EPA Report No.
                                                        July  1984
METHODOLOGY FOR THE MEASUREMENT OP AIRBORNE ASBESTOS
               BY ELECTRON MICROSCOPY
                         by

                    George  Yamate
                  Satish C. Agarwal
                  Robert D. Gibbons

               III Research Institute
              Chicago,  Illinois  60616
               Contract No. 68-02-3266
                   Project Officer

                  Michael  E.  Beard
              Quality  Assurance  Division
     Environmental Monitoring Systems Laboratory
         U.S.  Environmental Protection  Agency
    Research Triangle Park, North Carolina   27711
     ENVIRONMENTAL MONITORING SYSTEMS LABORATORY-
          OFFICE .OF .RESEARCH. AND DEVELOPMENT
         U.S.  ENVIRONMENTAL PROTECTION AGENCY
    RESEARCH TRIANGLE PARK, NORTH CAROLINA   27711

-------
                                  DISCLAIMER

     This report has been reviewed by the Environmental Monitoring  Systems
Laboratory, U.S. Environmental Protection Agency, and approved  for  publica-
tion.  Approval does not signify that the contents necessarily  reflect  the
views and policies of the U.S. Environmental Protection Agency, nor  does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
                                       ii

-------
FOREWORD
    iii

-------
PREFACE
    iv

-------
                                   ABSTRACT

     The provisional electron microscope methodology for measuring the
concentration of airborne asbestos fibers was refined.  The methodology is
divided into separate protocols.  The step-by-step procedures for each
protocol are nearly identical, so that cumulative data can be obtained and
uncertainties, especially in asbestos identification, can be clarified.  The
operational steps encompass (1) type of sample, (2) collection and transport,
(3) sample preparation, (4) examination under the transmission electron
microscope (TEM) and data collection, (5) data reduction and reporting of
results, and (6) quality control-quality assurance.

     The TEM analytical protocol is subdivided into three levels of anal-
ysis:  Level I, for screening many samples; Level II, for regulatory action;
and Level III, for confirmatory analysis of controversial samples.  Because
identification of asbestos structures is critical, the level of analysis is
directly related to the information sought:

       Level I—morphology and visual selected area electron
                diffraction (SAED) pattern recognition.
      Level II—morphology; visual SAED; and elemental analysis.

     Level III—morphology; visual SAED; a selected number of SAED
                micrographs of zone-axis patterns; and elemental
               .analysis.

     This report was submitted in fulfillment of Contract No. 68-02-3266 by
IIT Research Institute under the sponsorship of the U.S. Environmental
Protection Agency.  This report covers the period September  19,  1979, to
June 19, 1981, and work was completed as of September 30, 1981.

-------
                                   CONTENTS
Disclaimer	ii
Foreword	iii
Preface	iv
Abstract	v
Tables	ix
Figures	x
List of Abbreviations	xi

1.  Introduction	1

2.  Conclusions and Recommendations	3

3.  Guidelines for Understanding the Methodology	4

    Level of Analysis	5
    Order of Analysis	5
    Collection and Reporting	5
    Costs	5
    Application to Nonairborne  Sources	6
    Geographical Considerations	6
    Laboratory Conditions	6

4.  Level I Analysis	.	8

    Summary of Protocol	8
    Equipment, Facilities, and  Supplies	8
    Description of Methodology	9

       1.  Type of Samples—Source	9
       2.  Sample Collection and Transport	10
       3.  Sample Preparation  for Analysis—Grid  Transfer	13
       4.  TEM Examination  and Data Collection	16
       5.  Data Reduction  and  Reporting  of  Results	,21
       6.  Quality Control/Quality Assurance	22

5.  Level II Analysis	24

    Summary of Protocol	24
    Equipment, Facilities, and  Supplies	25
    Description of Methodology	25

       1.  Type of Samples—Source	25
       2.  Sample Collection and Transport	26
       3.  Sample Preparation  for Analysis—Grid  Transfer	29


                                       vi

-------
                             CONTENTS (continued)

                                                                           Page

      4.  TEM Examination and Data Collection	33
      5.  Data Reduction and Reporting of Results	39
      6.  Quality Control/Quality Assurance	42

6.  Level III Asbestos Analysis	44

    Discussion of Protocol	44
    Summary of Protocol	46
    Equipment, Facilities, and Supplies	47
    Description of Methodology	48
      1.  Crystallography and Morphological Properties	48
      2.  Chemical Properties—Elemental-Analysis  by  EDS	49
      3.  Selected Area Electron Diffraction  (SAED)	49
      4.  Use of Tilting to Acquire Exact Zone-Axis SAED  Patterns	50
      5.  Characteristics of SAED Patterns  Encountered	51
      6.  Determination of Camera Constant  and  SAED Pattern  Analysis	53
      7.  Determination of Camera Constant  Using Gold Rings	54
      8.  Measurement of d-Spacings and  Interplanar Angles	55
      9.  Identification of Unknown Fibers	58

7.  Archival Samples	60

    Discussion of Protocol	60
    Description of Methodology	61

      1.  Samples with Adequate Loading	61
      2.  Samples with Heavy Loading	62

8.  Bulk-Sample* Analysis	65
    Discussion of Protocol	65
    Description of Methodology	65

     * 1.  Polarized Light Microscopy	65
      2.  X-Ray Diffraction Analysis	66
      3.  Electron Microscopy	.•	66

9.  Numerical Relationships and Analytical  Aids	67
                                  1
    Limits of Detection	67
    Statistical Methodology	68

      1.  95% Confidence Limits for  a  Poisson Variate	70
      2.  Comparison  of  Two Poisson  Variates	70
    Magnification Calibration	73
    Preparation of Blanks	74
    Use of Computers	74

References	75
                                       vii

-------
                             CONTENTS (continued)

                                                                           Page
Appendices
    A.  Figures	76
    B.  Computer Printout of Level I Analysis (Example)	96
    C.  Computer Printout of Level II Analysis (Example)	99
                                      viii

-------
                                    TABLES
Number                                                                     Page
  1     Profile Comparison of Asbestos  Standards	40
  2     Determination of Camera Constant  (Example)	55
  3     Determination of Spot Spacings  (Examples)	57
  4     Comparison of d-spacings'  from SAED  File  and
        Powder Diffraction File (Example)	59
  5     Hypothetical Data	69
  6     95 Percent Confidence Limits	71
  7     Control and Test Sample Differences	72
                                       ix

-------
                                    FIGURES
Number
 Al  Vacuum evaporator	76
 A2  Multiple coating arrangement in evaporator	77
 A3  Close-up of multiple-coating arrangement	78
 A4  Modified Jaffe wick washer method (sketch)	79
 A5  Modified Jaffe wick washer	80
 A6  Transmission electron microscope	80
 A7  Morphology and counting guidelines used 1
     in determining asbestos structure	81
 A8  Level I data sheet Cexample)	82
 A9  Scanning of full-grid opening...	83
A10  Transmission electron microscope with energy dispersive  spectrometer...84
     iff
All  Spectra profiles of asbestos standards	85
A12  Level II data sheet (example)	86
A13  EM data report (example)	87
                •*'
A14  Sample summary report (example)	88
A15  Effects of tilting and alignment of  fiber	89
A16  Method of measuring two perpendicular diameters  for each ring	89
A17  Method of recognizing a horizontal row of spots	90
A18  Relationship of di , 62 , 6l.2>  and R	91
A19  Typical zone-axis  SAED patterns from araosite standard  specimen	92
                                  . ••
A20  Typical zone-axis  patterns  from crocidolite standard  specimen	93
A21  Typical zone-axis  patterns  from treraolite standard specimen	94
A22  Typical SAED patterns and EDAX spectra  from anthophyllite
     standard specimen	95

-------
                             LIST OF ABBREVIATIONS
AEM     analytical electron microscope
EDS.    energy dispersive spectrometer
EM      electron microscope
JCPDS   Joint Committee on Powder Diffraction Standards
LTA     low-temperature ashing
NIOSH   National Institute of Occupational Safety and Health
PLM     polarized light microscopy
QC/QA   quality control/quality assurance
SAED    selected area electron diffraction
SEM     scanning electron microscope
STEM    scanning transmission electron microscope
TEM     transmission electron microscope
TSP     total suspended particulates
U1CC    Union Internationale Contre  le Cancer
XRD     x-ray diffraction
XRF     x-ray fluorescence
                                       xi

-------
                                   SECTION  1

                                  INTRODUCTION

      Asbestos Is recognized as a health hazard, especially if inspired into
 the alveolar region of the respiratory tract.  Asbestos may be present in air
 samples,  water samples, biological or clinical samples, and other miscel-
 laneous bulk samples, such as ores and food.  These various types of samples
 require different collection methodologies and diverse preparation techniques.

      Asbestos analysis methodologies may be categorized as bulk-material
 analyses, or those providing concentration information, and single-fiber
 analyses, or those providing morphology, size distribution, and concentra-
 tion.  BuJLJk-material_analy.sis_ techniques, which include infrared spectroscopy,"
("differential thermal analysis, and x-ray diffraction analysis (XRD), are
ilimited by an inability to analyze concentrations of less than 1 ug» and by an
'inability to differentiate between fibrous and nonfibrous forms of minerals.
i                                                                                i

      Sing_le^fiber.,anal.ysis. techniques include optical microscopy and electron
 microscopy.  Optical microscopy employing phase contrast has been promulgated
 injto_a raonitorpTg~*inethod for the workplace environment "(NIOSH-rP&CAM 239).  In
 addition^promulgation of a monitoring method for bulk-material asbestos
 samples (building insulation) using polarized light, microscopy (PLM) is
 presently being considered.  However, optical microscopic techniques cannot;
[determine fibers of less than approximately  1 ym in diameter, and phase
;contrast cannot differentiate between asbestos and nonasbestos fibers.     :

      The electron microscope (EM) provides particle morphology and size, and a
 degree of identification.  A comprehensive study of various EM procedures
 (Samudra et al., 1977) was conducted in development of a provisional method-
 ology manual, Electron Microscope Measurement of Airborne Asbestos Concentra-^
 tions (Samudra, 1978).  Three EM methods are available:  the scanning electron
 microscope (SEM), the transmissio.ix electron microscope (TEM), and the analyti-
 cal electron microscope (AEM).  The SEM, with an x-ray energy-dispersive
 spectrometer (EDS), permits visual characterization (analogous to reflection
 optical microscopy) and fiber identification by elemental analysis.  The TEM,
 providing an increased data-acquisition capability, permits visual characteri-
 zation (in the transmitted mode) and fiber identification by crystal structure
 analysis.  The AEM is a TEM with an EDS, and with the added capability of
 SEM/STEM (scanning transmission electron microscope) operation, which permits
 visual characterization (morphology and size) as well as fiber identification
 using both crystal structure by selected area electron diffraction  (SAED)  and
 elemental analysis by EDS.

-------
j"     The original EM methodology was developed for the U.S. Environmental /
iProtection Agency (EPA) for measuring airborne asbestos concentrations,   j
ispecifically for ambient air and for use as a "screening" tool.  Development
 guidelines included attainable precision and accuracy of results; relative
 rapidness in use; cost-effectiveness; applicability to a large number of
 laboratories possessing a TEM (at that time, very few laboratories had TEM's
 with x-ray analysis capability or an AEM); and procedural steps to be inde-
 pendent of unique or exceptional in-house capabilities of a single laboratory
 (that is, interlaboratory precision rather than intralaboratory).

      In usage, the EM method was successful within its prescribed limita-
 tions—that is, the precision and accuracy of results between laboratories
 using the complete method was good.  However, problems __that had been recogr
 nized in the study developing the methodology (Samudra et at., .1977) arose in
 "the areas of (1) interpretation of airborne, (2) sample collection, (3) need
 for more exacting identification of asbestos, especially of amphibole type,
 and (4) use of only part of the methodology.

      The,.pr.esent .study..was .undertaken to.refine .the..methodology..  The problem
 areas and related criticisms were addressed within the underlying goals and
 guidelines set for optimizing the methodology.  Protocols similar to a cook-
 book were not possible since basic knowledge or training was required regard-
 ing (1) sample collection, (2) preparation of samples  for EM, (3) use of  the
 TEM-AEM, and (4) diffraction pattern analysis.  The refined methodology is
 based on an assumption that each intended user of a particular level of analy-
 sis has the necessary background and training to use it.

-------
                                   SECTION 2

                        CONCLUSIONS AND RECOMMENDATIONS
      The  EM methodology for  measuring the concentration of airborne asbestos
 fibers  has been refined and  specified, and is recommended for field evalua-  ._
,tion.   The methodology is based on a TEM analytical protocol that is divided  1
jinto  three levels of effort:  Level I, for screening many samples; Level II,  j
;,for regulatory action; and Level III, for confirmatory analysis of controver-.J
Isial  samples.   The three-level analytical methodology is cost-effective, and
 will  provide the required results for proper assessment of asbestos.

-------
                                   SECTION 3

                  GUIDELINES  POR UNDERSTANDING THE METHODOLOGY
      The methodology  is  divided  into  separate  protocols.   The step-by-step
procedures  for  each protocol  are nearly  identical,  so  that cumulative data can
be obtained  and uncertainties, especially  in asbestos  identification, can be
clarified.   These  operational steps are:

      (1)  Type  of  Sample—Source

      (2)  Sample Collection and  Transport

      (3)  Sample Preparation  for Analysis—Grid  Transfer

      (4)  TEM Examination and Data Collection

      (5)  Data  Reduction and  Reporting of  Results

      (6)  Quality  Control/Quality Assurance  (QC/QA).

      The analytical, protocol  under the TEM examination and data collection
procedure is subdivided  into  three levels  of  increasing analytical effort in
terms of requiring an instrument of greater  capability, an electron
tnicroscopist with  greater expertise,  and a longer  analytical time.  Level I, a
monitoring  or screening  methodology,  resembles the  present EPA provisional"
methodology (Samudra  et  al.,  1978; Anderson  and  Long,  1980).  Ley.e.l_.II, is a
regulatory  method  requiring additional analytical  criteria to establish
asbestos identification  limits,  and to provide guidance for Level I analyses
by confirming or clarifying visual SAED  patterns,   ^yel  IIIi the most
sophisticated and  the costliest  of the methods,  is  intended for confirming  ,
asbestos identification, especially in judicial  controversies and other
special situations.
                                                                             x ••
      In Sections 4,  5, and  6, the protocols  for each of the three levels of
.analysis are presented independently  of  each other, and thus procedures common
to each are repeated. All  figures are presented in Appendix A.

      Section 7  describes modifications for using the methodology on archival
samples, which  are samples  collected  on  nonprocedural filter substrates, or
samples collected  without regard to  filter loading levels.  Section 8
describes analysis of inorganic  sources  in bulk-air samples or in bulk form.
Section 9 concludes the  report with a discussion of analytical aids pertaining
to the limits of detection,  preparation  of blanks,  use of computers,
magnification calibration,  and statistical methodology.

-------
     General guidelines for understanding the methodology are discussed in
the following paragraphs.

LEVEL OP ANALYSIS

     Knowledge of the history, source, and location of the sample, and the
purpose and objective of the analysis aids in selecting the correct level of
analytical effort.  Simply "grinding the samples out" neither is cost-
effective nor produces the best results, especially for Level II and Level  III
analyses.  Instead of all Level I, all Level II, or all Level III, the
majority of the analyses may be Level I, followed by some Level II.  Level  III
could be used in its entirety or only at the analytical phase.  If the source
is known to contain no amphlbole-type interference, or if chrysotile is of
interest, gold-coating can be eliminated.

     If a legal proceeding is anticipated, Level III analysis will be required
where a chain-of-custody record is kept from collection, transport to the
laboratory, preparation, analysis, data reduction, and reporting of results.
EM finder grids oust be used for grid transfer.  In addition, for quality
assurance, a second laboratory must be available for analyzing a portion of
the sample using the same degree of custodial care.  QC/QA protocols must be
observed and records kept.

     Whenever possible, and especially for unknown source samples, 10 to 20%
of each set of samples should be analyzed by Level II analysis prior to using
Level I as a screening procedure.

     Level I is a relatively rapid procedure, and can be used by many
laboratories with access to a conventional TEM.  However, Level I results
should net be used in legal proceedings.  If "positives" or  "false positives"
are found, especially in areas where asbestos is known to be absent, and the
field blank and laboratory blank have been checked, Level II analysis, and
possibly Level III analysis, should be performed.

ORDER OF ANALYSIS                                                              ••'-

 .   . The -order of analysis is  (1) field  blanks,  (2) laboratory blanks  (if     .--_
needed), and (3) field samples.

COLLECTION AND REPORTING

     The counting rule,  "minimum  100  fibrous structures  per  known  area
(complete grid opening)  or  10  grid openings, whichever  is first,"  is a minimum
rule for cost  limitation.  For very  low asbestos presence, or  for  asbestos
contamination  studies, where  particulate  loading is high and  asbestos  presence
very low, counting  20  grid openings  from each of 2 grids (10 per  grid)  is
recommended.

     The EM  magnification  factor  is  very high or,  conversely,  the  area  of        '
deposit  examined  is  very small.   Therefore,  although  the electron  microscopist '
may  report  a zero count,  the  notation  "Below Detectable  Level"  is  more  appro-._$
priate  in  the  sample  report.   Along  the  same  lines,  the  electron  microscopist

-------
 should  report  observations,  measurements,  and  conclusions as objectively as
 possible,  realizing  the  subjective  nature  of his decision-making,  such as
 parallel-sided,  3:1  aspect  ratio, number count,  size measurements,
 recognition-discrimination  of  SAED  patterns, and categorizing of asbestos
 structure.

      Data  reduction  and  reporting of results must be consistent and stated.
 Dimensions  of  X-fibers  (unknown length since complete fiber is not visible)
 may  be  doubled,  not  counted at all, or presented separately.  Doubling of the
 visible portion  is  recommended, and should be  so stated in the report.

      Mass  or  conversion  of  size measurements to  an assumed shape-volume-
 density relationship,  is calculated, and thus  is the least reliable of the
 data, especially for X-fibers, bundles, clusters, and matrices.

      Although morphology, SAED, and XRF either singly or in combination will
 provide identification  of asbestos, not all structures will be identified.
 The  nature  of  the asbestos  structure prevents  analysis of all structures by
 SAED and/or by elemental analysis with EDS.  Such factors as specimen thick-
 ness, orientation,  and  proximity to other particulates or to the grid wire
 will prevent  attainment  of  good SAED patterns  and limit the effectiveness of
 chemical analysis.

 COSTS

      Levels I, II,  and  III  analyses are estimated to require 200, 400, and
 1200 min per  analysis,  respectively.  Additional costs will result from
 collection, preparation, and reporting of results.  The equivalent monetary
 costs will depend on the laboratory rates of  the personnel involved.

 APPLICATION TO -NONAIRBORNE SOURCES

f~    Although the methodology has been developed for airborne asbestos, other~~i
\types of samples from different sources can be analyzed if the samples are
\ finely  divided and placed with proper loading  and uniform distribution either_
ion .a.polycarbonate membrane filter  or on a carbon-coated EM grid.  Of course,
 the  limitations of the collection and preparation steps must be known and  ..
 accounted  for to prevent inaccuracies in comparing results.

