                                       January  1980
INTERIM METHOD FOR DETERMINING ASBESTOS IN WATER
                       by

   Charles H.  Anderson and J.  MacArthur Long
          Analytical Chemistry Branch
       Environmental Research  Laboratory
             Athens, Georgia  30605
       ENVIRONMENTAL RES2ARCH LABORATORY
       OFFICE OF RESEARCH AND DEVELOPMENT
      U.S.  ENVIRONMENTAL PROTECTION AGENCY
             ATHENS, GEORGIA  30605

-------
                           DISCLAIMS*
     This report has been reviewed by the Environmental
Research Laboratory, U.S. Environmental Protection Agency,
Athens, Georgia, and approved for publication.  Mention of
trade names or commercial products does not constitute
endorsement or recommendation for use.
                               ii

-------
                            FOREWORD


     Nearly every phase 
-------
                            PREFACE
     In July 1976, the Preliminary Interim Method for Deter-
mining Asbestos in Water was issued by the U.S. £avironmental
Protection Agency's Environmental Research Laboratory in
Athens, Georgia.  That method was perceived as representing
the current state-of-the-art in asbestos analytical method-
ology.  The objective of writing the method was to present a
procedure that analytical laboratories could follow chat woulJ
result in a better agreement of analytical results.  Since
that time, a significant amount of additional experimental
work has generated data that provide the basis for a more
definitive method than was possible previously.

     This revised Interim "ethod reflects the improvements
that have been made in asbestos analytical methodology since
the initial procedure was drafted.  The general approach to
the analytical determination, however, remains the same as
previously outlined.  That is, asbestos fibers are separated
from water by filtration on a sub-micron pore size membrane
filter.  The asbestos fibers are then counted, after
dissolving the filter material, by direct observation in a
transmission electron microscope.

     The major change in the initial procedure is the elimina-
tion of the condensation washer as a means of sample prepara-
tion.   Intra- and inter-laboratory precision data for the
method are presented.  Also, a suggested statistical evalua-
tion of grid fiber counts is included.
                               iv

-------
                            ABSTRACT
     An interim electron microscope  (KM) procedure for
measuring the concentration of asbestos in water samples is
described.  The main features of the method include filtering
the samples through a sub-micron polycarbonate membrane
filter, examining an EM specimen grid in a transmission
electron microscope (TRM), and verifying fiber identity by
selected area electron diffraction.

     This interim method is a revision of the procedure issued
in 1976 and reflects the improvements that have been made in
asbestos analytical methodology since that time.

     This report covers a period from July 1976 to December
1978 and work was completed as of December 1978.

-------
                        CONTENTS
Foreword	   iii
Preface   	    iv
Abstract	     v
Figures and Tables	    ix

     1.    Scope and Application  	     1
     2.    Summary of Method	     1
     3.    Definitions  	     2
     4.    Sample Handling and Preservation 	     3
                  4.1    Containment Vessel   ....     3
                  4.2    Quantity of Sample   ....     3
                  4.3    Sample Preservation  ....     3
     5.    Interferences  	     4
                  5.1    Misidentification  	     4
                  5.2    Obscuration  	     4
                  5.3    Contamination  	     5
                  5.4    Freezing	     5
     6.    Equipment and Apparatus  	     5
                  6.1    Specimen Preparation
                              Laboratory  	     5
                  6.2    Instrumentation    	     6
                  6.3    Apparatus Supplias and
                              Reagents    	     7
     7.    Preparation of Standards	.,  . . .    11
                  7.1    Chrysotile Sto< k
                              Suspensirn	    11
                  7.2    Amphibole Stozfc
                              Suspension	    11
                  7.3    Identification Standards . .    11
     8.    Procedure	    11
                  8.1    Filtration   	    11
                  8.2    Preparation of Electron
                              Microscope Grids  ...    13
                  8.3    Nuclepore Filter, Modified
                              Jaffe Wick Technique  .    14
                  8.4    Electron Microsccspic
                              Examination	    16
                  8.5    Ashing	    22
                  8.6    Determination of Blank
                              Level	    23
                         vii

-------
     9.
     10.
     11.
     12.
     13.
Calculations
       9.1
       9.2
       9.3
Reporting
Precision
       11.
       11.
Accuracy  .
       12
       12
                         Fiber Concentrations
                         Estimated Mas:-;
                              Concentration  .
                         Aspect Ratio    .  .  .
                         Intra-Laboratory
                         Inter-Laboratory
          1   Fiber Concentrations
          ,2   Mass Concentrations
Suggested Statistical Evaluation of
Grid Fiber Counts  ...  	
Bibliography
24
24

25
26
26
26
26
27
27
27
29

2?

32
                         Vlll

-------
                          FIGURES
Number
  1.
  2.
Modified Jaffe Wick Method
Illustration of Counting Rules for
   Field-of-View Method" 	
Page
  8

 20
                           TABLES
Number
  1.
  2.
Intra-Laboratory Precision
Inter-Laboratory Precision
 28
 28
                           IX

-------
                         INTERIM METHOD
               FOR DETERMINING ASBESTOS IN WATER
1.    Scope and Application

     1.1   This method is applicable to drinking water and
           water supplies.

     1.2   The method determines the number of asbestos fibers
           per liter, the size (length and width) of the
           fibers, the size distribution, and the total mass.
           The method distinguishes chrysotile from amphibole
           asbestos.   The detection limits are variable and
           depend upon the amount of total extraneous pacticu-
           late matter in the sample as well as the contamina-
           tion level in the laboratory environment.  Under
           favorable  circumstances, 0.01 million fibers per
           liter  (MFL) can be detected.  The detection limit
           for total  mass of asbestos ribers is also variable
           and depends upon the fiber size and size distribu-
           tion in addition to the factors affecting the total
           fibar count.  The detection limit under favorable
           conditions is in the order of 0.1 nanogram per
           liter  (ng/L).

     1.3   The method is not intended to furnish a complete
           characterization of all the fibers in water.

     1.4   It is beyond the scope of this method to furnish
           detailed instruction in electron microscopy,
           electron diffraction,  or crystallography.  It is
           assumed that those using this method will be suffi-
           ciently knowledgeable in these fields to understand
           th« methodology involved.

2.    Summary of Method

     2.1   A variable, known volume of water sample is
           filtered through a 0.1 micrometer (ura) Nuclepore
           filter to  trap asbestos fibers and the filter is
           then carbon coated.  A small portion of the carbon
           coated filter with deposited fibers is placed on- an
           electron microscope grid and the filter material is
           removed by gentle solution in organic solvent.  The

-------
           material remaining on the electron microscope grid
           is examined in a transmission microscope at a
           magnification of about 20,900X.  The asbestos
           fibers are identified by their morphology and
           electron diffraction patterns and their lengths and
           widths are measured.  The total arsa exaaiir.e<3 in
           the electron microscope is determined and the
           nurobsr of asbestos fibers in this area is counted.
           The concentration in MFL is calculated from the
           number of fibers counted, the amount of water
           filtared, and the ratio of the total filtered area/
           sampled filter area.  The mass per liter is calcu-
           lated from the assumed density and the volume of
           the fibers.