 GEOGRAPHICAL CONSIDERATIONS

      In some parts of the country,  such as the-Upper.Great Lakes  area,  the
 possibility of misidentification is much greater.because some.nonamphibole
 minerals have visual SAED patterns.that.closely resemble those of amphiboles.
 Gold-coating and Level II analysis will help in differentiating^ between  these
 minerals.                        .

 LABORATORY CONDITIONS

      Asbestos analysis involves sustained microscopy for periods  of  3 to
 7 hours with unscheduled rest breaks.   Subjective decisions  regarding such
 factors as morphology, size measurement, visual  identification, and  possible

                                        6

-------
         1, difficult to ore** a .anipul.tive
^r"o^V^»-i°«l^
redundant v>.^^	*-- •          -
deadlines may  contribute  to  poor  precision,

-------
                                   SECTION 4

                               LEVEL I ANALYSIS

SUMMARY OF PROTOCOL

     Level I analysis is a monitoring or screening technique.   It  assesses  the
amount and type of asbestos structures in the atmosphere through the  following
steps:
         ; - •.«_.
     (1) .A known volume of air is passed through a polycarbonate
         ;membrane filter (pore diameter, 0.4 ym; filter diameter,'
         137 or 47 mm) to obtain approximately 5 to 10 yg of
          Iparticulates per cm2 of filter surface.

     (2)  The particulate-laden filter is transported in its own
          filter holder.

     (3)  The filter is carbon-coated in the holder.

     (4) jThe particulates are transferred  to an EM grid using  a
         Irefined Jaffe wick washer.

     (5) ; The EM grid, containing the particulates, is gold-coated
         jj-ightly.

     (6)  The EM grid is examined under low magnification (250X to
          lOOOXffollowed by high-magnification  (16.000X on  the
          fluorescent screen) search and analysis.
         I »—•                                                         	
   :" (7) 'A known area (measured grid opening) is scanned, and  the   I
         '; fibrous structures (fibers, bundles, clusters, and          '
         .•matrices) are counted, sized, and identified as to
 •--y-K ~-.  • • asbestos type (chrysotile, amphibole,  ambiguous, or no
   /      identity) by morphology and by observing the SAED  pattern.-

     (8)  The observations are recorded—a  minimum of  100 fibrous
          structures or 10 grid openings, whichever is first.

     (9)  The data are reduced and the results reported.
EQUIPMENT, FACILITIES, AND SUPPLIES

     The following  items are  required  for  Level  I  analysis:

     (1)  An 80__qr  100-kV  TEM. with  a fluorescent viewing screen
          inscribed with graduations for estimating the length and
          width  of  fibrous particulates.

-------
     (2)  JA vacuum evaporator with -a turntable for rotating         7
         jspeciraens during coating,  for such uses as carbon-coating;
         [polycarbonate filters, gold-coating EM grids, and         i
         ^preparing carbon-coated EM grids.
     (3)   An EM preparation room adjacent to the room housing the
          EM.  This room should either be a  clean-room facility, or
          contain a laminar-flow class-100 clean bench to minimize
          contamination during EM grid preparation.  Filter handling
          and transfer to EM grids should be performed in a clean
          atmosphere.  Laboratory blanks should be prepared and
          analyzed weekly to ensure  quality  of work.
          Several refined Jaffe wick jwashers for dissolving membrane
          filters.
     (5)  Miscellaneous EM supplies and chemicals, including carbon-
          coated 200-mesh copper grids, grid boxes, and chloroform.
     (6)  Sample collection equipment, including 37-mm-diameter or
          47-mm-diameter filter holders, 0.4-ym (pore size)
          polycarbonate filters, 5.0-um (pore size) cellulose ester
          membrane filters for back-up, a sampling pump with
          ancillary equipment, a tripod, critical orifices or flow
          meters, and a rain/wind shield.
DESCRIPTION OF METHODOLOGY

1.  Type of Samples—Source

     This protocol was' originally developed for the EPA for measuring airborne
asbestos_..(Saraudra et al., 1977; Samudra et al., 19.7.8).  A broad interpretation
of airborne has been to apply the term to samples obtained from ambient air
(the original purpose), aerosolized source materials (such as  the asbestos
workplace environment, and fugitive dust emissions), bulk-air  material (such
as total suspended particulate (TSP) samples, dust, and powders) and any other
type of sample obtained by nonrestrictive use of (1) collection of  a volume  of
air, (2)'separation from the air, and (3) concentration of the particulates
onto a substrate.  The airborne protocol has also been applied to samples
collected in the regulatory areas of the EPA, as compared with, for example,
the workplace environment (National 'Institute of Occupational  Safety and
Health), mining activities (U.S. Bureau of Mines), and shipboard atmosphere
(Federal Maritime Administration).        	

     The present methodology has been optimized for application specifically
to samples collected from a volume of air in which the asbestos concentration
is considered a minor component of the total particulate  loading  (other analy-
tical methods are available for samples known to contain  high  concentrations
of asbestos); and in which the particles are less than  15 jjm  in diameter,
since particles~g"reater  than  15 gm either are not inhaled or  are  deposited  in
the upper respiratory tract and expelled, and preferably  less  than  10  ym  in
diameter as recommended  by the Clean Air Scientific Advisory  Committee
(Hileman,  1981), since particles up  to 10 pm can be absorbed  by  the alveolar

-------
  region  of the lung.   These concentration and size restrictions will preclude
  many  air samples collected in an asbestos-processing environment and in bulk-
  air material from the complete methodology.  However, such samples can still
  be  examined with the TEM,  within the limitations of the instrument, through
  changes in preparation techniques—provided the effects on the final results,
  such  as fractionation of size and representativeness of the sample, are
  carefully considered.

  2.  Sample Collection and Transport

  Sample  Collection—
       Sampling procedures vary depending on the nature of the sample, purpose
  of  collection, analytical method to be used, sample substrate, and time and
  cost  of sample collection relative to the total analytical effort.  Neverthe-
  less, the primary objective of sample collection always is to obtain a repre-
  sentative, unbiased  sample.

       Impingers, impaction devices, electrostatic precipitators, and thermal~1
  precipitators have been used in sample collection, but each has limitations.j
,  Presently, the preferred substrates are membrane filters^ which are manu-   "~*
  factured from different polymeric materials, including polycarbonate, mixed
  esters  of cellulose, polystyrene, cellulose acetate, and cellulose nitrate.
  Polycarbonate membrane filters..differ from the others in being thin, strong,
  and smooth-surfaced, and in having sieve-like construction (circular pores
  from  top surface to the bottom).  The other membrane filters are thicker, have
  irregular-surfaces,  and have depth-filter construction (tortuous paths from
  top surface to bottom).

       Consequently, polycarbonate filters have been selected for airborne f
 .asbestos analysis.  The collection of small-sized particles (prefer less than";
 !lO  urn in diameter),  the light loading of participates, the uniform distribu-
 tion  of particulates attainable using a depth-type backing filter, the smooth
  surface and circular holes (which aid in determining size and instrument tilt
'.axis),  and the relative ease in grid transfer (thin and strong) minimize     • -.-:«. **. •.
  disadvantages of lack of retention and/or movement of large particles during-'^SC^
  handling.  Other membrane materials,_ such as the cellulose.-ester....type,_.are__    ......
  recommended for phase" contrast and PLMj'neavy"Part*-cle- loadings, and physical   ,-,..
  • .  .. . ,_	 -  •--        _     .—	-         ->•-  .  . . .         .  .	    ...   "!'i.Vrp,!V '
  retention of large particles.                                                ••".>,-'"
 -r-    -   •                                                                      -••>•
       In microscopical analysis, uniformity of particulate distribution and
  loading is critical to  success.  Air samples are taken on 37-mm-diameter or
  47-mm-diameter, 0.4-ym  (pore size) polycarbonate membrane filters using  the
  shiny,  smooth side as the particle-capture surface.  Cellulose ester-type
  membrane filters (pore  size, 5.0 ytn) are used to support the polycarbonate
  filter  on the support pad (37-mra-diaraeter personal sampler) or on  the  support
  plate (47-mm-diameter holder).

       Air monitoring cassettes  (37-mm-diameter) of three-piece construction  are
  available from  several  manufacturers.  As with the 47-ram-diameter  filters,
  loading  the cassettes with the support pad, back-up  filter, and  0.4  ym  (pore
  size) polycarbonate filter should be carefully performed on a class-100  clean
  bench.    Since  the filters are  held  in place by pressure  fit rather  than  by

                                        10

-------
screw tightening, air must not enter from the sides of the unit; a plastic
band or tape (which can double as a label) should be used as a final seal.

     Collecting airborne samples with proper loading requires experience.
Each of the following techniques is useful in collecting airborne samples for
direct microscopy, preserving representative sizes, without diluting
participate deposits:

     (1)  For long-terra sampling at a site, test samples should be
          returned to the laboratory by express mail service, or air
          express service or by being hand-carried, and should then
          be analyzed by scanning electron microscopy.

     (2)  The estimated particulate loading (deposit is barely
          visible to the naked eye) should be bracketed by varying
          the filtration rate and using the same time, or by varying
          the time and using the same filtration rate.
     (3)  An automatic particle counter, such as a light-scattering
          instrument (0.3-ym detection) or a real-time mass monitor
          (0.1-ym detection), should be used to obtain an
          approximate particulate-loading level of the area.

     Although any one of the three techniques will work, the suggested  tech-
nique is to take the samples as a set, varying the sampling rates and using
the same time so as to obtain filter samples with different particulate
loadings.  Each set is composed of a minimum of four 37-mm-diameter or  47-mm-
diameter filter units — three for different particulate loadings  (low, medium,
high), and the fourth for a field blank.  Suggested sampling rates are  0  for
the field blank, 2.48. L/rain for the low loading, 7.45 L/min for  the medium,
and 17.62 L/min for the high, for a 30 min sampling period using a 47-mm-
diameter filter "holder.  Simultaneous sampling will provide at least  one
sample with a particulate loading suitable for direct EM analysis.
"• ''
     TSP's range from  10 yg/m3  in  remote, nonurban  areas,  to 60  yg/m3 in near-~7
                                                                               J
urban areas,  to  220 yg/m3  in  urban  areas.   However,  for heavily polluted
areas,,...TSP. levels may  reach  2000  yg/m3.   A loading of 5 to 10 _yg .per cm2 of
filter is  adequate for  EM  analysis;  values beyond  20 to 25 yg per cm"2 require'       '.'-,.''v:
a dilution treatment.   As  an  example,  for  47-mm-diameter filters at face
velocities of 3.0 cm/s  (2.48  L/min)t, 9.0 cm/s (7.45 L/min), and 21.2 cm/s
(17.62 L/min), respectively,  air  vo'lumes of 74.4 L,  223.5 L, and 528.6 L are
sampled  in 30 min.  For a  TSP level of 200 yg/m3,  14.88 yg (1.07 yg/cm2),
44.7 yg  (3.23 yg/cm2),  and 105.7  yg (7.63 yg/cm2), respectively, would be
collected  on  47-mm-diaraeter  filters (which would have effective filtration
areas of 13.85 cm2).   The  sampling  time  could be increased to 60 rain for areas
having lower  TSP levels, or  reduced in a heavily polluted area (source
emissions).

     Airborne samples  from emission sources contain coarse particles (above
the  respirable size) of large matrix structures, binder materials, road dust,
clay minerals, fillers, and  other materials.  For these samples, a fifth
filter unit can  be added that has a size-selective inlet (cyclone, impactor,  (
or  elutriator) attached prior to  the filter unit.  The flow pattern and flow

                                       11

-------
 rates  of  the tandem sampling arrangement must be checked before use.  A satis-
 factory,  tested combination presently used in California is a cyclone-filter
 unit with a DSQ cut-off of 2.5 ym at 21.7 L/rain, and a DSQ cut-off of 3.5 u™
 at  15.4 L/min (John and Reischl,  1980).   Additional sampling devices, such as
 impingers (used in biological sampling), impactors, and other designated
 filter units (for TSP,  XRD, or x-ray fluorescence (XRF), for example) can be
 added  to  the system to  obtain supplementary as well as interrelated data.

     This expandable multifilter  sampling unit, designated Hydra, offers the
 following advantages:

     (1)   It is small,  inexpensive, and compact, so that an adult
           can easily handle it.
     (2)   It is efficiently designed, and includes a tripod,
           sampling pump, manifold, critical orifices, and a row of
           preloaded 37-inm-diameter or 47-mm-diameter filter
           holders.  A rain/wind shield, size-selective cyclone-
           filter units, tubing, and other extras can be added as
           needed.
     (3)   Its sample preparation  steps and handling are minimized.
     (4)   It allows complementary as well as supplementary analysis
           (TSP, size fractionation, bacteria, and XRF, for example),
           although additional air sampling capacity is required.
     (5)   It accommodates ambient air and source emission samples,
           with or without a size-selective inlet.
     (6)   It allows synchronous sampling in several places in the
           vicinity following the  same sampling procedure, thereby
           accommodating particulate concentration fluctuations.
                •f
      (7)   It includes filter holders that serve as transport and
 . .         storage units.
$•&
 Hydra's disadvantages are a short sampling period, which may record  an
 episode; a small sampling quantity or volume, which may not indicate the
"'presence""6f ' asbestos fibers; £nd__a_detection...limit.pf_ 2 x_ 101*  fibers/in3  for ,..„..
 sampling 1 m3 of air with the 47-mm-diameter filter". .....       ..... ." '"  "''"r"iv v-'!'-4^'
 _ __ »1 --- II* — -- •"-" *"*•-._ ........    -. •   •.-•••«•.   .       __ .    . .                         _        1. .,!,'.,

      Using 8 inch x  10 inch, or 102-mm-dlameter filter sizes,  is not recom-
 mended.  The sampling units are designed for purposes other than microscopy.
 Interchanging the type of sample substrate filter  (glass fiber or  paper  to
 polycarbonate) does not correct the inherent problems of filter size and
 sampling unit.

 Sample Storage and Transport —
      Once  the sample is acquired, its integrity must  be assured, and contami-
 nation and loss of fibers prevented, until it is examined  under the  EM.   The
 low cost and small size of the 37-mm-diameter and  47-mm-diameter filter
 holders enables them to be used as combination storage and transport con-
 tainers.   The filter holders should be maintained  in  a horizontal  position
                                        12

-------
 during storage and transport to  the laboratory  so  that  the  particulate-loaded
 filters can be removed under optimally controlled  conditions  in  the  laboratory.

      For 47-mm-dianieter holders  (open-face)  to  be  used  in transport  or stor-
 age, the screw cap is carefully  removed, and  the shiny, waxy,  stiff  separator
 paper used to keep the polycarbonate filters  apart is carefully  placed on  the
 retaining ring.  The cap  is then carefully screwed back on  so  that  the sepa-
 rator paper seals and protects the particulate-loaded filter  without touching
 it.  The 37-mm-diameter,  three-piece filter  holder (aerosol monitor) is used
 in its open-face position, and capped after  usage  for transport  and  storage.

      When the more expensive 47-mra-diameter  holder is to be re-used  immedi-
 ately, the particulate-loaded  filter should  be  carefully removed and placed in
 a 47-mm-diameter Petri-slide (such as that manufactured by  the Millipore
 Corp.*)  This transfer takes place in the"field rather  than in the  laboratory,
 so that the Petri-slide should be taken  into the field. The  37-mm-diameter
 filter holder or the 47-mm-diameter holder/Petri-slide  should be secured and
 all necessary sample identification marks  and symbols applied to the holder.

 3.  Sample Preparation for Analysis—Grid  Transfer

 Carbon-Coating  the Filter—
 ~     The polycarbonate filter, with the  sample  deposit  and  suitable blanks,   f
 should be coated with carbon as  soon as  possible after  sampling  is  completed.;
 To begin this procedure,  the particulate-loaded 47-mra-diameter polycarbonate  j
 filter is removed from the holder and transferred  carefully to an open-faced  |
 47-mm-diameter  Petri-slide for carbon-coating in the vacuum evaporator (see —!
j_JFigure Al, Appendix A).   If  the  47-mm-diameter filter is already in the Petri-
 ^lide, the cover is replaced with an open-face cover, minimizing filter
 disruption.  The^37-mm-diameter  filter  is  left in  the holder, but the upper
 lid is removed  to create  an  open-faced  filter.   The open-faced holders are
 placed on the rotating turntable in  the  vacuum evaporator  for carbon-coating.
 Figure A2 shows  the multiple-coating arrangement in the evaporator; Figure A3
 shows a close-up of the  37-mm-diaraeter  and the modified 47-mm-diameter holders.
 for carbon-coating.
 '••"-'' For archival filters and those of'larger sizes, portions of about 2.5'cm
  x 2.5 cm should be cut midway between the center and edge using a scalpel.
  The portions are then attached 'withvcellophane tape to a clean glass micro-
  scope slide and placed on the turntable'in the"vacuum .evaporator for coating.

       Any high-vacuum carbon evaporator may.be used.to carbon-coat the.filters
  (CAUTION:  carbon sputtering devices should'not be'used),  typically, the
  electrodes are adjusted to a height of 10 cm above the level of the filters.
  A spectrographically pure carbon electrode sharpened to a neck of 0.1 cm x
  0.5 cm-is used as the evaporating electrode.  _The  sharpened electrode is
  * Millipore Corp., 80-T Ashby Rd., Bedford, Mass. 01730

                                        13

-------
                                • ' '
    placed  in its spring-loaded holder so that the 'rieck rests against the flat
    surface of a second carbon electrode.

         The manufacturer's instructions should be followed to obtain a vacuum of
    about 1.33 x 10~3  Pa (1 x 10~5  torr) in the bell jar of the evaporator.  'With
    the turntable in motion, the neck of the carbon electrode is evaporated by
    increasing the electrode current to about 15 A in 10 s, followed by 20 to 25 A
    for 25  to 30 s.  If the turntable is not used during carbon evaporation, the
    particulate matter may not be  coated from all sides, resulting in an undesir-
    able shadowing effect.  The evaporation should proceed in a series of short
    bursts  until the neck of the electrode is consumed.  Continuous prolonged
    evaporation should be avoided,  since overheating and consequent degradation of
    the polycarbonate filter may occur, impeding the subsequent step of dissolving
    the filter.  The evaporation process may be observed by viewing the arc
    through welders goggles (CAUTION:  never 'look at the arc without appropriate
    eye protection).  Preliminary calculations show that a carbon neck of 5 mm3
    volume, when evaporated over a spherical surface 10 cm in radius, will yield a
    carbon  layer that is 40 nm thick.

         Following carbon-coating,  the vacuum chamber is slowly returned to
    ambient pressure, and the filters are'removed and placed in their respective
    holders or in clean, marked Petri dishes for storage on a clean bench.

    Transfer of the Sample to the EM Grid—
         Transferring the collected particulates from the carbon-coated polycar-
    bonate  filter to an~EM~grid is accomplished in a clean room or on a class-100
    clean bench".  The transfer is made in a Jaffe wick washer, which is usually a
    glass Petri dish containing a substrate to support the EM grid/carbon-coated
    membrane filter combination.  Solvent is added to a level to just wet the   "~~1
   fcombination and cause gentle dissolution of the membrane with minimum loss or/
   .'dislocation of the particulates, resulting in a membrane-free EM grid with   '
   I particles embedded in the carbon film coating.  The substrate support can  be
  ;J:stainless steel mesh bridges, filter papers, urethane  foams, or combinations
V-joT these.                               -          .
 -  U   •
TV-?---.-.-. The. ref ined.,.Jaf fe wick washer is_ described as  follows:
          (1)  The glass Petri dish (diameter,  10 cm; height,  1.5  cm)  is
               made airtight by grinding the  top edge of  the  bottom dish
               with the bottom of the cover dish, with water  and  .  .
               Carborundum* powder (80 mesh);  this creates  a  ground-
               glass seal .(closer fit). ..and minimizes the  need  to-ref.ill
               the Petri dish with added solvent.  (The usual  glass.....
               Petri dish was found, .not to retain the solvent  for  long
               periods of time, and unless the  wicking substrate  is kept
               continuously wet, poor solubility of the membrane  filter
               results, leading to a poor-quality EM grid).
    * Carborundum  is  a  registered  trademark of  the  Carborundum Co.,  Carborundum
      Center,  Niagara Falls, N.Y.  14302.

                                           14

-------
      (2)   A combination of foam and a single sheet of 9-cm filter
           paper is used as the substrate support.  A 3-cm x 3-cm x
           0.6-cm piece of polyurethane foam (the packing in
           Polaroid film boxes) is cut and placed in the bottom
           dish.  A 0.5-inch V-shaped notch is cut into the filter
           paper; the notch is oriented in line with the side of the
           foam, creating a well for adding solvent.
           Spectrographic-grade chloroform (solvent) is poured into
           the Petri dish through the notch until it is level with
           the top of the foam (also level with the paper).  The
           foam will swell, and care is needed to avoid adding
           solvent above the filter paper.
      (3)   On top of the filter paper, pieces of 100-mesh stainless
           steel screen (0.6 cm x 0.6 cm) are placed, usually in two
           rows, to make several grid transfers at one time (for
           such uses as replicas), and to facilitate maintenance of
           proper identity of each transfer.
           A 3-mm section (usually midway between the center and
           edge) of the carbon-coated polycarbonate filter is cut in
           a rocking motion with a scalpel.  The section may be a
           square, rectangle, or triangle, and should just cover the
           3-mm EM grid.
      (5)  A section is laid carbon-side down on a 200-mesh carbon-
           coated EM grid.  (Alternatively, Formvar-coated* grids or
           uncoated EM grids may be used.  Here, the carbon coating
           on the polycarbonate filter forms the grid substrate.)
           Minor overlap or underlap of the grid by the filter
           section can be tolerated, since only the central 2-mra
           portion of the grid is scanned in the microscope.  The EM
           grid and filter combination is picked up at the edges
           with the tweezers and carefully laid on the damp 100-mesh
           stainless steel screen.  The EM grid-filter combination
           will immediately "wet out" and remain on the screen.

      (6)  Once all specimens are placed in the washer, more solvent
           is carefully added through the notch" to maintain the
           liquid level so that it just touches the top of the paper
           filter.  Raising the solvent level any higher may float
           the EM grid off the mesh or. displace the polycarbonate
           filter section.          .       - ...... ...      .  ..

      (7)  The cover is placed in the washer 'and oriented in place
           over the specimen, and a map of 'the fiTter/grid'/s'creen
           arrangement is made on the glass cover and in the
           logbook.                          • '•-"•-  •"-
* Forravar is a registered trademark of  the Monsanto  Company,  800  N.  Lindbergh
  Blvd., St. Louis, Mo.

                                       15

-------
       (8)   Solvent  (chloroform) is added periodically to maintain
            the  level  within the washer mun_the_'jfiJLter_ i£_
            completely dissolved by the wicking action (24 to 48 h).
            * ,_-•—i»-iftu- •^^"•"** "• '•*»•- . v f  ~  '       ^  |Ht,fci    _  - *•* !»•*•  -•*, ^ j  ««-  .,  „, i , . •  - •" r"»^
       (9)   The  temperature in the room must remain relatively
            constant  to minimize condensation of solvent on the
            bottom of  the cover and subsequent  falling of solvent
            drops on  the EM grid.  Should day-night or other
            temperature differentials occur, solvent condensation on
            the  under-surface of the cover can  be minimized by
            placing  the Jaffe washer at a slight tilt (three glass
            slides under one edge of the Petri  dish parallel to the
            row  of grids) to allow the condensation drops to flow
            toward the lower edge rather than fall on the EM grids.
            At temperatures lower than 20°C (68°F), the complete
            filter solution may take longer than 72 h.

      (10)   After the  polymer is completely dissolved, the stainless
            steel mesh screen with the EM grid  is picked up while wet
            and  set  on lens paper tacked to the bottom of a separate
            Petri dish.  The EM grid is then lifted from and placed
            next  to  the screen to dry.  When all traces of solvent
            have  evaporated, the grid is stored in a grid box and
            identified by location and grid box in the logbook.

      Figure A4  illustrates the Jaffe wick washer method; Figure A5 shows  the
washer.  The foam/filter combination is currently preferred, as is use  of a
closely  fitted  (by means of the ground-glass seal) Petri dish.

Gold Coating—
      Ajn_addi.t.i.o.nal...s,tep will aid jin_s_ubjectively_ evaluating  theJSAED  pattern.
This step  is required for specimens from the uppeT~Gfeat" Lakes area and for
those of unknown*origins.  After the particulates on the filter are
transferred to  the  EM grid, the grid is held to a glass slide with double
stick tape for  gold-coating in the vacuum evaporator.   Several EM grids may be.
taped to the glass slide for coating at one time.  Approximately  10 mm  of
0.015-cm-diameter (0.006-inch) pure gold wire is placed in a tungsten basket
(10.cm from.the  rotating table holding the EM grids) and evaporated onto the
grid.                                 	~~           	           •'"•-.T"

      The thin gold-coating establishes an internal standard  for SAED  analysis.
For some mineral species, an internal standard will clarify  visual identifica-
tion of  the pattern of a fibrous  particulate as  being or not being an amphi-
bole species (for example, minnesotaite as opposed to amosite).  With exper-
ience, differentiation in SAED patterns can be observed.   For  samples of known!
[geographic origins, gold-coating  is optional, since the additional coating
Ihinders  observation and identification of small-diameter chrysotile  fibers.   \

 4.  TEM Examination and Data Collection
                                                                                    . -.0.
                                                                                    v;s-.. .;
low Magnification Examination  of  Grids—
     Figure A6 shows a modern  TEM.   The  grid  is  observed in the TEM at
magnifications of 250X and  1000X  to  determine its suitability for detailed

                                      16

-------
 study at  higher magnification.  The grid is rejected and a new grid  used  if:
 (1)  the carbon film over a majority of the grid openings is damaged  and not
 intact;  (2)  the specimen is "dark due to incomplete dissolution of  the
 polycarbonate filter; or (3)  the particulate loading is too light  (unless  a
 blank) or too heavy with particle-particle interactions or overlaps.

 TEN Analysis (Morphology and SAED)—
      The  following guidelines are observed for consistency in the  analytical
 protocol:

       (1)  Magnification at the fluorescent screen is determined by
            calibration with a diffraction-grating replica in the
            specimen holder.

       (2)  A field o'f view or "gate" is defined.  On some
            microscopes, the central rectangular portion of the
            fluorescent screen, which is lifted for photographic
            purposes, is convenient to use.  On others, a scribed
            circle or the entire circular screen may be used as the
            field of view.  The area of the field of view must be
            accurately measurable.

       (3)  The grid opening is selected on a random basis.

       (4)  The analysis, morphology, and SAED are performed at a
            tilt angle of 0°.

       (5)  The recommended instrument settings are:  accelerating
            voltage,  100 kV; beam current,  100 pA; film magnifi-
            cation,  20JOOOX (which is equivalent to  16.000X on  the
            fluorescent screen for this instrument); and concentric
            circles  of radii  1, 2, 3, and A cm on  the fluorescent
   .  ,       screen.
                                                                                    ii"*». '".. T-7.-iV '
 "•"'','  •"• '•  •  •                                                                       •  .-.•"; >".'•.•£••>•
       (6)  The grid  opening  is measured at  1000X.                                    ' ": '  " '

•:• :." "|-5-(7)  'Since asbestos  fibers are found  isolated" as well  as with           •   .-:^K'"^.^.
            each other or with other particles in  varying  arrange-                  "" "  .''•': '
            ments, the fibrous particulates are characterized  as
            asbestos  structures:     v

            Fiber (F) is a  particle with an  aspect ratio of  3:1  or
            greater,  with substantially parallel sides.
            Bundle  (B) is a particulate composed of  fibers in a
            parallel  arrangement, with each  fiber  closer  than the
            diameter of  one fiber.

            Cluster  (Cl)  is a particulate with  fibers  in  a random
            arrangement  such  that all  fibers  are intermixed  and no
            single  fiber  is isolated  from  the group.
                                        17

-------
     Matrix (M)  is a fiber or fibers with one end free and the
     other end embedded or hidden by a  particulate.

     Combinations of structures,  such as matrix and cluster,
     matrix and  bundle, or bundle and cluster, are categorized
     by the dominant fiber quality—cluster, bundle, or
     matrix.

(8)  Counting rules for single fibers are:

     (a)  Particulates meeting the definition of fiber are
          isolated by themselves.  With this definition, edge
          view of flakes, fragments from cleavage planes, and
          scrolls, for example, may be counted as fibers.

     (b)  Count  as single entities if separation is equal to
          or greater than the diameter of a single fiber.

     (c   Count  as single entities if three ends can be seen.
     (d)  Count  as single entities if four ends can be seen.

     (e)  In general, fibers that touch or cross are counted
          separately.
     (f)  Two or more fibers are counted as a bundle if the
          distances between fibers are less than the diameter
          of a single fiber, or if the ends cannot be
          resolved.

     (g)  Fibrils attached longitudinally to a fiber are
          counted as part of the fiber and the size (width) is
          estimated based on the fiber-to-fibril relationship.

     (h)  A fiber partially hidden by grid wires (one or  two
          ends) is counted, but labeled as an X-fiber.  If  the
          number of X-fibers is more than 20% of the fibers
          identified as asbestos, a  larger-mesh EM grid should
          be used, such as 100 mesh  (about 200 pm wide).

(9)  Sizing rules for  asbestos structures are:
     (a)  For fibers, widths and lengths are obtained by
          orienting  the fibers to the inscribed circles on  the
          fluorescent  screen.  Since estimates are within
          ±1 mm,  small-diamete-r  fibers  have  greater margins of
          error.  Fibers  less  than  1 mm at  the fluorescent
          screen magnification level are characterized  as
          being  1 mm.  A  cylindrical shape  is assumed for
          fibers.   X-fibers  are  sized by measuring  their.
          entire visible  portions in the grid opening.

     (b)  Bundles  and  clusters are  sized  by estimating  their
          widths  and  lengths.  The  sum  of individual  diameters
          is  used  to  obtain  the  total width,  and  an average
          length  for  the  total length.   A laminar-sheet shape
                                 18

-------
                is assuned, with the average diameter of the
                individual fiber as the thickness.

           (c)  Matrices are sized by adding the best estimates of
                individual fiber components.  A laminar or sheet
                structure is assumed for volume calculation.

     (10)  The method of sizing is as follows:

           (a)  An asbestos structure is recognized, and its
                location in the rectangular "gate" relative to the
                sides, inscribed circles, and other particulates, is
                memorized.

           (b)  The structure is moved to the center for SAED
                observation and sizing.

           (c)  Sizing is performed using the inscribed circles.  If
                the structure, such as a fiber, extends beyond the
                rectangular "gate" (field of view), it is super-
                imposed across the series of concentric circles
                (several times, if necessary) until the entire
                structure is measured.

           (d)  The structure is returned to its original location
                by recall of the location, and scanning is
                continued.

Figure A7 illustrates some of the counting and morphology guidelines used  in
determining asbestos structures.

TEM Procedure—
     The TEM procedure is as follows:

      (1)  EM grid quality is assessed at 250X.

      (2)  Particulate loading is assessed at  1000X.

      (3)  A grid opening is selected  at random, examined at  1000X,
           and sized.

      (4)  A series of parallel traverses is made  across  the  grid
           opening at the film magnification of 20.000X.  Starting
           at one corner, and using  the  tilting section of  the
           fluorescent screen as a  "gate" or "chute,"  the grid
           opening is traversed.  Movement  through the  "gate"  is  not
           continuous, but rather is a stop/go motion.  On  reaching
           the end of one  traverse,  the  image  is moved  the  width  of
           one "gate," and the traverse  is  reversed.   These parallel
           traverses  are  made until  the  entire grid opening has  been
           scanned.

      (5)  Asbestos  structures are  identified  morphologically and
           counted as they enter  the  "gate."
                                       19

-------
(6)  The asbestos structure is categorized as fiber (with or
     without X-) bundle, cluster, or matrix, and sized through
     use of the inscribed circles.

(7)  The structure (individual fiber portion) is centered and
     focused, and the SAED pattern is obtained through use of
     the field-limiting aperture.

     (a)  SAED patterns from single fibers of asbestos
          minerals fall into distinct groups.  The chrysotile
          asbestos pattern has characteristic streaks on layer
          lines other than the central line, and some
          streaking also on the central line.  Spots of normal
          sharpness are present on the central layer line and
          on alternate lines (that is, 2nd, 4th etc.)  The
          repeat distance between layer lines is about
          0.53 nm.

     (b)  Amphibole asbestos fiber patterns show layer lines
          formed by very closely spaced dots, and have repeat
          distances between layer lines also of about
          0.53 nm.  Streaking in layer lines is occasionally
          present due to crystal structure defects.

     (c)  Transmission electron micrographs and SAED patterns
          obtained with asbestos standard samples should be
          used as guides to fiber identification.  An example
          is the "Asbestos Fiber Atlas" (Mueller et al.,
          1975).

(8)  From visual examination of the SAED pattern, the
     structure is classified as belonging to one of four
     categories:  (1) chrysotile, (2) amphibole group
     (includes amosite, crocidolite, anthophyllite, tremollte,
     and actinolite), (3) ambiguous (incomplete spot
    ^patterns), or (A) no identification.  SAED patterns  "~
    ""cannot be inspected for some fibers.  Reasons for the
     absence of a recognizable diffraction pattern include
     contamination of the fiber, interference from nearby
     particles, fibers  that are too small or too thick, and
     nonsuitable orientation of the fiber.   Some chrysotile""*
    ^fibers are destroyed in t,he electron beam, resulting in
     'patterns  that fade away within seconds  of being  formed.
     Some patterns are  very faint and can be seen only under
     the binocular microscope.  In general,  the shortest
     available camera length must be used, and the objective
     lens current may need to  be adjusted to give optimum
     pattern visibility for correct identification.   A 20-cm
     camera  length and  a  10X binocular are  recommended for
     inspecting the SAED pattern on the tilted screen.
                                 20

-------
        (9)  Additional grid  openings  are  selected,  scanned,  and
            counted  until  either  the  total number of  structures
            counted  exceeds  100 per known area  or a minimum  of  10
            grid  openings  has  been scanned, whichever is  first.

       ( 10)  The TEM  data should be recorded in  a systematic  form so
            that  it  can be processed  rapidly.   Sample information,
            instrument parameters, and  the sequence of operations
            should be tabulated for ease  in data reduction and
            subsequent reporting  of results.  Figure  A8 shows an
            example  of a data  sheet used  in Level I analysis.

       Figure  A9 illustrates  the method  of scanning  a  full-grid  opening.   The
  "field of  view"  method of counting previously  included in the  provisional
  methodology,  which  is based on randomly  selected fields  of  view,  has been
  discontinued.  Originally,  the method  was recommended for medium  loading level
  on  the filter  (50 to 300  fibers  per  grid opening).  However, if samples are
  collected  at  three  different  loading levels  and the  optimum is  selected, this
  medium loading on the filter  will not  be used. Samples  with grid openings
  containing 50  to 300 fibers may  be used  as laboratory fiber preparations or
  selected  source  samples,  but  in  field  samples  the  particulate  loading is
  usually of much  higher concentration than the  fiber.  Filter loading is
  characterized  by the particulate concentration, not  by fiber concentration.

  5.   Data  Reduction  and Reporting of  Results

  Data Reduction—
       From the data  sheet, size measurements  are converted  to microns (16,OOOX~|
  screen magnification), mass of asbestos  structure  is calculated,  and other    I
  characterizing parameters are calculated through  use of  a  hand calculator orI
\ computer.  (Appendix B, an  example of  a  computer  printout  from Level I
 ""analysis,  shows  reduced data—that is, what  was  found on the specified number
  of  grid  openings or area  examined.)  These measurements are summarized and
 ?,related  to the volume of  air  sampled and the total effective filtration area •
  (area of  deposit).  Size  measurements  of X-fibers  may be doubled  and noted, or
  kept as  a separate  category.
   •  '  Fiber number concentration is calculated from the equation

                                    t L
                  „.,    .  ,     Total no. of fibers
                  Flbers/ra3  =   No. of EH fields

                                  Total effective filter area, cm2
                                      Area of an EM field, cm2


                                  Volume of air sampled, m3


  The number of X-fibers,  bundles, clusters, and matrices are calculated in a
  similar manner.  X-fibers may be included with fibers if they are few in
                                        21

-------
number.  Similarly, their corresponding mass (from their size measurements)
may be included.

     Fiber mass for each type of asbestos (chrysotile or amphibole) in the
sample is calculated by assuming that both chrysotiles and amphiboles have
circular cross-sections (cylindrical shape) and that the width measurements
are one diameter.  The density of chrysotile is assumed to be 2.6 g/cm3, and
of amphiboles to be 3.0 g/cra3 .  The individual mass is calculated from the
equation
                Mass,  yg  = y x (length,  ym)  x (diameter,

                            x (density,  g/cm3) x 10 6


The total mass  concentration of  fibers for each  type  of asbestos is  then
calculated from the total mass of all the individual  fibers of that  type.

     The individual masses of bundles, clusters, and  matrices are calculated
by assuming a laminar or sheet-like structure  with an average thickness of  the
fiber make-up of the  structure.  Again, the density of chrysotile is  assumed
to be 2.6 g/cm3, and  of amphiboles to be  3.0 g/cm3.   The individual masses  are
calculated from the equation


           Mass, yg  =  (length, ym)  x (width, ym)  x (thickness,  ym)
                        x (density,  g/cm3) x 10 6
The total mass  for each  type of  structure for each  type  of  asbestos  is  the  sum
of all the individual masses.

     Other characterizing  parameters  of  the asbestos  structures  are:   (1)
length and width distribution of fibers, (2) aspect ratio distribution  of
fibers, and  (3) relationships of fibers, bundles, clusters,  and  matrices.

Reporting of Results—
     The data in their acquired  and reduced forms are reported  as  summarized,
or, depending on the purpose of  the analysis, are further reduced  to present
the interrelationships of  the various characterizing  parameters.  Again,  the
Level  1 methodology is a monitoring or screening  technique,  and  its  limita-
tions, such  as  the possibility of "false positives" and  misidentification,
should be noted.

6.  Quality  Control/Quality Assurance

     Sampling procedures will vary depending on the type of sample,  objectives
of  the sampling, and time/cost factors.  The primary  goals  of sampling are  to
obtain a  representative  sample at the location  and  time  of  sampling, and  to
maintain  sample integrity.  The  sampling team will  have  written sampling

                                       22

-------
procedures, and the field chief and/or designated individual will  be
responsible for all record-keeping  (including sample identification,  labeling,
logging of data, site description,  and meteorological conditions),  pre-  and
post-collection checks, and continuous sample custody and sign-outs until  the
sample is delivered to  the laboratory and  transferred to the appropriate
quality assurance officer (QAO).  Verification of sampling  times,  flow rates,
equipment calibration,  and taking of field  blanks will  be checked  arid recorded
in the field logbook.