3.    Definitions

     Asbestos - A generic term applied to a variety of
           commercially useful fibrous silicate minerals of
           th« serpentine or amphibole mineral groups.

     Fiber - Any particulate that has parallel sides and a
           length/width ratio greater than or equal to 3:1.

     Aspect Ratio - The ratio of length to width.

     Chrysotile - A nearly pure hydrated magnesium silicate,
           the fibrous form of the mineral serpentine,
           possessing a unique layered structure in which the
           layers are wrapped in a helical cylindrical manner
           about the fiber axis.

     Amphibole Asbestos - A double chain fibrous silicate
           mineral consisting of Si^O^^,   laterally linked by
           various cations such as aluminum/ calcium, iron,
           magnesium, and sodium.  Amphibole asbestos consists
           of crocidolite and araosite (the fibrous form of
           cummingtonite-gruenerite) , and the fibrous forms of
           tremolite, actinolite, and ant-' jphyllite.  These
           minerals consist of or contain fibers formed
           through natural growth processes.  Mineral frag-
           ments that conform to the definition of a fiber and
           that are formed through a crushing and milling
           process are analytically indistinguishable from the
           naturally formed fibers by this method.

     Detection Limit - The calculated concentration in MFL,
           equivalent to one fiber above the background or
           blank count (Section 8.6).

-------
     Statistically Significant - Any concentration based upon a
           total fiber count of 5 or more in 20 grid squares.

4.   Sample Handling and Preservation

     It is beyond the scope of this procedure to furnish
     detailed instructions for field sampling; the general
     principles of obtaining water samples apply.  Some
     specific considerations apply to asbestos fibers, how-
     ever, because they are a special type of part.-'.culate
     matter.   These fibers are small, and in water range in
     length from 0.1 ym to 20 um or more.  Because of the
     range of size there may be a vertical distribution of
     particle sizes.  This distribution will vary with depth
     depending upon the vertical distribution of temperature
     as well  as local meteorological conditions.  Sampling
     should take place according to the objective of the
     analysis.  If a representative .sample of a water supply
     is required, a carefully designated set of samples should
     be taken representing the vertical as well as the hori-
     zontal distribution and these samples should be
     composited for analysis.

     4.1   Containment Vessel

           The sampling container shall be a clean, screw-
           capped, polyethylene bottle capable of holding at
           least 1 liter.  The bottle should be rinsed at
           least two times with the water that is being
           sampled prior to sampling.

     4.2   Quantity of Sample

           A minimuir of approximately 1 liter of water is
           required.  Leave air space at the top of the
           container to allow for shaking the sample.  It is
           desirable to obtain two samples from one location.

     4.3   Sample Preservation

           No preservatives should be added during sampling
           and the addition of acids should be particularly
           avoided.  If the sample cannot be filtered in the
           laboratory within 48 hours of its arrival, suffi-
           cient amounts (1 milliliter per liter of sample)  of
           a  2.71% solution of mercuric chloride to give a
           final concentration of 20 ppm of Hg may be added tc
           prevent bacterial growth.

-------
           NOTE 1:   It has been reported that the growth of
           algae in water samples can be prevented by storing
           the samples in the dark.

           NOTE 2:   Refrigeration of sampler at about 5°C
           minimize bacterial and alga]  growth.

5.    Interferences

     5.1   Misidentification

           The guidelines set forth in this method for
           counting fibrous asbestos require a positive
           identification by both morphology and crystal
           structure as shown by an electron diffraction
           pattern.  Chrysotile asbestos has a unique tubular
           structure, usually showing the presence of a
           central  canal, and exhibits a unique characteristic
           electron diffraction pattern.  Although halloysite
           fibers may show a streaking similar to chrysotile,
           they do  not exhibit chrysotile's characteristic
           triple set of double spots or 5.3A layer line.  It
           is highly improbable that a nou-asbestiform fiber
           would exhibit the distinguishing chrysotile
           features.  Although amphibole fibers exhibit
           characteristic morphology and electron diffraction
           patterns, they do not have the- unique properties
           exhibited by chrysotiie.  It is possible,  there-
           fore, though not probable for misidentification to
           take place.

           It is important to recognize that a significant
           variable fraction of both chrysotile and axphibole
           asbestos fibers do not exhibit the required
           confirmatory electron diffraction pattern.  This
           absence  of diffraction is attributable to unfavor-
           able fiber orientation and fiber sizes.  The
           results  reported will be low, therefore, as
           compared to the absolute number of asbestos fibers
           that are present.

     5.2   Obscuration

           If large amounts of other materials are present,
           some small asbestos fibers may not be observed
           because  of physical overlapping.  This will result
           in low values for the reported asbestos content.

-------
     5.3   Contamination

           Although contam5.nat.ion is not strictly considered
           to be an interference, it is an important source of
           erroneous results, particularly for chrysotile.
           The possibility of contamination, therefore, should
           always be a consideration.

     5.4   Freezing

           The effect of freezing on asbestos fibers is not
           known but there is reason to suspect that fiber
           breakdown could occur and result in a higher fiber
           count than was present in the original sample.
           Therefore, the sample should be transported to the
           laboratory under conditions that would avoid
           freezing.

6.   Equipment and Apparatus

     6.1   Specimen Preparation Laboratory

           The ubiquitous nature of asbestos, especially chry-
           sotile, demands that all sample preparation steps
           be carried out to prevent the contamination of the
           sample by airborne or other source of asbestos.
           The prime requirement of the sample preparation
           laboratory is that it be sufficiently free from
           asbestos contamination that a specimen blank deter-
           mination using 200 ml of asbestos-free water yields
           no more than 2 fibers in 20 grid squares of a
           conventional 200 mesh electron microscope grid.

           In order to achieve this low level of contamina-
           .tion, the sample preparation area should be a
           separate conventional clean room facility.  The
           room should be operated under positive pressure and
           have incorporated electrostatic precipitators in
           the air supply to the room, or as an alternative,
           absolute (HEPA) filters.  No asbestos floor or
           ceiling tiles, transite heat-resistant boards, or
           asbestos insulation should be used in construc-
           tion.  Work surfaces should be stainless steel or
           Formica or equivalent.  A laminar flow hood should
           be provided for sample manipulation.  Disposable
           plastic laboratory coats and disposable overshoes
           are recommended.  Alternatively, new shoes for all
           operators should be provided and retained for clean
           room use only.  A mat (Tacky Mat, Liberty
           Industries, 589 Deming Road, Berlin, Connecticut
           06037, or equivalent) should be placed at the

-------
      entrance to the room to trap any gross contamina-
      tion inadvertently brought into the room on contam-
      inated shoes.   Normal electrical and water
      services, including a distilled water supply should
      be provided.  In addition, a source of ultra-pure
      water from a still or filtration-ion exchange
      system is desirable.