     Samples are turned over  to the QAO  for logging into a  project  logbook.
Each sample is carefully examined for gross features, such  as  tears,  breaks,
and overall condition of container. The  QAO registers  the  as-received sample
number and other designated information,  and assigns a  simple  internal code
number that will accompany the sample through the preparation  stage,  grid
transfer, grid analysis, data reduction,  and reporting  of results.

     After being logged into  the project  logbook, the sample is  transferred  to
the custody of the electron microscopy staff, where every precaution  is  taken
to maintain sample integrity  and to prevent contamination and  loss  of
collected particulates.  During storage  and transport,  the  filters  in their
respective holders are  maintained in a horizontal position  at  all  times.

     The sample logging, handling,  and storing procedures ensure that all
samples can be readily  located and  identified throughout  the course of a
program.  The QAO has divisional responsibility  for QC/QA activities, and  must
see that the laboratory maintains high standards.  He must  be  aware of current
standards of analysis,  and must ensure that internal quality control
standards, instrument calibration,  and records of samples and  completed
analyses are kept for ease of later retrieval and use.

     For quality*control, internal  laboratory blanks are  analyzed  at  least
once a week, which may  or may not coincide with  a sample  batch blank.  In
addition, a magnification calibration of the EM  using  a carbon grating replica.—.  ,
(2,160 lines per mm)  is performed once a week.   The  results are recorded in-aiy£.;; ,
EM instrument log, along with other routine instrumental  performance checks.
All photographs, TEM, SEM, and  STEM images are  recorded in  a  photo log.   These^...' ..
.QCI"fesu'lts are documented  for inspection by the  QAO.                        'v'v%;'-. •
                                       23

-------
                                    SECTION 5

                                LEVEL II ANALYSIS

  SUMMARY OF PROTOCOL

 I      Level II analysis is a regulatory technique  consisting of Level Ij
 [analysis plus chemical elemental analysis.   Morphology,  size,  SAED pattern,
 ~and  chemical analysis are obtained sequentially.   By a process of elimination,
  mineral fibers are identified  as chrysotile,  amphibole,  ambiguous, or "no-
 J.dentity" by morphology and SAED pattern.   X-ray  elemental analysis is used to
 | categorize the amphibole fibers, identify  the ambiguous  fibers, and confirm or
 {validate chrysotile fibers.

       Level II analysis is summarized as follows:

       (1)  A known volume of air is passed  through a polycarbonate
            membrane filter (pore diameter,  0.4 um; filter diameter,
            37 or 47 mm) to obtain approximately 5  to 10 ug of
            particulates per cm2 of filter surface.
       (2)  The particulate-laden filter is  transported in its own
            filter holder.
       (3)  The filte* is carbon-coated in the holder.
.-
       (4)  The particulates are transferred to an  EM grid using a
            refined Jaffe wick washer.
       (5)  The EM grid, containing the particulates, is gold-coated
            lightly.
       (6)  The EM grid is examined under low magnification (250X to
         pv.'lOOOX)- followed by high-magnification (16,OOOX on the
       ^iv  fluorescent screen) search and analysis.
       (7)  A known area (measured grid opening) is scanned, and the
            fibrous structures (fibers, bundles, clusters, and
            matrices) are counted, sized, and identified as to
            asbestos type (chrysotile, amphibole, ambiguous, or no
            identity) by morphology and by observing the SAED pattern;
            and finally by elemental analysis using EDS.
       (8)  The observations are recorded—a minimum of 100 fibrous
            structures or 10 grid openings, whichever is first.
       (9)  The data are reduced and the results reported.
                                        24

-------
EQUIPMENT, FACILITIES, AND SUPPLIES

     The following items are required for Level 1 analysis:

     (1)  A modern 100-kV TEM equipped with an EDS.  A scanning
          accessory as found in a STEM will increase the versatility
          and analytical capability for very small fibers and for
          fibers adjacent to other particulate matter.  The
          microscope should be equipped with the fluorescent viewing
          screen inscribed with graduation of known radii to
          estimate the length and width of fibrous particulates.

     (2)  A vacuum evaporator with a turntable for rotating
          specimens during coating, for such uses as carbon-coating
          polycarbonate filters, gold-coating EM grids, and
          preparing carbon-coated EM grids.

     (3)  An EM preparation room adjacent to the room housing the
          EM.  This room should either be a clean-room facility, or
          contain a laminar-flow class-100 clean bench to minimize
          contamination duing EM grid preparation.  Filter handling
          and transfer to EM grids should be performed in a clean
          atmosphere.  Laboratory blanks should be prepared and
          analyzed weekly to ensure quality of work.

     (A)  Several refined Jaffe wick washers for dissolving membrane
          filters.

     (5)  Miscellaneous EM supplies and chemicals, including carbon-
          coated 200-mesh copper grids, grid boxes, and chloroform.

     (6)  Sample collection equipment, including 37-mm-diameter or
          47-mm-diameter filter holders, 0.4-ym (pore size)
          polycarbonate filters, 5.0-ym (pore size) cellulose ester
          membrane filters for back-up, a sampling pump with
  1   ,;     ancillary equipment, a tripod, critical orifices or flow
 ••'"-  '-     meters, and a rain/wind shield.


DESCRIPTION OF METHODOLOGY                                                 -  •-- "?*;:•••; -V'>V-;>:
                                                                                • '..-'",'•;' '.' ^ '' '.•
1.  Type of Samples—Source

     This protocol is an expansion of the method originally developed  for the
EPA for measuring airborne asbestos (Samudra et al.,  1977; Samudra et  al.,
1978).   A broad interpretation of airborne has been to apply the  terra  to
samples obtained from ambient air (the original purpose), aerosolized  source
materials (such as the asbestos workplace environment, and fugitive  dust
emissions), bulk-air material (such as total suspended particulate  (TSP)
samples, dust, and powders) and any other type of  sample obtained by nonre-
strictive use of (1) collection of a volume of air, (2) separation from the
air, and (3) concentration of the particulates onto a substrate.  The  airborne
protocol has also been applied to samples collected in the regulatory  areas  of
the EPA, as compared with, for example, the workplace environment  (National


                                      25

-------
Institute of Occupational Safety and Health), mining activities (U.S. Bureau
of Mines), and shipboard atmosphere (Federal Maritime Administration).

     The present methodology has been optimized for application specifically
to samples collected from a volume of air in which the asbestos concentration
is considered a minor component of the total particulate loading  (other analy-
tical methods are available for samples known to contain high concentrations
of asbestos); and in which the particles are less than 15 ym in diameter,
since particles greater than 15 ym either are not inhaled or are  deposited  in
the upper respiratory tract and expelled, and preferably less than  10 ym  in
diameter as recommended by the Clean Air Scientific Advisory Committee
(Hileman, 1981), since particles up to 10 ym can be absorbed by the alveolar
region of the lung.  These concentration and size restrictions will preclude
many air samples collected in an asbestos-processing environment  and in bulk-
air material from the complete methodology.  However, such samples  can still
be examined with the TEM, within the limitations of the instrument  by changes
in preparation techniques—provided the effects on the final results, such  as
fractionation of size and representativeness of the sample, are carefully
considered.

2.  Sample Collection and Transport

Sample Collection—
     Sampling procedures vary depending on  the nature of the sample, purpose
of collection, analytical method to be used, sample substrate, and  time and
cost of sample collection relative to the total analytical effort.  Neverthe-
less, the primary objective of sample collection always is to obtain a
representative, unbiased sample.

     Impingers, impaction devices, electrostatic precipitators, and thermal
precipitators ha~ve been used in sample collection, but each has limitations.
Presently, the preferred substrates are membrane filters, which are manufac-
tured from different polymeric materials, including polycarbonate,  mixed
esters-of cellulose, polystyrene, cellulose acetate, and cellulose  nitrate.. •.>;•;.!
Polycarbonate membrane filters differ from  the others in being thin, strong,
and..smooth-surfaced, and in having sieve-like construction  (circular pores    ......
fr-om .top/ surface to the bottom).  The other membrane "filters "are' thicker; have'y.'
irregular-surfaces, and have depth-filter construction  (tortuous  paths  from
top surface to bottom).
                                  ., &

     Consequently, polycarbonate filters have been selected  for airborne
asbestos analysis.  The collection of small-sized particles  (prefer less  than
 10 ym in diameter), the light  loading of particulates,  the  uniform distribu-
tion of particulates attainable using a depth-type backing  filter,  the  smooth
surface and circular holes  (which aid in determining size and  instrument  tilt
axis), and  the relative ease  in grid  transfer  (thin and  strong) minimize
disadvantages of lack  of retention and/or movement of  large  particles  during
handling.   Other membrane materials,  such as the cellulose  ester  type,  are
recommended  for  phase  contrast  and PLM, heavy  particle  loadings,  and  physical
retention of  large particles.
                                       26

-------
     In microscopical analysis, uniformity of particulate distribution  and
loading is critical to success.  Air samples are taken on 37-mm-diaraeter  or
47-mm-diameter, 0.4-ym (pore size) polycarbonate membrane filters using the
shiny, smooth side as the particle-capture surface.  Cellulose ester-type
membrane filters (pore size, 5.0 um) are used to support the polycarbonate
filter on the support pad (37-mm-diameter personal sampler) or on the support
plate (47-mm-diameter holder).

     Air monitoring cassettes (37-mm-diameter) of three-piece construction  are
available from several manufacturers.  As with the 47-mm-diaraeter filters,
loading the cassettes with the support pad, back-up filter, and 0.4  urn  (pore
size) polycarbonate filter should be carefully performed on a class-100 clean
bench.  Since the filters are held in place by pressure fit rather than by
screw tightening, air must not enter from the sides of the unit; a plastic
band or tape (which can double as a label) should be used as a final seal.

     Collecting airborne samples with proper loading requires experience.
Each of the following techniques is useful in collecting airborne samples for
direct microscopy, preserving representative sizes, without diluting
particulate deposits:

     (1)  For long-term sampling at a site, test samples should be
          returned to the laboratory by express mail service, or air
          express service or by being hand-carried, and should then
          be analyzed by scanning electron microscopy.

     (2)  The estimated particulate loading (deposit is barely
          visible to the naked eye) should be bracketed by varying
          the filtration rate and using the same time, or by varying
          the time and using the same filtration rate.
     (3)  An automatic particle counter, such as a light-scattering
          instrument (0.3-um detection) or a real-time mass monitor
 	,. .   (0.1-ym detection), should be used to obtain an
 .';:-,;r.--;:••    approximate particulate-loading level of the area.

     Although any one of the three techniques will work, the suggested
'technique is to" take the samples as a set, varying ..the. .sampling rates and     ''":.-'
using the same time so as to obtain filter samples with different  particuiafe:V>v
loadings.  Each set is composed of a minimum of four 37-mm-diameter  or  47-min-  -l;-
diameter filter units—three for different particulate loadings (low, medium,
high), and the fourth for a field blank.  Suggested sampling rates  are  0 for
the field blank, 2.48 L/min for the low loading, 7.45 L/min for the  medium,
and 17.62 L/min for the high, for a 30 min sampling period using  a  47-mra-
diameter filter holder.  Simultaneous sampling will provide at  least one
sample with a particulate loading suitable for direct EM analysis.

     TSP's range from 10 ug/m3 in remote, nonurban areas,  to 60 vig/ra3 in near-
urban areas, to 220 ug/m3 in urban areas.  However, for heavily polluted
areas, TSP levels may reach 2000 ug/m3.  A loading of 5 to  10  ug  per cm2 of
filter is adequate for EM analysis; values beyond  20 to 25  ug  per  cm2 require
a dilution treatment.  As an example, for 47-mm-diameter filters  at  face
velocities of  3.0 cm/s (2.48 L/min), 9.0 cm/s (7.45 L/min), and 21.2 cm/s

                                      27

-------
(17.62 L/rain), respectively, air volumes  of  74.4  L-;  223.5  L,  and 528.6 L are
sampled in 30 min.  For a TSP level of  200 yg/m3,  14.88 yg (1.07 yg/cm2),
44.7 yg (3.23 Ug/cm2), and  105.7 vg (7.63 yg/cm2),  respectively, would be
collected on 47-mra-diameter filters (which would  have  effective filtration
areas of  13.85 cm2).  The sampling time could  be  increased to 60 min for areas
having lower TSP levels, or reduced in  a  heavily  polluted  area (source
emissions).

     Airborne samples from  emission sources  contain  coarse particles (above
the respirable size) of large matrix  structures,  binder materials,  road dust,
clay minerals, fillers, and other materials.   For these samples, a  fifth
filter unit can be added that has a size-selective  inlet (cyclone,  impactor,  .
or elutriator) attached prior to the  filter  unit.  The flow pattern and flow
rates of  the tandem sampling arrangement  must  be  checked before use.  A
satisfactory, tested combination presently used>in  California is a  cyclone-
filter unit with a DSQ cut-off of 2.5 ym  at  21.7  L/min, and a DSQ cut-off of
3.5 ym at 15.4 L/min (John  and Reischl, 1980).   Additional sampling devices,
such as impingers (used in  biological sampling),  impactors, and other
designated filter units (for TSP, XRD,  or x-ray  fluorescence (XRF), for
example)  can be added to the system to  obtain  supplementary as well as inter-
related data.

     This expandable multifilter sampling unit,  designated Hydra, offers the
following advantages:

     (1)  It is small, inexpensive, and compact,  so that an adult
          can easily handle it.
     (2)  It is efficiently designed, and includes a tripod,
          sampling pump, manifold, critical  orifices,  and a row of
          preloaded 37-mm-diameter or 47-mm-diameter filter
          holders.  A rain/wind shield, size-selective cyclone-
          filter units, tubing, and other extras can be added as
       .."''needed.                                '                      .  ...,....,.....-,._.x
       '•••'••   ' '• '  '                                                         •• ,-.' .;••>-: Tr-.'-i-
     (3)  .Its sample preparation steps  and handling are minimized.          -••''•^•.'.•i1'.1
                                                                              - • '-* "•' •
	-:-(-4).r.;*ic.-allows complementary, as  well as supplementary^ analysis
 .  '" '  .   '::(TSP, size fractionation, bacteria,  and XRF, for example),  " 
-------
     Using 8 inch x 10 inch, or 102-mm-diameter filter sizes, is not recom-
mended.  The sampling units are designed for purposes other than microscopy.
Interchanging the type of sample substrate filter (glass fiber or paper to
polycarbonate) does not correct the inherent problems of filter size and
sampling unit.

Sample Storage and Transport—
     Once the sample is acquired, its integrity must be assured, and contami-
nation and loss of fibers prevented, until it is examined under the EM.  The
low cost and small size of the 37-mm-diameter and 47-mm-diameter filter
holders enables them to be used as combination storage and transport con-
tainers.  The filter holders should be maintained in a horizontal position
during storage and transport to the laboratory so that the particulate-loaded
filters can be removed under optimally controlled conditions in the labora-
tory.

     For 47-mm-diameter holders (open-face) to be used in transport or
storage, the screw cap is carefully removed, and the shiny, waxy, stiff
separator paper used to keep the polycarbonate filters apart is carefully
placed on the retaining ring.  The cap is then carefully screwed back on so
that the separator paper seals and protects the particulate-loaded filter
without touching it.  The 37-mm-diameter, three-piece filter holder (aerosol
monitor) is used in its open-face position, and capped after usage for
transport and storage.

     When the more expensive 47-mm-diameter holder  is to be reused immedi-
ately, the particulate-loaded filter should be carefully removed and placed  in
a 47-ram-diameter Petri-slide (such as that manufactured by the Millipore
Corp.*).  This transfer takes place in the field rather than in the labora-
tory, so that the Petri-slide should be  taken into  the field.  The 37-mm-
diameter filter holder or the 47-mm-diameter holder/Petri-slide should be
secured and all necessary sample identification marks and symbols applied  to
the holder.                                                             .. ••• .--j
            "                                                          ""*••'•.
3.  Sample Preparation for Analysis—Grid Transfer

Carbon-Coating the Filter—                    	   ""'"-''
     The polycarbonate filter, with  the  sample deposit and suitable  blanks,
should be coated with carbon as soon as  possible after sampling is
completed.  To begin this procedure,* the' particulate-loaded  47-mm-diameter
polycarbonate filter is removed from the holder and transferred carefully  to
an open-faced 47-mm-diameter Petri-slide for carbon-coating  in  the vacuum
evaporator (see Figure Al, Appendix  A).  If  the 47-mm-diameter  filter  is
already in the Petri-slide,  the cover is replaced with an open-face  cover,
minimizing filter disruption.  The  37-mm-diameter filter is  left  in  the
holder, but  the upper  lid is removed to  create an openrfaced  filter.   The
open-faced holders are placed on  the rotating  turntable in the  vacuum
evaporator for carbon-coating.   Figure  A2  shows the multiple-coating
 *  Millipore  Corp.,  80-T Ashby Rd.,  Bedford,  Mass.  01730

                                       29

-------
                             I
arrangement in the evaporator; Figure A3 shows a close-up of  the  37-mm-
diameter and the modified 47-mm-diameter holders for carbon-coating.

     For archival filters and those of larger sizes, portions  of  about 2.5  cm
x 2.5 cm should be cut midway between the center and edge using a scalpel.
The portions are then attached with cellophane tape to  a clean glass.
microscope slide and placed on the turntable in the vacuum evaporator  for
coating.

     Any high-vacuum carbon evaporator may  be used to carbon-coat the  filters
(CAUTION:  carbon sputtering devices should not be used).  Typically,  the
electrodes are adjusted  to a height of 10 cm above the  level  of the  filters.
A spectrographically pure carbon electrode  sharpened to a neck of 0.1  cm x
0.5 cm is used as the evaporating electrode.  The sharpened electrode  is
placed in its spring-loaded holder so that  the neck rests against the  flat
surface of a second carbon electrode.

     The manufacturer's  instructions should be followed to obtain a  vacuum  of
about  1.33 x 10~3 Pa (1  x 10~5 torr) in the bell jar of the evaporator.  With
the turntable in motion, the neck of the carbon electrode is  evaporated by
increasing the electrode current to about 15 A in 10 s, followed  by  20 to  25 A
for 25 to 30 s.  If the  turntable is not used during carbon evaporation,  the
particulate matter may not be coated from all sides, resulting in an undesir-
able shadowing effect.   The evaporation should proceed  in a series of  short
bursts until the neck of the electrode is consumed.  Continuous prolonged
evaporation should be avoided, since overheating and consequent degradation of
the polycarbonate filter may occur, impeding the subsequent step  of  dissolving
the filter.  The evaporation process may be observed by viewing the  arc
through welders goggles  (CAUTION:  never look at the arc without  appropriate
eye protection).  Preliminary calculations  show that a  carbon neck of 5 mm3
volume, when evaporated  over a spherical surface  10 cm  in radius,-will yield  a
carbon layer that is 40  nm thick.

  '•/Following carbon-coating, the vacuum chamber is slowly  returned to
ambient .pressure, and the filters are removed and placed  in  their respective
.ho,lde.rs.,or .in, clean, marked Petri dishes for storage on a clean  bench.
 Transfer of the Sample to the EM Grid—
      Transferring  the  collected particulates from the carbon-coated polycar-
 bonate  filter  to an  EM grid  is accomplished  in a clean room or on a class-100
 clean bench.   The  transfer is made in a Jaffe wick washer,  which is usually a
 glass Petri dish containing  a substrate to support the EM grid/carbon-coated
 membrane filter combination.  Solvent is added to a level to just wet the
 combination and cause  gentle dissolution of  the membrane with minimum loss or
 dislocation of the particulates, resulting in a membrane-free EM grid with
 particles embedded in  the carbon film coating.  The substrate support can be
 stainless steel mesh bridges, filter papers, urethane foams, or combinations
 of  these.
                                       30

-------
        The refined Jaffe wick washer is described as follows:

         (1)  The glass Petri dish (diameter, 10 cm; height,  1.5 cm) is
              made airtight by grinding the top edge of the bottom dish
              with the bottom of the cover dish, with water and
              Carborundum* powder (80 mesh); this creates a ground-
              glass seal (closer fit) and minimizes the need  to refill
              the Petri dish with added solvent.  (The usual  glass
              Petri dish was found not to retain the solvent  for long
              periods of time, and unless the wicking substrate is kept
              continuously wet, poor solubility of the membrane filter
              results, leading to a poor-quality EM grid).

         (2)  A combination of foam and a single sheet of 9-cra filter
              paper is used as the substrate support.  A 3-cm x 3-cra x
              0.6-cra piece of polyurethane foam (the packing  in
              Polaroid film boxes) is cut and placed in the bottom
              dish.  A 0.5-inch V-shaped notch is cut into the filter
              paper; the notch is oriented in line with the side of the
              foam, creating a well for adding solvent.
              Spectrographic-grade chloroform (solvent) is poured into
              the Petri dish through the notch until it is level with
              the top of the foam (also level with the paper).  The
              foam will swell, and care is needed to avoid adding
              solvent above the filter paper.

         (3)  On top of the filter paper, pieces of 100-mesh  stainless
              steel screen (0.6 cm x .0.6 cm) are placed, usually in two
              rows, to make several grid transfers at one time (for
              such uses .as replicas), and to facilitate maintenance of
              proper identity of each transfer.

         (4)  A 3-mm section (usually midway between the center and
              edge) of the carbon-coated polycarbonate filter is cut in
              a rocking motion with a scalpel.  The section may be a
              square, rectangle, pr triangle, and should just cover the
              3-mm EM grid.
r - •;'..t-"-"."N (5')  A section  is  laid carbon-side down  on  a  200-mesh carbon-"
              coated EM  grid.  (Alternatively,  Formvar-coatedt grids  or
              uncoated EM grids may  b§  used.  Here,  the  carbon coating
              on the polycarbonate filter'forms the  grid substrate.)
              Minor overlap or underlap of  the  grid  by the filter
              section can be tolerated,  since only the central 2-mm
              portion of the grid is scanned in the  microscope.  The  EM
              grid and filter combination is picked  up at the edges
              with the tweezers and  carefully laid on  the damp 100-mesh
   *  Carborundum  is  a  registered  trademark of  the Carborundum Co., Carborundum
      Center,  Niagara Falls,  N.Y.  14302.

   t  Formvar  is a registered trademark of  the  Monsanto Company, 800 N. Lindbergh
      Blvd., St. Louis,  Mo.

                                          31

-------
           stainless steel screen.  The EM grid-filter combination
           will immediately "wet out" and remain on the screen.

      (6)   Once all specimens are placed in the washer, more solvent
           is carefully added through the notch to maintain the
           liquid level so that it just touches the top of the paper
           filter.  Raising the solvent level any higher may float
           the EM grid off the mesh or displace the polycarbonate
           filter section.

      (7)   The cover is placed in the washer and oriented in place
           over the specimen, and a map of the filter/grid/screen
           arrangement is made on the glass cover and in the
           logbook.

      (8)   Solvent (chloroform) is added periodically to maintain
           the level within the washer until the filter is
           completely dissolved by the wicking action (24 to 48 h).

      (9)   The temperature in the room must remain relatively
           constant to minimize condensation of solvent on the
           bottom of the cover and subsequent falling of solvent
           drops on the EM grid.  Should day-night or other
           temperature differentials occur, solvent condensation on
           the under-surface of the cover can be minimized by
           placing the Jaffe washer at a slight tilt (three glass
           slides under one edge of the Petri dish parallel to the
           row of grids) to allow the condensation drops to flow
           toward the lower edge rather than fall on the EM grids.
           At temperatures lower than 20°C (68°F), the complete
           filter solution may take longer than 72 h.
     (10)   After- the polymer is completely dissolved, the stainless
           steel mesh screen with the EM grid is picked up while wet
           and set on lens paper tacked to the bottom of a separate
           Petri dish.  The EM grid is then lifted from and placed
           next to the screen to dry.  When all traces of solvent
           have evaporated, the grid is stored in a grid box and
           identified by location and grid box in the logbook.

     Figure A4 illustrates the Jaffe wick washer method; Figure A5  shows  the
washer.  The foam/filter combination" is currently preferred, as is  use  of  a
closely fitted (by means of the ground-glass seal) Petri dish.

Gold Coating—
     An additional step will aid in subjectively evaluating the SAED  pattern.
This step is required for specimens from the upper Great Lakes area and for
those of unknown origins.  After the particulates on the filter are trans-
ferred to the EM grid, the grid is held to a glass slide with  double-stick
tape for gold-coating in  the vacuum evaporator.  Several EM grids may be  taped
to the glass slide with double-stick tape  for gold-coating in  the vacuum  evap-
orator.  For comparison,  one-half of the EM grids may  be coated and the other
one-half not coated; recognition of the gold-coating is helpful in  searching
and x-ray analysis.  Several EM grids may  be  taped to  the glass slide for

                                       32

-------
coating at one time.  Approximately  10 mm of  0.015-cm-diaraeter (0.006-inch)
pure gold wire is placed in a tungsten basket  (10  cm  from the  rotating table
holding the EM grids) and evaporated onto the  grid.

     The thin gold-coating establishes an internal standard  for SAED analysis.
For some mineral species, an internal standard will clarify  visual identifi-
cation of the pattern of a fibrous particulate as  being  or not being an
amphibole species (for example, minnesotaite  as opposed  to amosite).  With
experience, differentiation in SAED  patterns  can be observed.   For samples of
known geographic origins, gold-coating is optional, since the  additional
coating hinders observation and identification of  small-diameter chrysotile
fibers.

4.  TEH Examination  and Data Collection
     Figure A10  shows  a modern  TEM jwi_th_capabilities for elemental., arialysis^
wiih__an_EDS..  The grid is observed in  the  TEM""a"f"magnifications of ^50X and
1000X to determine  its suitability for detailed study at higher magnifica-
tion.  The grid  is  rejected and a  new  grid used if:   (1) the carbon film over
a majority of the grid openings is damaged and  not intact;  (2) the specimen is
dark due to incomplete dissolution of  the  polycarbonate filter; or (3) the
particulate loading is too light (unless a blank)  or too heavy with particle-
particle interactions  or overlaps.

TEM Analysis  (Morphology, SAED, and X-Ray  Analysis)—
     The following  guidelines are  observed for  consistency in the analytical
protocol:

      (1)  Magnification at the fluorescent screen is determined by
           calibration with a diffraction-grating  replica in the
           specimen holder.

      (2)  A  field  of  view or "gate" is defined.'  On some                 ....  .,
           microscopes,  the central rectangular portion of the              .-.^-.?'
           fluorescent screen,  which is lifted  for photographic
... ,:,',-., 	.purposes, is  convenient, to  use.  On  others, a scribed            „ ";
  ; ;-       circle or the entire circular screen may be used as the ' "  " ' r;;iu-;'"
           field of view.  The  area of the field of view must be
           accurately measurable.      '     -•••--    	
       (3)   The  grid opening is selected on a random basis.'

       (4)   The  analysis,  morphology,.and SAED are performed at a
            tilt angle of  0".              -- --  -.•...: =	

       (5)   The  recommended instrument settings"are:  accelerating
            voltage, 100 kV; beam current,_100_yA; film.magnifi-
            cation, 20.000X (which is equivalent to U.OOOX on' the
            fluorescent screen for this instrument); and concentric
            circles of radii 1, 2, 3, and A cm on the fluorescent
            screen.
                                       33

-------
(6)  The grid opening is measured at low magnification (about
     1000X).

(7)  Since asbestos fibers are found isolated as well as with
     each other or with other particles in varying arrange-
     ments, the fibrous particulates are characterized as
     asbestos structures:

     Fiber (F) is a particle with an aspect ratio of 3:1 or
     greater with substantially parallel sides.

     Bundle (B) is a particulate composed of fibers in a
     parallel arrangement, with each fiber closer than the
     diameter of one fiber.
     Cluster (Cl) is a particulate with fibers in a random
     arrangement such that all fibers are intermixed and no
     single fiber is isolated from the group.

     Matrix is a fiber or fibers with one end free and the
     other end embedded or hidden by a particulate.

     Combinations of structures, such as matrix and cluster,
     matrix and bundle, or bundle and cluster, are categorized
     by the dominant fiber quality—cluster, bundle, and
     matrix.

(8)  Counting rules for single fibers, which are illustrated
     in Figure A7 are as follows:

     (a)  Particulates meeting the definition of fiber are
          isolated by themselves.  With this definition, edge
          view of flakes, fragments from cleavage planes, and
          scrolls, for example, may be counted as fibers.
     (b)  Count as single entities if separation is equal to
          or greater than the diameter of a single fiber.

   1 ' (c)  Count as single entities if three ends can be  seen.
     (d)  Count as single entities if four ends can be seen.

'•'   ' (e)  In general, fibers that touch or cross are counted
          separately.
     (f)  Two or more fibers-are counted as a bundle if  the
          distances between  fibers are less than the diameter
          of a single fiber, or if the ends cannot be
          resolved.

     (g)  Fibrils attached longitudinally to a  fiber are
          counted as part of the fiber and the  size (width)  is
          estimated based on the fiber-to-fibril relationship.

-------
      (h)   A fiber partially hidden by grid .wires (one or two
           sides of the grid opening) is counted, but labeled
           as an X-fiber (X-F)  in the structure column.  If the
           number of X-fibers is high enough to affect the size
           distribution (mass,  etc.), a large-mesh EM grid
           should be used, such as 100 mesh (about 200 ym
           wide).

 (9)  Sizing rules for asbestos structures are:

      (a)   For fibers, widths and lengths are obtained by
           orienting the fibers to the inscribed circles on the
           fluorescent screen.   Since estimates are within
           ±1 mm, small-diameter fibers have greater margins of
           error.  Fibers less  than 1 mm at the fluorescent
           screen magnification level are characterized as
           being 1 mm.  A cylindrical shape is assumed for
           fibers.  X-fibers are sized by measuring their
           entire visible portions in the grid opening.

      (b)   Bundles and clusters are sized by estimating their
           widths and lengths.   The sum of individual diameters
           is used to obtain the total width, and an average
           length for the total length.  A laminar-sheet shape
           is assumed, with the average diameter of the
           individual fiber as  the thickness.
      (c)   Matrices are sized by adding the best estimates of
           individual fiber components.  A laminar or sheet
           structure is assumed for volume calculation.

(10)  The  method of sizing is as follows:

      (a)  "An asbestos structure is recognized, and its
           location in the rectangular "gate" relative to the
           sides, inscribed circles, and other particulates is
       -,'.  memorized.

      (b)   The structure is moved to the center for SAED
      .-•.:.• -observation and .sizing.                            „
       "                                            •—••-...,J*
      (c)  Sizing is performed using the inscribed circles.   If:;
           the structure, such as a fiber, extends beyond the
           rectangular gate (field of view), it is superimposed
           across the series of concentric circles (several
           times, if necessary) until the entire structure is
           measured.
      (d)  The structure is returned to its original  location
           by recall of the location, and scanning is
           continued.
                                 35

-------
Analytical Procedure—
     The analytical procedure is as follows:

      (1)  EM grid quality is assessed at 250X.

      (2)  Parciculate loading is assessed at 1000X.
      (3)  A grid opening is selected at random, examined at 1000X,
           and sized.
      (A)  A series of parallel traverses is made across the grid
           opening at the film magnification of 20.000X.  Starting
           at one corner, and using the tilting section of the
           fluorescent screen as a "gate" or "chute," the grid
           opening is traversed.  Movement through the "gate" is not
           continuous, but rather is a stop/go motion.  On reaching
           the end of one traverse, the image is moved the width of
           one "gate," and the traverse i's reversed.  These parallel
           traverses are made until the entire grid opening has been
           scanned.
      (5)  Asbestos structures are identified morphologically and
           counted as they enter the "gate."
      (6)  The asbestos structure is categorized as fiber (with or
           without X-) bundle, cluster, or matrix, and sized through
           use of the inscribed circles.

      (7)  The structure (individual fiber portion) is centered and
           focused, and the SAED pattern is obtained  through use of
           the field-limiting aperture.

           (a)  SAED patterns from single fibers of asbestos
                minerals fall into distinct groups.   The chrysotile
                ^asbestos pattern has characteristic streaks on layer
                lines other than the central line, and some
                streaking also on the central line.   Spots of normal
                sharpness are present on the central  layer line and
                on alternate lines (that is, 2nd, 4th etc.)  The   ''•-.'
                repeat distance between layer lines is about 0.53 nra.
           (b)  Amphibole asbestos fiber patterns show layer lines  -.
                formed by very closely spaced dots, and have repeat.;
                distances between layer lines also of about 0.53 nm.
                Streaking in layer lines is occasionally present due
                to crystal structure defects.
           (c)  Transmission electron micrographs and SAED patterns
                obtained with asbestos standard samples should be
                used as guides to fiber identification.  An example
                is the "Asbestos Fiber Atlas"  (Mueller et al.,
                1975).
                                       36

-------
      (8)   From visual examination of the SAED pattern, the struc-
           ture is classified as belonging to one of four cate-
           gories:  (1) chrysotile, (2) amphibole group (includes
           araosite, crocidolite, anthophyllite, tremolite, and
           actinolite), (3) ambiguous (incomplete spot patterns), or^
           (4)  no identification.  SAED patterns cannot be inspected
          •for  some fibers.  Reasons for the absence of a recog-
          ,nizable diffraction pattern include contamination of the
           fiber, interference from nearby particles, fibers that
           are  too small or too thick, and nonsuitable orientation
          'of the fiber.  Some chrysotile fibers are destroyed in
           .the  electron beam, resulting in patterns that fade away
           within seconds of being formed.  Some patterns are very
           faint and can be seen only under the binocular micro-
           scope.  In general, the shortest available camera length
           must be used, and the objective lens current may need to
           be adjusted to give optimum pattern visibility for
           correct identification.  A 20-cm camera length and a 10X
           binocular are recommended for inspecting the SAED pattern
           on the tilted screen.

      (9)   The  specimen holder is tilted for optimum x-ray detection
           (40° tilt for the JEOL* 100C instrument's Tracer
           Northern! NS 880 analyzer and Kevex* detector).  The
           categorized asbestos structure is maintained in its
           centered position for x-ray analysis by means of the Z-
           control.
     (10)
     (11)
     (12)
The spot size of the electron beam  is  reduced  and
stigmated to overlap the fiber.  As an option  for  STEM
instruments, the electron beam may  be  used  in  the  spot
mode 'and the x-ray analysis performed  on  a  small area  of
the structure.
• •«
The EDS is used to obtain a spectrum of the x-rays~~f
generated by the asbestos structure.                \  '">""'"
=-                                                  —**    i'; -".":
The profile of the spectrum is compared with prof ilesl ..VT^.-;
obtained from .asbestos  standards; the  best  (closest)   .;v<';  ,
match identifies and categorizes the structure.  The""    '"'''
image of the spectrum may be photographed,  or  the  peak
heights (Na, Mg, Si, Ca.'Fe) recorded  for normalizing  at
a later time.  No background spectra or constant acquisi-
tion time is required since the  shape  of the spectrum
(profile) is the criteria.  Acquisition of  x-ray  counts
may be to a constant time;  to a  constant peak  height  for
a selected element, such as silicon (1.74 keV);  or just
                                                                       !* !
* JEOL (U.S.A.) Inc.,  11 Dearborn Road, Peabody, Mass.   01960

t Tracor Northern Inc., 2551-T.W. Beltway Hwy., Middleton,  Wis.  53562

* Kevex Corp., Chess Dr., Foster City, Calif.  94404

                                       37

-------
           long enough to get an adequate idea of the profile of the
           spectra, and then aborted.  Figure All illustrates
           spectra obtained from various asbestos standards and used
           as referenced profiles.

     (13)  The specimen holder is returned to 0° tilt to examine
           other asbestos structures.

     (14)  Scanning is continued until all structures are
           identified, measured, analyzed, and categorized in the
           grid opening.

     (15)  Additional grid openings are selected, scanned, and
           counted until either the total number of structures
           counted exceeds 100 per known area, or a minimum of  10
           grid openings has been scanned, whichever is first.

     (16)  The TEM data should be recorded in a systematic form so
           that they can be processed rapidly.  Sample information,
           instrument parameters, and the sequence of operations
           should be tabulated for ease in data reduction and
           subsequent reporting of results.  Figure A12 shows an
           example of a data sheet used in Level II analysis.

     Figure A9 illustrates the method of scanning a full-grid opening.  The
"field of view" method of counting, which is based on randomly  selected fields
of view, has been discontinued.  Originally, the method was recommended for
medium loading level on the filter (50 to 300 fibers per grid opening).   How-
ever, if samples are collected at three different loading levels and  the  opti-
mum is selected, this medium loading on the filter will not be  used.   Samples
with grid openings containing 50 to 300 fibers may be used as laboratory  fiber
preparations or selected source samples, but in field samples,  the  particulate
loading is usually of much higher concentration than the fiber.  Filter load-
ing is characterized by particulate concentration, not by fiber concentration.

     EDS is relatively time-consuming, and becomes redundant, if. used. as"-. " |
repetitive analysis for a confirmatory check on chrysotile fibers.   Chrysotile
identity by morphology., and visual SAED analysis is not as controversiai^as :
amphibole identification and categorization.               •'."."'	     ;'*"'.   .  ' "'

     The following rules are recommended for EDS analysis (Level II):.
                                        •,

     (1)  For chrysotile structure identification, the first  five
          are analyzed by EDS,  then one out of every  10.
     (2)  For amphibole structure identification, the first  10  are
          analyzed by  EDS,  then one out of every  10.
     (3)  For amphibole structure identification and categorization,
          all confirmed amphiboles are analyzed by EDS.

     (4)  For ambiguous structure identification and categorization,
          all are  analyzed  by EDS.
                                       38

-------
     Energy dispersive x-ray analysis as used in asbestos analysis  is
seraiquantitative at best.  X-ray analyzer manufacturers may  claim quantitative
results based on calibration standards and sophisticated computer software,
but such claims are based on stoichioraetric materials and extension of  work
with XRF instrumentation.  Asbestos has a varying elemental  composition.   The
electron beam in an EM is of varying size, and not all instruments  are
equipped to measure the beam current hitting  the specimen.   The  size of the      :
"specimen has an effect on the x-ray output, and nearby materials may fluoresce
and add to the overall x-ray signals being generated.  Moreover, specimen
tilting results in a loss of x-ray acquisition from  particles  hidden by grid
wires or by other particles.
••«i
     The only consistency in x-ray analysis is that  the intensity of the
output, within restrictions, is proportional  to the  mass, therefore providing
the semiquantitative analytical possibility.  Asbestos minerals  .have been
found to have a characteristic profile, although not an exact  duplicate of
each other.  For example, the Mg:Si ratio of  chrysotile may  vary from 5:10 to
10:10, averaging about 7:10.  The ratio can be used  to confirm the  morphology j
and visual SAED analysis.                                                     "
^—

     Table 1 illustrates the phenomena of variability with  resemblance  for
some of the amphibole fibers.  Peak heights and profile measurements were
taken.