6.2   Instrumentation

      6.2.1   Transmission Electron Microscope.  A trans-
              mission electron microscope thac operates
              a): a minimum of 80 kv and has a resolution
              of better than 1.0 nm and a magnification
              range of 300 to 100,000 is required.  If
              the upper limit is not attainable directly
              it may be attained through the use of
              auxiliary optical viewing.  Jt is mandatory
              that the instrument &
-------
6.2   Apparatus/ Supplies and Reagents

      6.3.1   Jaffe Wick Washer.  The J'affe Wick Washer
              for dissolving Nuclepore filter is
              described in 8.3.1, and is illustrated in
              Figure 1.

      6.3.2   Filtering Apparatus.  A 47-mm funnel  (Cat
              No. XX1504700, Millipore Corporation, Order
              Service Dept . , Bedford, MA  01730) or
              equivalent is  used to filter water
              samples.   A 25-mm funnel (Millipore Cat No.
              XX1002500) or  equivalent is used to filter
              dispersed ash  samples.

      6.3.3   Vacuum Pump.  A pump, for use in sample
              filtration, should provide vacuum up  to
              about 500 mm of mercury.

      6.3.4   e>l Grids.  Grids of 200 -mesh copper or
              nickel covered with formv.ir film for  use
              with the Nuclepore-Jaf f e sample preparation
              method are required.  These grids may be
              purchased from manufacturers of electron
              microscopic supplies or prepared by
              standard electron microscopic grid prepara-
              tion procedures.  Finder grids may be
              substituted and are useful if the re-exami-
              nation of a specific grio opening is
              desired.

      6.3.5   Membrane Filters.

              47-mm diameter Millipore membrane filter,
              type HA,  0.45-um pore size.  Used as a
              Nuclepore filter support on top of glass
              frit.

              47-mm diameter Nuclepoire membrane filter;
              0.1-um pore site (Nuclepore Corp., 7035
              Commerce Circle, Pleasanton, CA 945C.6) .
              Used to filter tha water sample.

              25-mm diameter Millipore membrane ::ilter,
              type HA;  0.45-ym pore size.  Used as  Nucle-
                   filter support on top of glass frit.
              25-mm diameter Nuclepore membrane ::ilter;
              0.1-um pore size.  Used to filter dispersed
              ashed Nuclepore filter.

-------
 Screen Support
     with Grid
              Petri Dish
     Ridge
Glass Slides

       A
  Layer of
Filter  Papers
Nuclepore  Filter
      \	
                        Carbon
                  Carbon
                  Chloroform
                   Forrnvar
                   Grid
                     B
          Figure 1.  Modified Jaffe Wick Method.
                  A. Washing Apparatus
                  B. Washing Process
                       8
                           Grid

-------
6.3.6   Glass Vials.  SO-.iun diameter x 80-ram long.
        Used to hold filter during ashing.

6.3.7   Glass Slides.  Used to support Nuclepore
        filter during carbon evaporation.

6.3.8   Scalpels.

6.3.9   Scissors.

6.3.10  Tweezers.  Several pairs are needed for the
        many handling operations.

6.3.11  Double-sided Tape.  Used to hold filter
        section flat on glass slide while carbon
        coating is applied.

6.3.12  Disposable Petri Dishes, 50-mm diameter.
        Used for storing membrane filters.

6.3.13  Static Eliminator, 50C microcuries Po-210.
        INuclepore Cat. No. V090POL00101) or equiv-
        alent.  Used to eliminate static charges
        from membrane filters.

6.3.14  Carbon Rods.  Spectrochemically pure, 1/8
        in. dia., 3.6 mra x 1.0 mm neck.  Used for
        carbon coating.

6.3.15  Carbon rod sharpener.  (Cat. No. 1204,
        Ernest F. Fullam, Inc., P. 0. Box 444,
        Schenectady, NY 12301) or equivalent.  Used
        for sharpening carbon rods to a neck of
        specified length and diameter.

6.3.16  Ultrasonic Bath.   (50 watts, 55 kHz).  Used
        for dispersing ashed sample and for general
        cleaning.

6.3.17  Graduated Cylinder, 500 ml.

6.3.18  Spot Plate.

6.3.19  10-yl Microsyringe.  Used for administering
        drop of solvent to filter section during
        sample preparation.

6.3.20  Carbon Grating Replica, 2160 lines/mm.
        Used for calibration of EM magnification.

-------
6.3.21  Filter Paper.  S & S #589 Black Ribbon
        (9-cm circles) or equivalent absorbent
        filter paper.  Used for preparing Jaffe
        Wick Washer.

6.3.22  Screen supports  (copper or stainless steel)
        12 mm x 12 mm, 200 mesh or equivalent.
        Used to support specimen grid in Jaffe Wick
        Washer.

6.3.23  Chloroform.  Spectro grade, doubly
        distilled.  Used for dissolving Nuclepore
        filters.

6.3.24  Asbestos.  Chtysotile  (Canadian), Crocido-
        lite, Amosite.  UICC (Union Internationale
        Centre le Cancer) Standards.  Available
        from Duke Standards Company, 445 Sherman
        Avenue, Palo Alto, CA 94306.

6.3.25  Petri Dish.  Glass dish (100 mm diameter x
        15 mm high).  Used for modified Jaffe Wick
        Washer.

6.3.2(>  Alconox.   (Alconox, Inc., New York, NY
        10003) or equivalent.  Used for clean-ng
        glassware.  Add 7.5 g Alconox to a liter of
        distilled water.

6.3.27  Parafilm.   (American Can Company, Neenah,
        WI) or equivalent.  Used as protective
        covering for clean glassware.

6.3.28  Pipets.  Disposable, 5 ml and 50 ml pipets
        are required.

6.3.29  Distilled or Deionized Water.  Filter if
        necessary through 0.1-ym Nuclepore filter
        for making up all reagents, for final
        rinsing of glassware, and for preparing
        blanks.

6.3.30  Mercuric chloride, 2.71% solution w/v.
        Used as sample preservative.  See 4.3.  Add
        5.42 g or reagent grade mercuric chloride
        (HqCl2) to 100 ml distilled water and
        dissolve by shaking.  Dilute to 200 ml with
        additional water.  Filter through O.i-ym
        Nuclepore filter paper before using.
                   10

-------
7.   Preparation of Standards

     Reference standard samples of asbestos that can be used
     for quality control for a quantitative analytical method
     are not available.  It is, however, necessary for each
     laboratory to prepare at least two suspensions, one of
     chrysotile and another of a representative amphibole.
     These suspensions can then be used for intra-laboratory
     control and to furnish standard morphology photographs
     and diffraction patterns.