     To aid in the visual perspective of  the  spectrum profile, the  peak
heights were normalized to a silicon value of 10, resulting  in a five-number
series that is relatively easy to visualize—as in  the following examples:

           chrysotile          ~ 0-7-10-0-0

           treraolite           ~ 0-4-10-3-
-------
TABLE 1.  PROFILE COMPARISON OF ASBESTOS STANDARDS
Asbestos Type
Amoslte (GF-38A)








Anthophylllte (AF-45)




Crocldollte (CR-37)











,. •* '.T • '(- i-' • _'•'
Treoollte (T-79)











Size, u
0.19 x 1 .44 (stlpnate)
0.19 x 0.75 (STE»)
0.19 x 1.25
0.19 x 0.88 (100 s)
0.25 x 1.81 (100 s)
0.12 x 1.56
0.31 x 2.38
0.19 x 1.56
Repeat
0.56 x 2.38 (stlgmate)
0.31 x 2.38 (stlgmate)
0.31 x 5.19 (stlgmate)
0.19 x 1.56 (stipnate)
0.19 x 1 .88 (stlgmate)
0.19 x 0.81 (stleroate)
0.06 x 0.50 (stignate)
0.06 x 0.69 (stlpnate)
0.12 x 1 .00 (stlgmate)
Repeat (STEM)
0.12 x 0.62 (stlgaate)
0.12 x 1.12 (self-mate)
0.19 x 1 .56 (stlpraate)
0.06 x 1.69 (stlgmate)
Repeat (STEM)
Repeat (STEM)
.Repeat (STEM)

0.38 x 2.19 (stlpmate)
0.38 x 2.19 (spot) v
0.25 x 1.75 (stipmate)
0.25 x 1.75 (spot)
Repeat (stiRmate)
(STE«-inn S)
(STEM-inO s)
( STEM- 1 no s)
(STEM-100 s)
(STEM-100 s)
(STEM-40 s)
(STEM-40 s)
Na Kg
182
18ft
18]
226
57ft
253
256
27ft
• 477
631
64n
1064
507
787
131 inn
28 2R
37 35
44 53
70 64
56 65
53 5ft
76 83
45 4R
72 85
35 42
16 22

138
114
an
95
70
37ft
135
14S4
64
1072
4ft
12.1
Si
497
521
352
870
42n7
2049
2127
169ft
2945
2577
1670
3ftlO
2191
2286
885
205
17]
37«
612
479
326
735
290
892
373
166

368
327
197
252
211
1118
364
4Rin
1«1
3114
113
333
Ca Fe
38ft
387
289
674
3338
1515
1M3
lllft
1959
349
71
46ft
309
257
50,
115
9ft
204
333
2ftO
16ft
421
159
.463. .
237 .
104

03 ' - '\
8"
f>5
62
M
245
' 72
1235
48
BR2
27
41
Profile
0-4-10-0-R
0-4-]0-n-7
n-s- 10-0-1
p_3_,n_n-c
n-l-io-o-fi
p. j_j 0-0-7
o-i-in-n-8
0-2-10-0-7
0-2-10-0-7
0-2-10-0-1
n-4-m-p-n
o-3-m-o-i
o-2-in-o-i
O-3-in-n-i
2-1-10-O-ft
i-i-io-n-ft
2-2-in-n-ft
l-l-m-o-5
1-,-io-n.',
1-1-10-0-5
2-2-10-0-5
l-l-l o-n-f,
2-2-in-n-f,
l-l-in-O-5
l-l-!0-n-ft'.':.. ' '
i-i-io-n-6 , ., .

:: ,.$',*•- •;
o-i-io-2-n
O-i-jn-3-o
p-4-]o-2-n
n-3-10-;-n
i-3-in-2-n
n_i-m-2-n
n-3-in-3-n
n-3.]n-2-.T
O-l-in-3-n
n-i-io-:-o
ft-i-in-i-n

-------
(area of deposit).  Size measurements of X-fibers may be doubled  and  noted,  or
kept as a separate category.

     Fiber number concentration is calculated from the equation


            Fibers/m3  =  Total no. of fibers
                           No. of EM fields

                            Total effective filter area, cm2
                          X     Area of an EM field, cm2

                                        1
                            Volume of air sampled, m3
The number of X-fibers, bundles, clusters, and matrices are calculated  in  a
similar manner.  X-fibers may be included with fibers if they are  few in
number.  Similarly, their corresponding mass (from their size measurements)
may be included.

     Fiber mass for each type of asbestos (chrysotile or amphibole)  in  the
sample is calculated by assuming that both chrysotiles and amphiboles have
circular cross-sections (cylindrical shape) and that the width measurements
are one diameter.  The density of chrysotile is assumed to be 2.6  g/cm3, and
of amphiboles to be 3.0 g/cm3.  The individual mass is calculated  from  the
equation


                Mass,  ug  = -7 x (length,  vm)  x (diameter,  utn)2
                •T             1

                            x (density, g/cm3) x  10  6


The total mass concentration of fibers for each type of asbestos is  then
calculated, .from .the. total .mass.of all the individual ..fibers of that  type.  , ... M

    " The individual masses of bundles, clusters, and matrices are  calculated
by assuming a laminar or sheet-likexstructure with an average thickness of the
fiber make-up of the structure.  Again, the density of chrysotile  is assumed
to be 2.6 g/cm3, and of amphiboles to be 3.0 g/cm3.  The individual  masses are
calculated from the equation


           Mass, u§  =  (length, ym)  x (width,  um)  x  (thickness,  um)

                        x  (density,  g/cm3)  x 10~6
The total mass for each type of structure for each type of asbestos  is  the  sun
of all the individual masses.

-------
     Other characterizing parameters of the asbestos structures are:  (1)
length and width distribution of fibers, (2) aspect ratio distribution of
fibers, and (3) relationships of fibers, bundles, clusters, and matrices.

Reporting of Results—
     The data and their subsequent reduction are reported as summarized, or
can be further reduced to present the interrelationships of the various
characterizing parameters.  Figure A13 is an example of the EM data report;
Figure A14 is an example of the sample summary report.

     The methodology can establish the limits of identity for unknown samples,
act as a QC/QA method for Level I analysis, and satisfy most of the
identification criteria for asbestos.

6.  Quality Control/Quality Assurance

     Sampling procedures will vary depending on the type of sample, objectives
of the sampling, and time/cost factors.  The primary goals of sampling are  to
obtain a representative sample at the location and time of sampling, and to
maintain sample integrity.  The sampling team will have written sampling
procedures, and the field chief and/or designated individual will be respon-
sible for all record-keeping (including sample identification, labeling,
logging of data, site description, and meteorological conditions),  pre- and
post-collection checks, and continuous sample custody and sign-outs until  the
sample is delivered to the laboratory and  transferred to the appropriate
quality assurance officer (QAO).  Verification of sampling times, flow rates,
equipment calibration, and taking of field  blanks will be checked and recorded
in the field logbook.

     Samples are turned over to the QAO for logging into a project  logbook.
Each sample is carefully examined for gross features, such as tears, breaks,
and overall condition of container.  The QAO registers the as-received sample
number and other designated information, and assigns a simple internal code
number-that will accompany the  sample through the preparation stage, grid   --*-.
transfer, grid analysis, data reduction, and reporting of results.

     After 'being logged' into the project logbook, the sample is "transferred do
the custody of the electron microscopy  staff, where every precaution is  taken..
to maintain sample integrity and to prevent contamination and loss  of
collected particulates.   During storage,and transport, the filters  in  their
respective holders are maintained in a  horizontal position at all  times.

     The sample logging, handling, and  storing  procedures ensure  that  all
samples can be readily located  and identified throughout  the course of a
program.  The  QAO has divisional responsibility for QC/QA .activities,  and  must
see  that  the  laboratory maintains high  standards.  He must be aware of  current
standards of analysis, and must ensure  that internal  quality control
standards, instrument calibration, and  records  of  samples  and completed
analyses  are kept for ease of  later  retrieval and use.

-------
     For quality control, internal laboratory blanks are analyzed at least
once a week, which may or may not coincide with a sample batch blank.  In
addition, a magnification calibration of the EM using a carbon grating replica
(2,160 lines per mm) is performed once a week.  The results are recorded in an
EM instrument log, along with other routine instrumental performance checks.
All photographs, TEM, SEM, and STEM images are recorded in a photo log.  These
QC results are documented for inspection by the QAO.

-------
                                   SECTION 6

                          LEVEL III ASBESTOS ANALYSIS

DISCUSSION OF PROTOCOL
                                                                              ^
     The Level III protocol is an extension of the Level II analysis  proce-
dures described in Section 5.  This extension may be necessitated by  the need
for positive identification of. the specific amphibole species  in situations
where (1) fundamental disagreements between parties  involved in a litigation
require further clarification; (2) for identification purposes, e.g.,  as
causative agents in medical diagnosis or studies; (3) for quality control  of
Level II analysis in special situations, and/or; (A) for source samples
whether as bulk material or bulk-air type where a legal judgment is antici-
pated.

     Since an SAED pattern may be considered as a signature of  the crystal
structure of the diffracting crystal (mineral fiber  or particulate),  the
mineral giving the pattern can be identified by comparison of  measured and
standard sets of d-spacings and interplanar angles (0) from SAED patterns
obtained in near-exact zone axis orientations.  Such identification,  however,
may not be absolute without the provision of SAED patterns from more  than  one
zone-axis orientation.

     The Level III analysis is an objective, confirmatory-type analysis  and
consists of Level II analysis plus quantitative SAED analysis  from two
different near-exact zone-axis orientations on a selected number of  fibers
identified for detailed SAED analysis during the course of Level II  analysis.

     The Level III analytical procedure consists of  locating  the selected
fibers contained in gold-coated grid openings (for internal calibration);
photographing the fibers under bright-fsld  illumination; obtaining  (by
tilting) and recording two zone-axis SAED patterns from each  selected fiber;
and obtaining (recording and photographing) representative EDS spectra from
the subject fiber.

     The present Level III protocol is based on the  following  guidelines:

     (1)  Maintenance of procedural continuity so that results of
          Level  II analytical  effort will aid in conducting  the Level
          III effort.

     (2)  Since  detailed  SAED  analysis on all the fibers measured  in
          Level  II analysis is not possible due to time and  cost
          restraints, a selection  criterion  is needed  to assure
          representative analyses.


                                       44

-------
(3)  The primary emphasis in Level III analysis is on the
     positive identification of the amphibole type.

(4)  The present protocol is designed to allow greater flexibil-
     ity and freedom of decision for the microscopist in deter-
     mining the selection criteria since, due to practical
     constraints (position, orientation, contamination, etc.)i
     all fibrous particulates may not be suitable for detailed
     SAED work.

(5)  It is recommended that approximately 20% (at least 10%) of
     the fibers examined in Level II analysis be selected for
     Level III SAED analysis.  Fibers which would be classified
     as "amphiboles" or "ambiguous" in Level II analysis should
     be more often included for Level III analysis as compared
     to those fibers which could be readily identified as "not
     asbestos."  In cases where the majority of the fibers in
     Level II belong to a single, easily-identifiable species
     (e.g., chrysotile), fibers that are different should be
     more often selected for detailed Level III analysis.  This
     flexibility in selection criteria will maximize the gain
     (meaningful information) from Level III effort beyond what
     would be achieved from the analysis of 10-20% randomly
     selected fibrous particulates.

(6)  The electron microscope grids used in Level III analysis
     (also Level II if Level III is anticipated) should be
     finder grids so that location of fibers examined could be
     referenced for quantitative SAED and for future re-
     checking.

(7)  Level III analysis should always be conducted by or under
     the close supervision of a professional electron micro-
     scopist knowledgeable in crystallography, SAED analysis,
     mineralogy, plus Level I and Level II asbestos analyses.
     If such expertise is not available in-house, an outside
     consultant should be retained.

(8)  If enforcement proceedings and possible legal involvement
     may be part of the analytical procedure, the sample collec-
     tion procedure entails additional record-keeping to
     maintain sample integrity.  The field crew chief or a
     designated individual initiates, in addition to normal
     QC/QA activities, a chain-of-custody record.  The sample is
     collected by the field team or by a representative of  the
     adversary party in the presence of each other, and is
     sealed and signed for with the date and time.  The desig-
     nated individual acknowledges receipt of the collected
     sample.  In transferring the sample, the designate signs a
     release of the sample in the presence of the new recipient,
     who notes the date and time, and signs for acceptance  in
     the designate's presence.  The chain of custody ensures
     that only responsible pers-nnel have access to and control
     of the sample, thereby avoiding the possibility of


                                  45

-------
          contamination before and after transport to the labora-
          tory.  At the laboratory, the QAO has first access to the
          sealed sample container, and signs for it after obtaining a
          signed release by the hand-carrier.

SUMMARY OF PROTOCOL

      (1)  An EM grid is prepared as directed in Level II analysis
           using finder or locator grids instead of regular 200-mesh
           grids.

      (2)  The particulate-loaded grid is then one-half or completely
           coated with a thin layer of gold.

      (3)  The gold-coated grid is placed in a tilt-rotation or a
           double-tilt specimen holder, and examined in the AEM or
           STEM.

      (A)  At low magnification the specimen grid is examined, and a
           grid opening is selected and identified for reference.

      (5)  Fibers identified for detailed Level III SAED work during
           Level II analysis, employing the selection criteria
           described under Level III guidelines, are now examined one
           at a time.
      (6)  A bright-field image of the fiber is taken at 0° tilt and
           at the magnification of analysis (20.000X).

      (7)  With the tilt-rotation or double-tilt combination, well-
           defined SAED patterns of two different zone-axis orienta-
           tions are observed and photographed.  The fiber location
           with respect to the edges of the grid opening or to other
           particulates may prevent more than one zone-axis orienta-
           tion from being obtained for some fibers.

      (8)  X-ray elemental analysis is taken of the fiber after  the
           SAED patterns.  The EDS analysis also may be affected by
           proximity of the fiber  to  the edge of the grid opening or
           to other particles if tilting of  the specimen is required
           for efficient use of  the EDS.  An image of  the spectra is
           taken along with a record of the peak heights (the
           presence of grid peaks, such as  Cu or Ni, as well as  gold-
           coating may serve as markers).

      (9)  As  explained earlier, due  to time and cost  considerations,
           at  least  10% (preferably 20%) of  the fibers examined  in
           Level II are analyzed in Level  III work.

      (10)  Those fibers whose EDS  elemental  analysis points  to  a
           possible amphibole identification are  selected  for  SAED
           pattern indexing.

-------
     (11)  Parameters of interest obtained form zone-axis SAED
           patterns are:  the camera constant, CC (obtained from the
           gold ring); the diffraction spot spacing dj (along the
           slant vector), dj (along a row); the inter-row spacing, R;
           and the interplanar angle 61,2.  See Figure A18 for
           details.

     (12)  The reciprocal lattice values of the d-spacings dj and d2
           and the inter-row spacing (R) are converted into direct
           lattice spacings and then dj, d2, R, and 61,2 are compared
           to those of standard amphibole species listed in JCPDS
           Powder Diffraction Files, values computed from lattice
           parameters and crystal structures, or SAED Standard
           Pattern File developed internally from known amphibole
           minerals regulated by EPA.
EQUIPMENT, FACILITIES, AND SUPPLIES

     Essential items required for a Level III analysis are:

       •  A 100-kV AEM equipped with the fluorescent viewing screen
          inscribed with graduations of known radii to estimate the
          lengths and widths of fibrous particulates; or a modern
          100-kV TEM equipped with an EDS.  A scanning accessory as
          found in an STEM will increase the versatility and
          analytical capability for very small fibers or for fibers
          adjacent to other particulate matter.  This microscope
          should also be equipped with the fluorescent viewing
          screen inscribed with graduations of known radii to
          estimate the lengths and widths of fibrous particulates.

       •  A specimen holder with tilt-rotation or double-tilt
          capability to obtain diffraction patterns at different
          zone-axis orientations.

       •  Darkroom facilities for developing negatives, making
          enlarged prints of patterns, and facilitating measurement
          of distances, spots, lines, and circles.

       •  A vacuum evaporator with a turntable for rotating
          specimens during coating, 'for such uses as carbon-coating
          polycarbonate filters, gold-coating EM grids, and
          preparing carbon-coated EM grids.

       •  An EM preparation room adjacent to the room housing  the
          EM.  This room should either be a clean-room facility, or
          contain a laminar-flow class-100 clean bench to minimize
          contamination during EM grid preparation.  Filter  handling
          and  transfer to EM grids should be performed in a  clean
          atmosphere.  Laboratory blanks  should be prepared  and
          analyzed weekly to ensure quality of the work.  In addi-
          tion, a sample preparation room with a laminar-flow  class-
          100  clean bench should be available for handling bulk-air


                                      47

-------
          samples, ashing procedures, sedimentation, ultrasonifi-
          cation, filtration, and other prefilter activities.

       •  Several modifed Jaffe wick washers for dissolving membrane
          filters.

       •  Miscellaneous supplies and chemicals, such as membrane
          filters, EM grids, films, gold wire, chloroform, and
          carbon rods.

       •  Sample collection equipment, such as filter holders,
          sampling pumps, critical orifices, and tripods.

DESCRIPTION OF METHODOLOGY

     A detailed discusson of the morphology, crystallography and chemistry of
asbestos minerals, electron microscopy, and SAED analysis is outside the scope
of the present protocol.  Basic knowledge in these areas and an adequate level
of comprehensive knowledge of TEM and SAED are prerequisites for the micro-
scopists participating in asbestos analysis, especially at Level III stage.

     Since Level III analysis is an extension of Level II analysis, common
methodological details dealing with type of samples (source), sample collec-
tion and transport, sample preparation, TEM examination and data collection,
data reduction and reporting of results, and quality control/quality assurance
(QC/QA) program, which were discussed in detail in Section 5 (Level II
Asbestos Analysis) will not be repeated here and the users are advised  to
refer to Section 5 for details in these areas.  Differences, if any, between
Level II and Level III protocols in common areas have been dealt with earlier
under "Guidelines" and "Summary of Protocol."

     The following provides brief descriptions of some of the essential areas
of the Level III protocol that were not covered under Level II protocol.

1.  Crystallography and Morphological Properties

     Both crystallographic and morphological characteristics of asbestos
minerals can help considerably in asbestos indentification and analysis.
Chrysotile displays a unique narrow tubular morphology.  The amphibole
asbestos minerals have very  similar morphologies—they are elongated along  the
z-axis (the chain direction) and generally lie with (100) planes approximately
perpendicular to  the electron beam.  All varieties of amphiboles exhibit  these
Wadsley faults parallel to the length of the fiber.

     Chrysotile  possesses a  cylindrical lattice which produces a unique SAED
pattern.  All the amphiboles, except anthophyllite, which is orthorhombic,
have a monoclinic crystal structure.  The amphiboles are double-chain
silicates in which the  fiber axis,  z, has a repeat of 0.53 nm  (inter-row
spacing  'R1 in real space, Figure  A18).  Since  the other  lattice parameters
are also very similar,  detailed  zone-axis SAED analysis  in more  than one
orientation is needed  for positive  identification.  The  non-asbestos  forms  of
amphiboles have  properties very  similar to  their asbestos counterparts, thus
they must be distinquished  from  asbestos on the  basis of morphology alone.


                                      48

-------
2.  Chemical Properties—Elemental Analysis by EDS

     Araphiboles are nonstochioraetric minerals and often contain substitutional
cations in varying amounts.  Therefore, precise determination of their chem-
istry is difficult and positive identification based on chemistry alone is not
reliable.  This may be particularly pertinent when dealing with asbestos
minerals present as minor constituents in mineral samples.

     Elemental ratios, which are sometimes used to distinquish between
asbestos types, often vary over wide ranges even in standard samples.  The
presence of gold coating, which would tend to preferentially absorb x-rays
from lighter elements more than heavier elements, may make the situation even
worse.  In view of these ambiguities, and due to inherent practical difficul-
ties in obtaining representative quantitative EDS elemental analyses from
submicroscopic fibers, the present Level II and Level III protocols specify
the use of only qualitative EDS spectra, which are often very valuable for
screening purposes in the Identification procedure.  For example, in distin-
guishing between tremolite and actinolite type of amphibole, actinolite
usually contains Fe, but tremolite does not.

3.  Selected Area Electron Diffraction (SAED)

     The method of obtaining an SAED pattern of a randomly oriented specimen
is usually described in the EM instruction manual.  The general directions  for
using the instrument to obtain and photograph SAED patterns are:

      (1)  Select the image magnification for the selected area.

      (2)  Bring the desired field of view to the center of the
           screen.
      (3)  Insert the appropriate field-limiting aperture  (according
           to  the desired field of view) into the beam path.

      (A)  Obtain the sharpest field-limiting aperture shadow.

      (5)  Confirm that the desired field of view is in the field-
           limiting aperture.
      (6)  Focus the specimen  image; a photograph of the  selected
           area image can be taken.

      (7)  Obtain the SAED pattern, remembering  to retract  the
           objective lens aperture  from  the beam path.  The SAED
           pattern will be observed on the  fluorescent screen.

      (8)  Select the desired  camera length  (the shorter  the  length,
           the better  for  SAED patterns  of  asbestos  taken  at  high   -.-..
           magnification).
      (9)  Focus  the  SAED  pattern  sharply.   The  beam  stopper  is  used
           to  intercept  the  bright  center  spot.

-------
     (10)  For photography, the illumination is expanded (condenser.
           reduced) after focusing the pattern, so that the pattern
           becomes barely visible (indistinct).  A manual time
           exposure of approximately 20 to 30 s (maybe more
           depending on such factors as specimen and film) is
           required.  The beam stopper can be left in place or
           removed from the beam path 1 to 2 s before closing the
           shutter.  A double exposure of the specimen image and the
           SAED pattern can be taken if particle-to-particle spacing
           is adequate.

4.  Dse of Tilting to Acquire Exact Zone-Axis SAED Patterns

Determination of the Tilt Axis—
     In the side-entry type electron microscopes, the instrument tilt  axis  is
always fixed.  However, the position of the tilt axis on the viewing screen
shifts with magnification.  Also, there is always an angular rotation  between
the image and the SAED pattern.  It is highly desirable to know the location
of the tilt axis on the viewing screen and its relationship vis-a-vis  SAED
pattern under the operating conditions to make effective use of specimen
tilting for obtaining exact zone-axis orientations.  The following  steps  can
be used to locate the position of the tilt axis:

     (1)  A gold-coated EM grid with a standard asbestos mineral
          specimen on a polycarbonate replica film is placed in a
          tilt-rotation or double-tilt holder and inserted at 0°
          tilt into an aligned TEM set at  100 kV, 100 uA,  20.000X
          magnification,  and  20-ym camera  length operation.

     (2)  The image is focused on the fluorescent screen, which is
          at approximately  16.000X magnification.

     (3)  A circular hole in  the polycarbonate replica  is  positioned
          in the center of  the field of view.

     (4)  On tilting,  the circular feature changes to an ellipse
          with the major  axis unchanged,  and  indicates  the position
          (direction)  of  tilt axis at that magnification.  The minor
          axis shows  the  perpendicular direction  to  the tilt  axis.
          A high tilt  angle defines,  the  tilt  axis more  accurately
          than a small tilt angle. Figure  A15 illustrates  the  effect '
          of tilt.
          •
      (5)  A double-exposure photograph at  0°  tilt and  at  some  high
          tilt angle,  such  as  30°, is  taken  of  the focused circular
          hole for  reference.

Tilting—for zone-axis SAED Patterns—
      Quantitative  SAED requires  knowledge of  crystallography  to  obtain useful
zone  axis diffraction patterns  from  which precise measurements  can be made for
comparison with  known asbestos  standards  on  file.  Thus the method of obtain-
ing  the  visual SAED pattern of  randomly  oriented  specimens,  as  in  Level I and
II analysis,  is  modified  for  quantitative SAED pattern analysis.   It requires
tilting  of the specimen  to align major  crystallographic directions with the

                                       50

-------
electron beam.  The zone axis is a line parallel to a set of intersecting
crystal planes and nearly parallel to the electron beam.  A zone-axis pattern
thus gives regular repeat distances and even intensities of spots throughout
the pattern.

     Either a double-tilt or a tilt-rotation type specimen holder can be used
for obtaining zone-axis patterns.  A double-tilt holder is often preferred
because tilt-rotation combination involves translational movement of the fiber
during tilting, necessitating constant adjustment of the specimen-positioning
controls to keep the specimen centered in the SAED aperture.  On the other
hand, it is much easier to obtain an accurate measure of the degree of tilt
and perform systematic tilting with the tilt-rotation specimen holder.  It is
only necessary to rotate the specimen (fiber) until the tilt axis (as deter-
mined earlier) coincides with a major row of spots and then tilt until a major
zone axis is parallel to the incident electron beam.  Alternately, fiber axis
of the fiber can be oriented either parallel or perpendicular to the tilt axis
and then further tilting is used to obtain exact zone-axis orientations.

     In order to avoid flip-flopping between image and diffraction modes while
tilting, a recommended procedure is to defocus the diffraction pattern (the
aperature becomes visible and the specimen/fiber can be seen in it) so that a
double image of fiber in aperture can be seen with a poorly focused diffrac-
tion pattern.  The movement of the fiber can then be tracked in relation to
the spot pattern during tilting and kept centered in the SAED aperture by use
of the specimen-positioning controls (knobs) of the microscope.  Sometimes a
larger aperture aids in the tracking-pattern recognition process.

     An experienced electron microscopist can readily recognize the geometri-
cal features like Klkuchi lines or Laue zones in the SAED pattern and use
these to obtain the exact zone-axis SAED patterns.  A detailed discussion of
Kikuchi patterns and Laue zones and their utility in tilting experiments may
be found in any standard text book on electron microscopy.  Use of the double-
tilt specimen holder is very helpful and less tedious in tilting experiments.
However, all laboratories may not have both  types of specimen holders avail-
able.  A skilled microscopist can use either specimen holder without much
difficulty.  Experience and skill are more important factors in SAED analysis
than the type of specimen holder used.

5.  Characteristics of SAED Patterns Encountered in Asbestos Analysis

     Successful application and exploitation of SAED analysis in asbestos
analysis needs prior knowledge of the general appearance and distinguishing
characteristics of other SAED patterns which are often encountered. . The
following discussion summarizes some of  the  observed SAED  features of  asbestos
and other related minerals.  This discussion is by no means  comprehensive  and
assumes  that  the reader  is familiar with general crystallography and  the
nomenclature  pertaining  to varous aspects of SAED patterns.

Minnesota!te  and Stilpnomelane—
     These  iron-rich non-asbestos layer minerals are often  encountered  in
asbestos analysis of specimens  from certain  geographic  locations.   Particu-
lates of these minerals  lie near  their basal  (001) planes.   Stilpnomelane  and

                                       51

-------
minnesotaite both possess large superlattices and their commonly observed  SAED
patterns are easily distinguishable from araphibole patterns.  The spacing  (in
reciprocal space) is about half (for minnesotaite) or less than that for most
amphiboles.  These minerals can be readily distinguished in Level I or Level
II analyses if a gold coating (optional) is applied to the specimen grids.  A
visual inspection of the number of rows of spots inside the (111) gold ring is
sufficient to distinguish minnesotaite and stilpnomelane from amphiboles.

Chrysotile—
     Due to the cylindrical lattice of chrysotile the SAED pattern is unique.
The SAED pattern observed is symmetrical about the cylinder axis, x, and the
spacing of the rows of spots is proportional to  I/a, where a_ is 0.53 nm.   The
most distinguishing features of the pattern are  the flared spots of the type
(130) which occur in the firt layer line.  The flaring is due to the cylin-
drical lattice.  A typical EDS spectra shows the presence of only Mg and Si
(Figure All).

Amphiboles—Systematic Absences, Twinning, and Double Diffraction—
     The most commonly observed row of diffraction shots found in SAED
patterns in araphiboles is in the y* or b* direction, representing the shortest
reciprocal spacing between the spots (18.4 A in  real space).  There are many
strong zone axis orientations containing the y*  row of spots.  The lattice of
amosite, crocidolite, tremolite, and actinolite  is c-centered, and for such a
lattice the h + k odd spots are absent along the y* or b* row.  In practice,
however, weak spots may be present in forbidden  positions due to the presence
of thin multiple twinning on (100), which cause  streaking parallel to a*.
Often, reciprocal nets from both twins are present in the same SAED pattern.
In a twinned crystal, the number of important diffraction nets containing  b*
is doubled, leading to the observation that the  diffraction patterns appear
insensitive to tilt.

     In some cases SAED patterns can contain spots from both  twin individuals
which overlap.  However, not all the spots present in the composite  SAED
patterns are generated by the overlapping nets;  some spots may be present
because of double diffraction where a diffracted beam from one twin  becomes
the transmitted beam when it enters the other twin.

     The purpose of the above discussion is to point out that although many
complications exist in the analysis of  SAED patterns, these can be overcome;
in a good goniometric tilting stage most amphiboles can be identified by  SAED
analysis.

Amosite—
     The nearest reciprocal lattice section to  the  (100) direct lattice  plane
in amosite is  (301)* and it is also the most commonly observed section.   Due
to the presence of the thin (100)  twins,  this section closely resembles  (100)*.

     Typical EDS spectra from amosite  fibers  (Figure All)  show mainly  Si  and
Fe with  smaller amounts of Mg and  Mn.   Mn  is  frequently  observed  as  a  substi-
tutional cation in amosite.
                                       52

-------
Crocidolite—
     Most of the commonly observed patterns are asymmetrical and cannot be
indexed easily.  However, they all show rows of spots separated by a
reciprocal repeat (R) corresponding to the fiber axis (0.53 nm).

     The main elements observed in typical EDS analysis are Mg, Si, Ca, and
Fe.  Na, which is usually present in crocidolite, may not be detected in gold-
coated specimens because of absorption, or because of overlapping secondary
peaks from the copper grid.

Tremollte-Actinolite—
     Temolite and actinolite show a variety of SAED patterns which have very
similar appearances.  In actinolite some of the Mg is replaced by Fe, with the
result that interplanar d-spacings of actinolite are slightly larger than
tremolite.  In both tremolite and actinolite, the main elemental constituents
are Mg, Si, and Ca.  Actinolite also contains some Fe.

Anthophyllite—
     Even though anthophyllite has an orthorhombic crystal structure, its
commonly observed patterns are similar to the monoclinic amphiboles.  Antho-
phyllite fibers dehydrate more easily in an electron beam and are, therefore,
more difficult to study.

     EDS elemental analysis shows the main constituents to be Si and Mg with a
small amount of Fe.

6.  Determination of Camera Constant and SAED Pattern Analysis

     As mentioned earlier, a thin film of gold is evaporated on the specimen
EM grid to obtain zone-axis SAED patterns superimposed with a ring pattern
from the polycrystalline gold film.  Since d-spacings corresponding to identi-
fiable gold rings are known, these can be used as an internal standard in
measuring unknown d-spacings on an SAED pattern from a fiber.  The precision
of measurement is as good as the quality of the photograph (or negative) and
usually the measurements should be in the order of 0.1-0.2 mm with an angular
tolerance of 0.5-1.5 degrees.  The measurements can be made by several
methods:  manually with a ruler, with a mechanical aid, or a densitometer,
etc.  The patterns can be read directly on the developed negative or on an
enlarged non-glossy print.

     In practice, it is desirable to optimize the thickness of the gold film
so that only one or two sharp rings are obtained on the superimposed SAED
pattern.  Thicker gold film would normally give multiple gold rings, but it
will tend to mask weaker diffraction spots from the unknown fibrous particu-
lates.  Since  the unknown d-spacings of most interst in asbestos  analysis  are
those which lie closest to the transmitted beam, multiple gold rings are
unnecessary on zone-axis SAED patterns.

7.  Determination of Camera Constant Dsing Gold Rings

     An average camera constant using multiple gold rings can be  determined  as
explained below.  However, in practice, in most cases determination of  the

                                       53

-------
average camera constant is not necessary and  thicker gold  films  are  not
desirable.  The camera constant, CC, is 1/2 the diameter,  D, of  the  rings
times the interplanar spacing, d, of the ring being measured and is  expressed
as :
                           CCdnm-JL)   =        x d(0
     The value of d for each ring can be obtained  from  the  JCPDS  file.

     (a)  Measure the diameters  (two perpendicular locations)  of  the
          gold rings in mm as precisely as possible  (see  Figure
          A16).

     (b)  Measure as many distinct  rings as  possible to minimize
          systematic errors.

     (c)  Example:  if the measured values in  mm  are Dj ,  T>i,  03,  D^ ,
          and DS , these will represent, respectively, d-spacings  of

           4.079  A. 079  A. 079   4.079      4.079
             , —  »   ;  »   , —  >   / —  » and   / —  A
            /3      2     /8     /u         /12

     (d)  The camera constants will be:
                     l    4.079     l
           CC,  =  -=- x  -^2-  = ^- x  2.355
                    D2    4.079    °2
            CC2  =   — x  -^~  = — x 2.04
                    D3    4.079   °3    ,  ...
            CC3   =   -=- x  	 = -=— x 1.442
                          /ff
                          .
            CCU   =  ^-x       =    x 1.23
                    2     /H    2     -•
                    D5    .
            CC5   =  .^-x  i     = _x 1.178
      (e)  The camera constant f or the SAED patterh'is the average of
           CCi ,  CC2 ,  CCa ,  CCi, ,  and CC$ .  Table 2 presents an example
           of camera-constant determination.

                                       54

-------
             TABLE 2.  DETERMINATION OP CAMERA CONSTANT (EXAMPLE)

Ring
No.
1
2
3
4
Dj readings
(am)
23.0,
27.4,
37.8,
44.6,
22.0
27.6
38.2
45.4
Mean Dj
(mm)
22.5
27.5
38.0
45.0
d-spacing, di
(A)
2.355
2.04
1.44
1.23
Camera constant
Cj = Dt/2 x dt
26.5
28.0
27.4
27.7
  Mean Value of     L i     26.5 + 28.0 + 27.4 + 27.7     my/    IN
Camera Constant  =  ~  =  	5	  =  27>4 (mm~^
8.  Measurement of d-Spacings and Interplaaar Angles

     The gold film, because of its small, randomly oriented crystallites,
produces a ring pattern superimposed on the SAED pattern from the fibers.   The
diameters of the gold rings correspond to known values of d-spacings, and  this
provides an internal standard to correct for inherent uncertainties  present
due to variations in instrumental and/or operating conditions.  Since the  d-
spacings of interest on SAED patterns are usually the ones that lie  closest to
the center spot (transmitted beam), a camera constant measured from  the  first
gold ring in the direction of measurement of d-spacings will usually give
better accuracy in computed spacings than the use of an average camera
constant.  This method will account for any distortions in the symmetry  of the
gold ring pattern.  The zone-axis SAED pattern usually has several rows  of
spots within the circular pattern of the gold rings.  These rows of  spots
contain information about the two sets of planes in the crystal structure  and
the angle between them.  The following procedure outlines the steps  necessary
to obtain the distances between planes (d-spacings) and the corresponding
interplanar angle, 6 (see Figure A17):

     (1)  From the spot pattern, determine the row with spots most
          closely spaced, and designate this as a horizontal row.
          Draw a fine line to show the row through the origin, and
          designate this the zeroeth row.  Draw fine lines to show
          the first and succeeding horizontal rows.  For a few
          horizontal rows, measure the mean spacing between adjacent
          spots (or the minimum vector):


                  Distance between spots m units apart
           A,  =   _ _  _    —-         - -   - - - _ _
             i                      m

          where m is chosen as an optimum number to minimize
          measurement errors.  The mean horizontal spot distance, X,

                                       55

-------
     equals the summation of Xi divided by the number, n, of
     rows measured.   The d-spacing in A corresponding to this
     vector is the camera constant divided by X, and is labeled
     d2•  Table 3 presents an example of spot spacing
     measurement within a horizontal row.
(2)  The perpendicular distance between two adjacent horizontal
     rows is similarly measured.  This interrow spacing, Z, is
     the mean separation between horiziontal rows, and equals
     the distance between a number of rows divided by the
     number of spaces.  This distance is an additional vector
     for comparison that coincides with the slant vector, di~
     spacing, when angle 81,2 is 90°.  The row-spacing 00
     equals the camera constant divided by Z.  Table 3 presents
     an example of perpendicular spacing between horizontal
     rows; Figure A17 illustrates spot and row spacing.

(3)  To obtain the d\-spacing and corresponding angle 9i,2» a
     perpendicular is drawn to the zeroeth horizontal row
     through the origin.  A line is drawn to the first spot to
     the right of the perpendicular in the first row and
     extended through the succeeding rows.  This line, called
     the slant vector, forms the acute angle 9i,2«  The mean
     spacing, Y, between spots on the slant vector can be
     measured by dividing the maximum distance between spots by
     the number of spaces between them, or by calculating from
     the interrow spacing:
                           sin G! ,2
     The d-spacing in A corresponding to this vector  is  the
     camera constant, CC, divided by Y and  labeled dj.
                     CC x sin 61 ,2
     Figure A18 illustrates  the  relationship  of  dj,  d2,  61,2
     and  R.   In some cases,  the  interplanar angle  81,2 nay be
     more  than 90 degrees  (not shown  in  Figure A18).