     7.1   Chrysotile Stock Suspension.

           Grind about 0.1 g of UICC chrysotile to a powder in
           ar. agate mortar.  Transfer 10 mg to a clean 1-liter
           volumetric flask, add several hundred ml of
           filtered distilled water containing 1 ml of a stock
           mercuric chloride solution and then make up to 1
           liter with filtered distilled water.  To prepare a
           working solution, transfer 10 ml of the above
           suspension to another 1-liter flask, add 1 ml of a
           stock mercuric chloride solution and make up tc 1
           liter with filtered distilled water.  This suspen-
           sion contains 100 yg per liter.  Finally transfer 1
           ml of this suspension to a 1-liter flask, add 1 ml
           of a stock mercuric chloride solution and make up
           to volume with filtered distilled water.  The final
           suspension will contain 5 to 10 MFL and :'s suitable
           for laboratory testing.

     7.2   Amphibole Stock Suspension.

           Prepare amphibole suspensions from UICC amphibole
           samples as in Section 7.1.

     7.3   Identification Standards.

           Prepare electron microscopic grids containing the
           UICC asbestos fibers according to Section 8 and
           obtain representative photographs of each fiber
           type and its diffraction pattern for future
           reference.

8.   Procedure

     8.1   Filtration.

           The separation of insoluble material, including
           asbestiform minerals, through filtration and subse-
           quent deposition on a membrane filter is a critical
           step in the procedure.  The objective of the
                              11

-------
filtration is not only to separate, but also to
distribute the particulate matter uniformly such
that discrete particles are deposited with a
minimum of overlap.

The volume filtered will range from 50 to 500 ml.
In an unknown sample, the volume can not be speci-
fied in advance because of the presence of variable
amounts of particulate matter.  In general, suffi-
cient sample is filtered such that a very faint
stain can be observed on the filter medium.  The
maximum loading that can be tolerated is 20 yg/cm2,
or about 200 yg on a 47-mm diameter filter; 5
pg/cm2 is near optimum.  If the total solids
content is known, an estimate of the maximum volume
tolerable can be obtained.  In a sample of high
solids content, where less than 50 ml is required,
the sample should be diluted with filtered
distilled water so that a minimum total of 50 ml of
water is filtered.  This step is necessary to allow
the insoluble material to deposit uniformly on the
filter.  The filtration funnel assembly must be
scrupulously cleaned before each filtration.  The
filtration should be carried out in a laminar flow
hood.

NOTE:  The following cleaning procedure has been
found to be satisfactory.

Wash each piece of glassware three times with
distilled water.  Following manufacturer's recom-
mendations, use the ultrasonic bath with an
Alcor.ox-water solution to clean all glassware.
After the ultrasonic cleaning, rinse each piece of
glassware three times with distilled watc»r.  Then
rinse each piece three times with deionized water
that has been filtered through O.l-ym Nuclepore
filtar.  Dry in an asbestos-free oven.  After the
glassware is dry, seal openings with parafilm.

        a.  Assemble the vacuum filtration
        apparatus incorporating the 0.1-ym Nucle-
        pore backed with 0.45-ym Millipore filter.
        See 8.3.2.

        b.  Vigorously agitate the water sample in
        its container.  Treatment of the sample in
        an ultrasonic bath may be required to
        evenly disperse the particulate material.
                   12

-------
              c.  If the required filtration volume can
              be estimated, either from turbidity esti-
              mates of suspended solids or previous
              experience, immediately withdraw the proper
              volume f.rom the container and add the
              entire volume to the 47-mm diameter
              funnel.  Apply vacuum sufficient for
              filtration but gentle enough to avoid the
              formation of a vortex.  If a completely
              unknown .sample is being analyzed, a
              slightly modified procedure must be
              followed.  Pour 500 nil of a well-mixed
              sample into a 500-ml graduated cylinder and
              immediately transfer the entire contents to
              the prepared vacuum filtration apparatus.
              Apply vacuum gently and continue suction
              until all of the water has passed through
              the filter.  If the resulting filter
              appears obviously coated or discolored,
              another filter should be prepared in the
              same manner, but this time using only 200
              or 100 ml of sample.

              NOTE 1:  Do not add more water after
              filtration has started and do not rinse the
              sides of the funnel.

              NOTE 2:  Nuclepore filter is basically a
              hydrophobia material.  The manufacturer
              applies a detergent, to the surface of the
              filter in order to render it hydrophilic;
              this process, nowever, does not appear to
              be entirely satisfactory in some batches.
              Pretreatment of the filter in a low temper-
              ature asher at 10 watts for 10 seconds can
              be used to render the surface of the filter
              hydrophilic.  This process will signifi-
              cantly decrease the islands of sparse
              deposit that are frequently observed.

              d.  Disassemble the funnel, remove the
              filter, and dry it in a covered petri dish.

8.2   Preparation of Electron Microscope Grids.

      Prsparation of the grid for examination in the
      microscope is a critical step in the analytical
      procedure.  The objective is to remove the organic
      filter material from the asbestos fibers with
      minimum loss and *'.ovement =>nd with minimum breakage
      of the grid support film.
                          13

-------
      If the sample contains large amounts of organic
      matter that interfere with fiber counting and
      identification a preliminary ashing step is
      required.  See 8.5.

8.3   Nuclepore Filter, Modified Jafie Wick Technique.

      8.3.1   rreparation of Modified Jaffe Washer

              Place three glass microscope slides (75 mm
              x 22 mm)  one on top of the other in a petri
              dish (100 mm x 15 mm) along a diameter.
              Place 14 S & S 1589 Black Ribbon filter
              papers; (9-cm circles) in the petri dish
              over \:he stack of microscope slides.  Place
              three copper mesh screen supports (12 mm x
              12 mm)  along the ridge formed by the ?,tack
              of slices underneath the layer of filter
              papers.   Place an EM specimen grid on each
              of the screen supports.  See Fig. 1.

              NOTE:   A stack of 30 to 40 S & S filters
              (7-cm circle), or equivalent, can be
              substituted for the 14 filters and micro-
              scope slides in preparing the Jaffe washer.

      8.3.2   Vacuum Filtration Unit

              Assemble the vacuum filtration unit.  Place
              a 0.45-um Millipore filter type HA on the
              glass frit ar.d then position a 0.1-ym
              Nuclepore filter, shiny side up, on top of
              the Millipore filter.  Apply suction to
              center  the filters flat on the frit.
              Attach  the filter funnel and shut off the
              suction.

      8.3.3   Sample  Filtration

              See 8.1.

      8.3.4   Sample  Drying

              Remove  the filter funnel and place the
              Nuclepore filter in .a loosely covered pe'cri
              dish to dry.  The petri dish containing the
              filter  may be placed in an asbestos-free
              oven at 45° C for 30 minutes to shorten
              the drying time.
                         14

-------
        Alternatively, the Nuclepore filter section
        may be mounted on a glass slide prior to
        drying the filter.

8.3.5   Selection of Section for Carbon Coating

        Using a small pair of scissors or sharp
        scalpel, cut out a rectangular section of
        the Nuclepore filter.  The minimum approxi-
        mate dimensions should be 15 mm long and 3
        mm wide.   Avoid selection near the perim-
        eter of the filtration a.vea.

3.3.6   Carbon Coating the Filter

        Tape the two ends of the selected filter
        section to a glass slide using double-sided
        tape.  Take care not to stretch the filter
        section.   Identify the filter section using
        a china marker on the slide.  Place the
        glass slide with the filter section into
        the vacuum evaporator.  Insert the necked
        carbon rod and, following manufacturer's
        instructions, obtain high vacuum.  Evapo-
        rate the neck, with the filter section
        rotating, at a distance of approximately
        7.5 cm from the filter section to obtain a
        30 to 50 nm layer of caroon on the filter
        paper.  Evaporate the carbon in several
        short bursts rather than continuously to
        prevent overheating the surface of the
        Nuclepore filter.

        NOTE 1:  Overheating the surface tends co
        crosslink the plastic, rendering :he filter
        dissolution in chloroform difficult.

        NOTE 2:  The thickness of the carbon film
        can be monitored by placing a drop of oil
        on a porcelain chip that is placed at the
        same distance from the carbon electrodes as
        the specimen.  Carbon is not visible in the
        region of the oil drop thereby enabling the
        visual estimate of the deposit thickness by
        the contrast differential.

8.3.7   Grid Transfer

        Remove the filter from the vacuum evapo-
        rator and cut out three sections somewhat
        less than 3 mm x 3 mm and such that the
                   15

-------
              square of Nuclepore fits within the circum-
              ference of the grid.  Pass each of the
              filter sections over a static eliminator
              and then place each of the three sections
              carbon-side down on separate specimen grids
              previously placed in the modified Jaffe
              Washer.  Using a microsyringe, place a
              10-ul drop of chloroform on each filter
              section resting on a grid and then saturate
              the filter pad until pooling of the solvent
              occurs below the ridge formed by the glass
              slides inserted under the layer of filter
              papers.  Place the cover on the petri dish
              and allow the grids to remain in the washer
              for approximately 24 hours.  Do not allow
              the chloroform to completely evaporate
              before the grids are removed.  To remove
              the grids from the washer, lift the screen
              support with the grid resting upon it and
              set this in a spot plate depression to
              allow evaporation of any solvent adhering
              to the grid.  The grid is now ready for
              analysis or storage.

8.4   Electron Microscopic Examination

      8.4.1   Microscope Alignment and Magnification
              Calibration

              Following the manufacturer's recommenda-
              tions carry out the necessary alignment
              procedures for optimum specimen examination
              in the electron microscope.  Calibrate the
              routinely used magnifications using a
              carbon grating replica.

              NOTE:  Screen magnification is not neces-
              sarily equivalent to plate magnification.

      8.4.2   Grid Preparation Acceptability

              After inserting the specimen into the
              microscope, adjust the magnification low
              enough (300X-10COX)  to permit viewing
              complete grid squares.  Inspect at least 10
              grid squares for fiber loading and distri-
              bution, debris contamination, and carbon
              film continuity.
                         16

-------
        Reject the grid for counting if:

        1)  The grid is too heavily loaded with
        fibers to perform accurate counting and
        diffraction operations.  A new sample
        preparation either from a smaller volume of
        water or from a dilution with filtered
        distilled water must then be prepared.

        2)  The fiber distribution is noticeably
        uneven.  A new sample preparation is
        required.

        3)  The debris contamination is too severe
        to perform accurate counting and diffrac-
        tion operations.  If the debris is largely
        organic the filter must be ashed and redis-
        persed (see 8.5).  If it is inorganic, the
        sample must be diluted and again prepared.

        4)  The majority of grid squares examined
        have broken carbon films.  A different grid
        preparation from the same initial filtra-
        tion must be substituted.

8.4.3   Procedure for Fiber Counting

        Two methods are conunonly used for fiber
        counting.  In one method (A), 100 fibers
        contained in randomly selected fields of
        view are counted.  The number of fields
        plus the area of a field of view must be
      .  known when using this irethod.  In the other
        method (B), all fibers (at least 100) in
        several grid squares or 20 grid .squares are
        counted.   The number of giid squares
        counted and the average area of one grid
        square must be known when using this method.

        NOTE:  The method co use depends upon the
        fiber loading on the grid and it is left to
        the judgment of the analyst to select the
        optimum method.  The following guidelines
        can be used:  If it is estimated that a
        grid square (80 ;im x 80 um)  contains 50 to
        100 fibers at a screen magnification of
        20000X, it is convenient to use the field-
        of-view counting method.   It' the estimate
        is less than 50, the grid square method of
        counting should be chosen.   On the other
                   17

-------
        hand, if the fiber count is estimated to be
        over 300 fibers per grid square/ a new grid
        containing fewer fibers must be prepared
        (through dilution or filtration of a
        smaller volume of water).

8.4.3A  Field-of-View Method

        After determining that a fibeL count can be
        obtained using this method, adjust the
        screen magnification to 15,000 to 20,OOOX.
        Select a number of grid squares that would
        be as representative as possible of the
        entice analyzable grid surface.  Prom each
        of these squares, select a sufficient
        number of fields of view for fiber
        counting.  The number of fields of view per
        grid square is dependent upon the fiber
        loading.  If more than one field of view
        per grid square is selected, scan the grid
        opening orthogonally in an arbitrary
        pattern that prevents overlapping of fields
        of view.  Carry out. the analysis by
        counting, measuring and identifying (see
        8.4.4) approximately 50 fibers on each of
        two grids.

        The following rules should be followed when
        using the field of view method of fiber
        counting.  Although these  rules were
        derived for a circular field of view they
        can be modified to apply to square or
        rectangular designs.

        1)  Count all fibers contained within the
        counting area and not touching the circum-
        ference of the circle.

        2)  Designate the upper right-hand quadrant
        as I and number in clockwise order.  Count
        all fibers touching or intersecting the arc
        of quadrants I or IV.  Do  not count fibers
        touching or intersecting the arc of quad-
        rants II or III.

        3)  If a fiber intersects  the arc of both
        quadrants III and IV or I  and II count it
        only if the greater length was outside the
        arc of quadrants IV and I, respectively.
                   18

-------
        4)   Count fibers intersecting the arc of
        both quadrants I and III but not those
        intersecting the arc of both II and IV.

        These rules are illustrated in Fig. 2.

8.4.3B  Grid Square Method

        After determining that a fiber count can be
        obtained using this method adjust the
        screen magnification to 15,OOC to 20,OOOX.
        Position the grid square so that scanning
        can be started at the left upper corner of
        the grid square.  While carefully examining
        the grid, scan left to right, parallel to
        the upper grid bar.  When the perimeter of
        the grid square is reached, adjust the
        field of view down one field width and scan
        in the opposite direction.  The tilting
        section of the fluorescent screen may be
        used conveniently as the field of view.
        Examine the square until all the area has
        been covered.  The analysis should be
        carried out by counting, measuring, and
        identifying  (see 8.4.4) approximately 50
        fibers on each of two grids or until 10
        grid squares on each of two grids have been
        counted.  Do not count fibers intersecting
        a grid bar.

8.4.4   Measurement and Identification

        Measure and record the length and width of
        each fiber having an aspect ratio greater
        than or equal to three.  Disregard obvious
        biological-bacteriological fibers and
        diatom fragments.  Examine the morphology
        of each fiber using optical viewing if
        necessary.  Tentatively identify, by
        reference to the UICC standards, chrysotile
        or possible amphibole asbestos.  Attempt to
        obtain a diffraction pattern of each fiber
        utilizing the shortest camera length
        possible.  Move the suspected fiber image
        to the center of the screen and insert a
        suitable selected area aperture into the
        electron beam so that the fiber image, or a
        portion of it, is in the illuminated area.
        The size of the aperture and the portion of
        the fiber should be such that particles
        other than the one to be examined are
                  19

-------
—  Counted

 -  Not  Counted
Figure 2.  Illustration of Counting Rules for
           Field of View Method.
                      20

-------
excluded from the selected area.  Observe
che diffraction pattern with the binocular
attachment.  It an incomplete diffraction
pattern is obtained,  move the particle
image around in the selected area to get a
clearer diffraction pattern or to eliminate
possible interferences from neighboring
particles.

Determine whether the fiber is chrysotile
or an araphiboie by comparing the diffrac-
tion pattern obtained to the diffraction
patterns of known standard asbestos
fibers,  Confirm the tentative identifica-
tion of chrysotile and amphibole asbestos
fi.om their electron diffraction patterns.
Classify each fiber as chrysotile, amphi-
bole, non-asbestos, no diffraction, oc
ambiguous.

NOTE 1:  It is convenient to use a tape
recorder during the examination cf the
fibers to record all pertinent data.  This
information can then be summarized on data
sheets or punched cards for subsequent
automatic data processing.

NOTE 2:  Chrysotile fibers occur as single
fibrils or in bundles.  The fibrills
generally show a tubular structure with a
hollow canal, although the absence of the
canal does not rule out its identifica-
tion.  Amphibole asbestos fibers usually
exhibit a lath-like structure with irreg-
ular ends, but occasionally will resemble
chrysotile in appearance.

NOTE 3:  The positive identification of
asbestos by electron diffraction requires
some judgment on the part of the analyst
because some fibers give only partial
patterns.  Chrysotile shows unique promi-
nent streaks on the layer lines nearest the
central one and a triple set of double
spots on the second layer line.  The
streaks and the set of double spots are the
distinguishing characteristics of chryso-
tile required for identification.  Amphi-
bole asbestos requires a more complete
diffraction pattern to be positively iden-
tified.  As a qualititative guideline,
           21

-------
              layer lines for amphibole, without the
              unique streaks  (some streaking may be
              present) of chrysotile, should be present
              and the arrangement of diffraction spots
              along the layer lines should be consistent
              with the amphibole pattern.  The pattern
              should be distinct enough to establish  ,
              these criteria.

              NOTE 4:  Chrysotile and thin amphibole
              fibers may undergo degradation in an elec-
              tron beam; this is particularly noticeable
              in small fibers.  It may exhibit a pattern
              for 1 to 2 seconds and disappear and the
              analyst must be alert to note the charac-
              teristic features.

              NOTE 5:  An ambiguous fiber is a fiber that
              gives a partial electron diffraction
              pattern resembling asbestos, but is insuf-
              ficient to provide positive identification.

      8.4.5   Determination of Grid Square Area

              Measure the dimensions of several represen-
              tative grid squares from each batch of
              grids with an optical microscope.  Calcu-
              late the average area of a grid square.
              This should be done to compensate for
              variability in grid square dimensions.
8.5   Ashing
      Some samples contain sufficiently high levels of
      organic material that an ashing step is required
      before fiber identification and counting can be
      carried out.

      Place the dried Nuclepore filter paper containing
      the collected sediment into a glass vial (28 mm
      diameter x 80 mm high).  Position the filter such
      that the filtration side touches the glass wall.
      Place the vial in an upright: position in the low
      temperature asher.  Operate the asher at 50 watts
      (13.56 MHz) power and 2 torr oxygen pressure.  Ash
      the filter until a thin film of white ash remains.
      The time required is generally 6 to 8 hours.  Allow
      the ashing chamber to slowly reach atmospheric
      pressure and remove the vial.  Add 10 ml of
      filtered distilled water to the vial.  Place the
      vial in an ultrasonic bath for 30 minutes to
      disperse the ash.  Dilute the sample if required.

                         22

-------
      Assemble the 25-mm diameter filtering apparatus.
      Center a 25-rnrr. diameter 0.1-um Nuclepore filter
      (with the 0.45-um Millipore backing) on the glass
      frit.  Apply suction and recenter the filter if
      necessary.  Attach the filter funnel and turn off
      the suction.  Add the wator containing the
      dispersed ash from the vial to the filter funnel.
      Apply suction and filter the sample.  After drying,
      this filter is ready to be used in preparing samrle
      gr ids as in 8.3.

      NOTE 1:   In specifying a 25-mm diameter filter it
      is assumed that the ashing step is necessary mainly
      because of the presence of organic material and
      that the smaller filtering area is desirable from
      the point of view of concentrating the fibers.  If
      the sample contains mostly inorganic debris such
      that the smaller filtering area will result in
      overloading the filter, the 47-ram diameter filter
      should be used.

      NOTE 2:   It will be noted that a 10-.ml volume is
      filtered in this case instead o£ the minimum 50-ml
      volume specified in 8.1.  These volumes are consis-
      tent when it is considered that there is approxi-
      mately a 5-fold difference in effective filtration
      area between the 25-mm diameter and 47-mm diameter
      filters.

      NOTE 3:   Cross contamination is probable when
      ashing more than one sample at a time.

8.6   Determination of Blank Level

      Carry out a blank determination with each batch of
      samples prepared, but a minimum of one per week.
      Filter a fresh supply (500 ml) of distilled,
      deionized water through a clean 0.1-um membrane
      filter.   Filter 200 ml of this water through a
      0.1-um Nuclepore filter, prepare the electron
      microscope grid, and count exactly as in the proce-
      dures 3.1 - 8.4.  Examine 20 grid squares and
      record this number of fibers.  A maximum of two
      fibers in 20 grid squares is acceptable for the
      blank sample.

      NOTE:  Monitoring the background level of asbestos
      is an integral part of the procedure.  Upon initia-
      ting asbestos analytical work, blank samples must
      be run to establish the initial suitability of the
      laboratory environment, cleaning procedures, and
                         23

-------
      reagents for carrying out asbestos analyses.
      Analytical determinations of asbestos can  be
      carried out only aft^r an acceptably low level of
      contamination has been established.

Calculations

9.1   Fiber Concsntrfitions

      Gr id Square Counting Method - If the Grid  Square
      Method of counting is employed, use the following
      formula to calculate the total asbestos fiber
      concentration in MFL.

              C = (F x Af)/(Ag x V0 x 1000)       (1)

      where:  C = Fiber concentration (MFL)

              F = Average number of fibers per grid
                  opening

              A,  = Effective filtration area of  filter
                   paper (mm2) used in grid preparation
                   for fiber counting

              A  = Average area of one grid square  (mm2)

              VQ » Original volume of sample filtered  (ml)

      If ashing is involved, use the same formula but
      substitute the effective filtration area of the
      25-nun diameter filter for Af instead of that  for
      the 47-mm diameter filter.  If one-half the filter
      is ashed, multiple C by two.

      Field-of-View Counting Method - If the Field-of-
      View Me-thod of counting is employed, use the
      following formula to calculate the total asbestos
      fiber concentrations  (MFL}.

              C = (F x Af x 1000)/(AV x V0)       (2)

      where:  C = Fiber concentration

              F = Average number of fibers per field of
                  vii»w

              AJ  = Effective filtration ar^a of  filter
                   paper (mm2) used in grid preparation
                   for fiber counting
                         24

-------
              Ay = Area of one field of  view  (ym2)

              V0 = Original volume of sample  filtered  (ml)

      If ashing is involved, use the same  formula  but
      substitute the effective filtration  area of  the
      25-mm diameter filter for Ar instead of that  for
      the 47-mir diameter filter.

9.2   Estimated Mass Concentration

      Calculate ".he mass (pg) of each fiber counted  using
      the following formula.

              M = L x W2 X D x 1CT6

      If the fiber content is predominantly chrysotile,
      the following formula may be used.

              M=lxLxW2xDx i.0~6             (3)
                  4
      where:  M = Mans  (ug)                  '

              L = Length (um)

              W = Width (MTU)

              D = Density of fibers  (g/cm^)

      Then calculate the mass concentration  (yg/1)
      employing the following formula.

              Mc = C x Mf x 106

      where:  M  = mass concentration (yg/1)

              C  =» fiber concentration  (MFL)

              Mf = mean mass per fiber  (ug)

      To calculate Mf use the following  formula.



              Mf -  1   MVn                        (4)
      where:  M^= mass of each fiber, respectively

              n = number of fibers counted
                          25

-------
           NOTE 1:  Because many of the amphibole fibers are
           lath shaped rather than square in cross section the
           computed mass will tend to be high because laths
           will, in general, tend to lie flat rather than on
           edge.

           NOTE 2:  Assume the following densities:  chryso-
           tile 2.5, amphibole 3.25.

     9.3   Aspect Ratio

           The aspect ratio for each fiber is calculated by
           dividing the length by the width.

10.  Reporting

     10.1  Report the following concentration as MFL for
           sample and blank using 95% confidence intervals.

           a.  Chrysotile

           b.  Amphibole

           c.  Total asbestos fibers

     10.2  Use two significant figures for concentrations
           greater than 1 MFL, and one significant figure for
           concentrations less ihan 1 MFL.

     10.3  Tabulate the size distribution, length and width.

     10.4  Tabulate the aspect ratio distribution.

     10.5  Report the calculated mass as ug/1.

     10.6  Indicate the detection limit in MFL.

     10.7  Indicate if less than five fibers were counted.

     10.8  Include remarks concerning pertinent observations,
           (clumping, amount of organic matter,, debris)  amount
           of suspected though not identifiable as asbestos
           fibers (ambiguous).

11.  Precision

     11.1  Intra-Laboratory

           The precision that is obtained within an individual
           laboratory is dependent upon the number of fibers
           counted.   If 100 fibers are counted and the loading
                              26

-------
           is at least 3.5 fibers/grid square, -computer
           modeling of the counting procedure shows that a
           relative standard deviation of about 10% can be
           expected.

           In actual practice some degradation from this
           precision will be observed but should not exceed ^
           15% if several grids a-'e prepared from the same
           filtered sample.  The relative standard deviation
           of analyses of the same water sample in the same
           laboratory will increase as a result of sample
           preparation errors and a relative standard devia-
           tion of about about + 25 to 35% will occur.  As the
           number of fibers counted decreases, the precision
           will also decrease approximately proportional to /N
           where N is the number of fibers counted.  The
           precision for mass concentration is generally
           poorer than that for fiber concentration.

           Based upon the analysis of one laboratory utilizing
           a different analyst for each of three water
           samples, intra-laboratory precision data are
           presented in Table 1.

     11.2  Inter-Laboratory

           Based upon the analysis by various government and
           private industry laboratories cf filters prepared
           from nine water samples, inter-laboratory precision
           data of the method are presented in Table ^.

12.   Accuracy

     12.1  Fiber Concentrations

           As no standard reference materials are available,
           only aoproximate estimates of the accuracy of the
           procedure can be made.  At 1 M?L, it is estimated
           that the results should be within a factor of 10 of
           the actual asbestos fiber content.

           This method requires the positive identification of
           a fiber to be asbestos as a means for its quantita-
           tive determination.  As the state-of-the-art
           precludes the positive identification of all of the
           asbestos fibers present, the results of this
           method, as expressed as MFL, will be biased on the
           low side and, assuming no fiber loss, ^represent 0.4
           to 0.8 of the total asbestos fibers present.
                              27

-------
                              TABLE  1.   INTRA-LABORATORY PRECISION
NJ
CO

Sample Number of Mean Fiber Precision, Mass Precision,
Type Sample Concentration Relative Concentration Relative
Aliquots MFL (millions of Standard (py/1) Standard
Analyzed asbestos fibers/1) Deviation Deviation
Chrysotile
(UICC)
Crocidolite
(UICC)
Taconite
(raw water)
26 23
2C 8
2U 16
37%
36%
24%
0.32 71%
1.5 48%
10.5 65%

TABLE 2. INTER-LABOPATORY PRECISION




Sample Number of
Type Labs
Reportiny
Chrysotile 1C
9
11
9
9
3
Amphibole 11
4
14
Mean Fiber
Concentration,
MFL (millions of
asbestos fibers/1)
877
119
59
31
28
25
139
95
36
Precision,
Relative
Standard
Deviation
35%
43%
41%
65%
32%
35%
50%
52%
66%

-------
     12.2  Mass Concentrations

           As in the case of the fiber cone^ntirat ions,  no
           standard samples of the size distribution found in
           water are available.  The estimated mass concentra-
           tion is oft^n very inaccur~.-~ because or pc-cr
           counting statistics assoo^ai-ed with larce r.bers
           that are few in number out represent most or the
           actual mass concentration.

13.   Suggested Statistical Evaluation of Grid Fiber Counts

     13.1  Because the fiber distribution on the sample
           filter, resulting from the method of filtration,
           has not been fully characterized, the fiber distri-
           bution obtained on the electron microscope grids
           for each sample should be tested statistically
           against an assumed distribution and a measure of
           the precision of the analysis should be provided.

     13.2  Assume that the fibers are uniformly and randomly
           distributed on the sample filter and grids.   One
           method for confirming this assumption is given
           below.

           Using the chi-square test, determine whether the
           total number of fibers found in individual grid
           openings are randomly and uniformly distributed
           among the openings using the following formula.
                       (n.-np.)

                                                      (5)
                         np^
           where:  X  = Chi-square statistic

                   N  = Number of grid openings examined for
                        the sample

                   n- = Total number of fibers found in each
                        respective grid opening

                   n  = Total number of fibers found in N grid
                        openings

                   p- = Ratio of the area of each respective
                        grid opening to the sum of the areas
                        of all grid openings examined
                               29

-------
13.3
NOTE:  If an average area for the gri^ squares has
been measured as outlined in 8.4.5, the tarm np^
represents the mean fiber count per grid square.

If the value for X2 exceeds the value listed in
statistical tables for the 0.1% significance level
with N-l degrees of freedom, the fibers are not
considered to be uniformly and randomly distributed
among the grid openings.  In this case, it is
advisable to try to improve the uniformity of fiber
deposition by filtering another aliquot of the
sampla and repeating the analysis.

If uniformity and randomness of fiber deposition on
the microscope grids has been demonstrated as in
13.2, and the fiber concentration is assumed to be
normally distributed aocut the mean value, the 95%
confidence interval about the mean fiber concentra-
tions for chrysotile, amphibole, and total asbestos
fibers may be determined using the following
formulae.
              s  '
                        N

                      N I
                           N(N-l)
                                            1/2
                                                    [6)
      where:   SG = Standard deviation of the chrysotile
                   fiber count

              N  = Number of grid openings examined for
                   the sample

              X^ = Number of chrysotile fibers in each
                   grid opening, respectively

      Obtain  the standard deviations of the fiber counts
      for  amphibole asbestos fibers and for total
      asbestos fibers by substituting the corresponding
      value of X  into equation (6).
                 X  = X +
                  u
                          tS
                                              7)
                      -   tSc
                    X— V —	
                  —  ~~ f\    __
                  L        /N
                                             (8)
                         30

-------
where:  Xu = Upper value of 95% confidence interval
             for chrysotile

        XL = Lower value of 95% confidence interval
             for chrysotile

        X  = Average number of fibers per grid
             opening

        t  = Value listed in t-distribution tables
             at the 95% confidence level for a two
             tailed distribution with N-l degree of
             freedom

        Sc - Standard deviation of the fiber counts
             for chrysotile

        N  = Number of grid openings examined for
             the sample

The values of Xu and XT can be converted to concen-
trations in millions of fibers per liter using the
formula in section 9 and substituting either Xu or
XL for the term F.

Obtain the upper and lower values of the 95% confi-
dence interval for amphibole asbestos fibers and
total asbestos fibers by substituting the corres-
ponding values of X and S into equations (7) and
(8).

Report the precision of the analysis, in terms r>f
the upper and lower limits of the 95% confidence
interval, for chrysotile, amphibole, and total
asbestos fiber content.  If a lower limit is found
to be negative, report the value of the limit as
zero.
                    31

-------
                     SELECTED BIBLIOGRAPHY
                                       '  «^

Beaman, D. R. and D. M. File.  Quantitative Determination of
Asbestos Fiber Concentrations.  Anal.  Cnem. 48(1):  101-110,
1976.

Lishka, R. J., J. R. Millette, and E.  F. McFarren.  Asbestos
Analysis by Electron Microscope.  Proc. AWWA Water  Quality
Tech. Conf. American Water Works Assoc./ Denver, Colorado
XIV-1 - XIV-12, 1975.

Millette. J. R. and E. F. McFarren.  EDS of Waterborne
Asbestos Fibers in TE.M, SEM and STEM.   Scanning Electron
Microscopy/1976 (Part III) 451-460, 1976.

Cook, P. M., I. B. Rubin, C. J. Maggiore, and W. J.
Nicholson.  X-ray Diffraction and Electron Beam Analysis of
Asbestiform Minerals in Lake Superior  Waters.  Proc.  Inter.
Conf. on Environ. Sensing and Assessment 34(2): 1-9,  1976.

McCrone, W. C. and I. M. Stewart.  Asbestos.  Amer. Lab. 6(4):
10-18, 1974.

Mueller, P. K., A. E. Alcocer, R. L. Stanley, and G.  R.
Smith.  Asbestos Fiber Atlas.  U.S. Environmental Protection
Agency, Research Triangle Park, North  Carolina, Technology
Series, EPA 65U/2-75-036, 1975.

Glass, R. W.  Improved Methodology for  Determination  of
Asbestos as a Water Pollutant.  Ontario Research Foundation
Report, April 30, 1976,  Mississauga,  Ontario, Canada.

Saroudra, A. V.  Optimum Procedure for  Asbestos Fibers Identi-
fication from Selection Area Electron  Diffraction Patterns in
a Modern Analytical Electron Microscope Using Tilted  Speci-
mens.  Scanning Electron Microscopy, Vol. I, proceedings of
the Workshop on Analytical Electron Microscopy, March, 1977.
Chicago, Illinois.

Chatfield, E. J., R. W. Glass, and M.  J. Dillon.  Preparation
of Water Samplss for Asbestos Fiber Counting by Electron
Microscopy.  U.S. Environmental Protection Agency,  Athens,
Georgia.  EPA-600/4-78-011, January 1978.
                              32

-------
Asher, I. M. and, P. P. McGrath.  Symposium on Electron
Microscopy of Microf iberi,.  Proceedings of the First PDA
Office of Science Summer Symposium, August 1976.  Pennsylvania
State University.

Chopra, K. S.  "Interlaboratory Measurements of Amphibole and
Chrysotile Fiber Concentrations in Water."  Journal of Testing
and Evaluation, JTEVA, Vol. 6, No. 4, July 1978, pp. 241-247.

Chatfield, E. J.  Preparation and Analysis of Particulate
Samples by Electron Microscopy, with Special Reference to
Asbestos.  Scanning Electron Microscopy, April 16-20, 1979,
Washington, DC.

National Bureau of Standards Special Publication 506.
Proceedings of the Workshop on Asbestos:  Definitions and
Measurement Methods held at NBS, Gaitnersbarg, MD, July 18-20,
1977.  (Issued March 1978).
                              33

-------


-------