 Summary of Data from Each  SAED Pattern:

 (a)  The  camera constant,  CC, as determined from the gold
     rings, normalizes  the distances  on  the SAED pattern
     regardless of  such factors  as  magnification and tilting.
                                  56

-------
TABLE  3.   DETERMINATION OF  SPOT SPACINGS (EXAMPLES)
            Separation                 Mean  spacing,
Reading        (mm)         Units          XL (a)

      Spot  spacing within  a horizontal  row,  dj :
   1           49            16        3.006
   2           42.7          14        3.05
   3


                                       3.028 = Mean
                           27 4
            d-spacing  =     '-  = 9<05
  Perpendicular spacing between horizontal rows, R:
                43            8         5.0375
                                        5.0375  = Mean
          d-spacing, R   =      '    =  5.44  A
  Note:   It  is  preferable  that  the  camera constant
  values  used in  computing d-spacings  are measured
  from  the  first  one  or  two gold  ring  diameters in
  the direction of  d-spacing measurement.
                          57

-------
     (b) The parameters of interest are:
         0   d-spacing of spots in a horiziontal row:  CC/X = 62
         •   d-spacing of spots in the slant vector:  CC/Y = di
         •   angle 01>2 formed between a horizontal row and slant
             vector
         •   d-spacing corresponding to row separation as an
             additional parameter of interest:  CC/Z = R.

     It should be noted that the use of camera constant in the form used here
in calculating di, 62, and R, which are measured in reciprocal space on SAED
patterns, automatically converts the calculated numbers into real space
spacings, which are then compared to those from a suitable standard file.

9.  Identification of Unknown Fibers

     Unknown d-spacings (di and d2>, interrow spacing (R), and interplanar
angles (9) measured from zone-axis SAED patterns of unknown fibers are
compared with corresponding known values tabulated in JCPOS powder diffraction
files, or those computed using lattice parameters and crystal structures of
candidate asbestos minerals, or with the values contained in an internally
developed file from standard specimens of candidate minerals.  Table 4 is an
example of the TITRI standards file (Jones et al.,  1981).  Figures A19 to A22
are examples of zone-axis SAED patterns.

     Unknowns are matched as closely as possible to  the file parameters for
positive identification.  However, considerable care and competent judgment
are required in Level III confirmatory analysis.  For example, amphiboles are
usually nonstochiometric minerals, and thus a perfect match may not be possi-
ble between the d-spacings and interplanar angles determined from unknown
fibers and those available from standard minerals.  JCPDS Powder Diffraction
files do not list interplanar angles.  Since amphiboles have low-symmetry
crystal structures, tabulated values of d-spacings and interplanar angles
would be extensive and very expensive to generate, and to get an accurate
match may not be possible because these tables are derived assuming certain
lattice parameters which may not be the same as those of the unknown fibers
being analyzed.  Given these inherent uncertainties, it would seem that use of
internally developed SAED  files consisting of several readily accessible
orientations (by virtue of natural habit of amphibole fibers) from standard
amphibole species could eliminate a lot of tedious  unnecessary work and yet
provide reliable data  for comparison and identification of unknown fibers.

     In practice, SAED analysis combined with qualitative EDS analysis may
help resolve certain cases where a close match in d-spacings and interplanar
angles  is not possible.   For difficult  specimens or  SAED patterns of contro-
versial nature, a second  opinion may be necessary, especially if a legal case
is  involved.                       		           . ."".
                                       58

-------
TABLE 4.  COHPARISION OP d-SPACINGS FROM SAED FILE
      AND  POWDER  DIFFRACTION  FILE  (EXAMPLE)

Internal Standard File Data
AraphLbole
type
Amosite




Crocidolite




Tremolite



Anthophyllite

Zone
axis
flOO]
[307]
[101]
[Toil
[310]
[100]
[101]
[Tio]
[30T]
[3ioj
[100]
[101]
[loT]
[30T]
[100]
[T42]
dl
(A)
5.3
1.79
4.88
4.14
5.22
5.22
4.94
4.79
1.75
5.12
5.04
4.83
2.59
1.72
—
4.56
d2
(A)
9.14
9.26
9.23
9.11
5.13
8.97
9.05
8.19
8.97
5.12
9.03
9.03
8.97
8.98
—
4.56
9
(deg)
90.0
84.0
74.0
78.0
95.0
90.0
75.0
79.0
83.5
96.0
90.0
75.0
80.5
83.5
90.0
60.0
Interrow
spacing, R
(«)
5.3
—
5.17
4.21
—
5.22
5.19
5.23
—
—
—
—
—
—
5.24
—
Powder Diffraction
File Data (1975)
dl
(A)
5.22
1.76
4.84
4.10
5.22
5.20
5.89
4.89
1.76
—
5.07
4.87
2.59
1.69
5.28
4.50
d2
(A)
9.20
9.20
9.20
9.20
5.12
9.02
9.02
8.40
9.02
—
8.98
8.98
8.98
8.98
8.90
4.50
File
index
no.
17-725
17-725
17-725
17-725
17-725
19-1061
19-1061
19-1061
19-1061
19-1061
13-437
13-437
13-437
13-437
9-455
9-455
                         59

-------
                                   SECTION 7

                                ARCHIVAL SAMPLES
 DISCUSSION OF PROTOCOL

      Samples  that  have been  collected  on  filter  substrates  other than poly-
 carbonate,  or that have been collected without regard  to filter loading
 levels,  are referred  to as archival  samples.  These  samples were usually
 collected  for other analytical  objectives,  such  as for optical microscopy or
 gravimetric analysis,  for defined  sampling  periods without  regard to concen-
 tration  levels in  the  air, or for  collection  of  particles larger than 10 um in
 diameter.   Such samples were historically collected, and are of value and
 interest in determining the  presence of asbestos fibers and/or structures.
 Filter substrates  designated as archival  samples include glass fiber filters;
 cellulose  or modified  paper  filters; cellulose ester filters; other organic
 polymeric  membranes,  such as polystyrene, nylon, and polyvinyl chloride; and
 all  overloaded organic polymeric membrane filters.

      The purpose of the preparation  step  is to  transfer particles from a
 filter surface to  an  EM grid with  a  minimum of distortion in morphology and
 size distribution.  The nature  of  non-polycarbonate  filter substrates or
'particle loading makes it sometimes  necessary to transfer a satisfactory
 quantity of particles  to a  polycarbonate  filter  prior to transfer to the EM
 grid.  At  present, only transfer to  an EM grid  from  a polycarbonate filter has
 been standardized.

      A modified preparation  technique is  recommended for archival samples,
 followed by the analytical  methodology using Level  I,  Level II, or Level III
 effort—with the understanding  that  these samples will indicate the presence
 of asbestos, and secondarily the number,  size,  distribution, and morphology.
 The results from sample to  sample  are less  precise  due to problems in
 standardizing the preparation procedures  used for archival samples.

      The archival filter samples are prepared for analysis based on the
 information sought, type of filter material, and particle loading on the
 filter.  The various preparation techniques for these filters include:

      (I)   Individual particle picking and/or reverse washing of the
            filter, with subsequent filtration of the filtrate using a
            polycarbonate filter.

      (2)   Collapsing the membrane filter structure  by exposure  to
            solvent vapor (surface fusion), to produce a more uniform
            substrate for replication and grid transfer.


                                        60

-------
     (3)  Solubilizlng filter material in selected solvents,
          followed by separation of participates.

     (4)  Low temperature ashing (LTA).

Two preparation methods are recommended based on filter loading and type of
filter material:  surface fusion, and LTA.

DESCRIPTION OF METHODOLOGY

     Because the greatest number of archival samples have cellulose ester
substrates, this type of filter material is used in examples describing the
methodology.

1.  Samples with Adequate Loading

Discussion—
     As an example, samples collected on cellulose ester filters have been
received by a laboratory.  Direct transfer of the particulates to  the EM grid
is possible using acetone as the solvent in a modified Jaffe wick  washer.
However, a question arises concerning indeterminate particle loss  in the
transfer.  Carbon-coating the cellulose ester filter prior  to grid transfer
minimizes particle loss.  However, this Improvement in particle count is
offset  by difficulty in visually observing and counting the fibrous particles
against a replica background of the uneven surface topography of the cellulose
ester filter, and by indeterminate loss of very  small particles hidden  in  the
crevices of the uneven filter surface.  The NIOSH method of surface fusion
(Zurawalde and Dement,  1977) is  relatively reliable, and produces a more
consistent result, although the question of loss of the very small particles
has not been resolved.  LTA, described later in  this section, may  also  be  used
for these samples.

Procedure—
     The NIOSH  technique, a modification of a particle-transfer  technique
developed at Los  Alamos Scientific Laboratory (Ortiz and  Isom,  1974),  is
described as follows:

     (1)  A section  of  the membrane  filter is cut with a  scalpel,  and
          placed  on  a  clean microscope slide with the  sampled side
           facing  up.

      (2)   The  cut section  is  fastened on  all sides  to  the  slide  with
           narrow  strips of transparent  tape.

      (3)   The  slide,  with  the  cut  section, is exposed  to  acetone  vapor
           (not  liquid)  for approximately  10 min. The  acetone vapor
           collapses  the  structure  of the  filter  and  produces  a  fused,
           relatively smooth-surfaced film.  The  size  of  the acetone
           vapor bath and  time  of  filter  response to  the  vapors  are
           critical in obtaining the  desired  smooth,  fused surface;
           each laboratory must determine  its own optimum conditions.

      (4)   The  fused  filter section is placed on  the  rotating stage of
           the  vacuum evaporator for  carbon-coating.

                                       61

-------
     (5)  A 3-mm-diameter portion of the carbon-coated filter is
          transferred to a carbon-coated EM grid in the modified
          Jaffe wick washer.

     (6)  Acetone is used in dissolving the fused membrane filter.

     (7)  Transfer to the grid and options for analytical efforts
          were described previously.

2.  Samples with Heavy Loading

Discussion—
     As an example, samples collected with a heavy deposit of particulates
have been received by a laboratory.  These particulates may be organic in
nature (for example, pollen or soot), or of mineral matter.  LTA is used to
remove the organic material (filter as well as particulates), leaving the
inorganic residue.  The residue is gently resuspended and dispersed in
filtered distilled water by low-wattage, short-time ultrasonification.  The
resuspension is then filtered onto a O.l-um (pore size) polycarbonate
filter.  The dry, particulate-loaded polycarbonate filter is then carbon-
coated and transferred to EM grids for analysis as described previously.

     Low temperature ashers are available with one, two, or four chambers.
The following modifications minimize contamination in using these units:           /-,

     (1)  A single chamber is dedicated for ashing samples for EM
          analysis.
     (2)  An in-line filter is placed in the oxygen supply between
          the regulator and entry into the asher.                                   ,;

     (3)  For models with direct access to ambient laboratory air on
          completion of ashing and return to ambient  pressure, a
          filter  is placed in the inlet line to prevent  Laboratory
          air from .being sucked into the chamber.    . .  .    •

In  using the single-chamber method,  a blank test  tube and  the sample  tubes  (up
to  four, for a  total of five  in a  IQr-cm-diaraeter. chamber.) are placed  in-the  •••' :.-•  ..>
chamber "lengthwise, with  the  opening facing-the door. •                      .'.V ,'..".".
 4   ^    •
     The filtration step  is also used in diluting the initial heavy
particulate loading.  Filtration of  aliquots is not recommended  to  obtain
different levels  of loading on the new  filters; a representative  sample  from
each aliquot  in the filtration of  suspensions  is  difficult to obtain.
Instead, for heavy  loadings,  different known areas of filter segments  (one-
eighth, one-fourth, or  one-half  of  the  filter)  should be  ashed  so  that  the
entire contents of  the  resuspension  tube can be  filtered  onto either  a  25-mm-,
37-mm-, or  47-mra-diameter polycarbonate  filter for the desired  dilution.

      Distilled  water  is  filtered  through a  0.1-um (pore  size)  polycarbonate
filter  prior  to use.  All glassware  is  washed  with soap  and  water,  rinsed with
acid, and  then  rinsed with particle-free distilled water.  The  dedicated asher
chamber  is  carefully  wiped with  damp lens  paper.
                                       62

-------
LTA Procedure—
     The LTA manufacturer's instructions are followed since the power  required
for one, two, or four chambers, the mass (glassware plus sample) placed  in  the
chamber, and the desired rate of ashing all vary.   In general, the following
steps are performed:

     (1)  Each filter segment with a known deposit area is carefully
          placed in a clean test tube (13 mm x 80 mm) using a clean
          tweezer.

     (2)  With forceps, the tubes containing the sample, and one lab
          blank (unused filter segment of the same size and type of
          filter material as the sample) are placed lengthwise, side
          by side in the chamber, with the mouths of the tubes facing
          the open end (door) of the asher chamber.  The tubes are
        '  laid in the center of the chamber within the region of the
          coils surrounding the chamber.  Up to four sample tubes and
          one blank can be laid like logs inside the chamber.

     (3)  The power is slowly and carefully increased to prevent
          "flashing" of the filter, which would result in  loss of
          sample.
     (4)  The filter membrane vanishes in about 30 rain; ashing is
          continued for another 2 to 3 h to ensure complete ashing.
          The chamber is slowly allowed to reach ambient pressure.
     (5)  The test  tubes are carefully removed and placed  in a
          beaker, covered, and stored on a class-100 clean bench  for
          resuspension.

 Resuspensioa (SonlficaClon) Procedure—
     Ultrasonification is used in resuspending and redispersing  the  ash
 resulting from "the  LTA.  The superiority of a probe-type ultrasonic  device
 over a  bath-type device has not been demonstrated.  However,  the  criteria of
 low energy  and minimum Bonification  time appear valid.  The  probe-type
 instrument  is more  readily calibrated (desired reproducibility),  but  requires
 a  larger volume of  suspension  to work with.  The  bath-type instrument is more
 difficult to calibrate, and is usually of  fixed wattage.   A generalization is
 that probes are used  for dispersing, baths  for cleaning. .  Recently,  a bath-
 type ultrasonic unit  (Ladd Research  Industries,  Inc.,  Burlington,  Vermont)
 with a  variable power  source and  timer has  become  available that  appears to
 have the advantages of  both types of ultrasonic devices.

     The  resuspension  procedure  is  as  follows:

     (1)   10 mL of  filtered distilled  water  is added  to each test
           tube.
     (2)   Each  tube is  placed  in a  100-mL  beaker  containing 50 mL of
          water.
      (3)   The beaker,  with the tube,  is  placed  in the  low-energy
           ultrasonic bath.
                                       63

-------
     (4)  Ultrasonic energy of about 50-to 60 w'(60%) is applied for
          3 rain.

Filtration Procedure—
     Liquid filtration of suspensions for EM examination using polycarbonate
filters is one of the more difficult procedures to standardize.  Variations in
the nature of the filter material, geometry and distribution of pores, and
method of manufacturing make it difficult to obtain a uniform deposit of
particulates on the filter.  The following procedure is used for consistency
in the filtration procedure:

     (1)  A filtering apparatus having a filter size adequate for  the
          desired dilution—preferably 25-mm or 47-mm diameter—is
          assembled.  A polycarbonate filter (O.l-gm pore size) is
          used shiny side up for the deposit, with a cellulose ester
          filter (5~um pore size) as a backing filter on the glass
          frit.

     (2)  While dry, the filters are centered and suction is applied.
          The filter funnel is mounted on the centered, perfectly
          flat filters with the vacuum on.

     (3)  The vacuum is then turned off.  A 2 mL amount of  particle-
          free distilled water is added to the filter funnel,
          followed by careful addition of all the water in  the test
          tube containing the dispersed ash.  The test  tube is rinsed
          twice with particle-free distilled water, and the contents
          are carefully added to the filter funnel.

     (4)  Suction is then applied; neither rinsing i:he  filter funnel
          nor adding extra liquid is permitted during the entire fil-
          tration process.

     (5)  At  the end of filtration, suction is stopped.

    •(6)  If  possible, the filter is dried on a glass slide or holder
          that can be placed directly in  the vacuum evaporator for
          carbon-coating.

    1 (7)  The dry filter  is stored, in a .disposable Petri dish  (taped':   ' •  '/
          on  a glass slide), or in the special holder,  until ready
          for carbon-coating, grid ^transfer, and  EM analysis.

     (8)  The effective area of the redispersion  filter and the  area
          of  original filter deposit cut  for ashing must be recorded
          (ashing factor)  for inclusion in analytical data  reduction
          and reporting.
                                        64

-------
                                   SECTION  8

                             BULK-SAMPLE ANALYSIS
DISCUSSION OF PROTOCOL

     Bulk samples may be air samples collected in large volumes using electro-
static precipitators, bag collectors, or high-volume samplers, for example; or
chey may be original pieces of source material containing asbestos, such as
insulation, asbestos paper products, and asbestos cement products.

     Efficient usage of the three levels of analysis requires effective com-
munication between those requesting an analysis and those responsible for  con-
ducting the analysis.  Personnel requesting an analysis must understand the
limitations of each level of analysis by EM.  For example, requesting EM
analysis of a bulk-material (solid) sample, where there is marked disagreement
regarding the presence of asbestos (amphibole), and using Level I (screening)
analysis, are incompatible.  Bulk-material samples require grinding for analy-
sis; grinding requires care to minimize such problems as contamination, change
in size of the asbestos fiber, increase in fragments that morphologically  meet
the criteria of a fiber, possible change in the relationship of asbestos to
nonasbestos components, and possible destruction of asbestos fiber crystal-
Unity.

     Following grinding, bulk-material samples should first be analyzed by
PLM, followed by XRD, if necessary.  XRD provides information on samples
having asbestos concentration levels of at least' 2%.  PLM provides information
on asbestos and nonasbestos components, as well as on the size of the asbestos
fibers in the solid-bulk phase.  The additional information aids in EM . • ..•..;:.."£
analysis of these samples at the selected, level. of .analysis.        .    .   ...','•':«

DESCRIPTION OF METHODOLOGY

1.  Polarized Light Microscopy

     Analysis of bulk samples, such as insulation material,  for component
identification and for determination of the type and concentration of asbestos
present is best accomplished by  PLM.  With  the  polarized  light microscope,
particle properties—such as color, morphology, refractive  index, bire-
fringence  (which indicates  a crystalline substance  rather  than an amorphous
substance), surface  texture, reflectivity,  and magnetism—can be observed  and
determined.   Determination  of such  a large  number of  particle properties
allows  identification of  specific particle  types, in most cases.  For example,
amorphous  slags and  crystalline  minerals are  common nonfibrous  filler compon-
ents  of insulation materials that can be easily distinguished by  PLM.

                                       65

-------
2.  X-Ray Diffraction Analysis

     XRD has been successfully used to measure asbestos concent In both
aerosol samples and bulk samples.  The reported values for sensitivity and
accuracy vary depending on the exact technique, but recent reports quote
sensitivity values of about 1% chrysotile, whereas 5% was more common when  the
technique was first demonstrated.  Much of the improved sensitivity derives
from sample preparation techniques, which are important in XRD, but digital
data collection and use of such accessories as x-ray monochromators to reduce
background are also important.

     While 1? sensitivity has not been demonstrated for asbestos materials
other than chrysotile, the same sample preparation procedures are applicable
to other mineral forms, with comparable sensitivities expected in cases with
serious interfering lines.  Interferences would hinder the analysis for
chrysotite as well.as for other asbestos minerals.  Considering that  the  high-
sensitivity procedures have been only partially demonstrated at the 17. level,
a sensitivity of, say, 2% is probably a more  realistic expectation for
asbestos minerals in general.

3.  Electron Microscopy

     For bulk-air samples, asbestos analysis  by EM entails an addition to the
sample preparation procedure to attain a representative powder sample at  a
suitable concentration level to be placed on  the  EM grid.  This additional
step is similar to the method used in preparing standards of known asbestos.
The finely divided powder samples are split into  representative fractions,  and
a small, weighed portion is suspended in a known  volume of filtered distilled
water containing.0.1% Aerosol OT.*  A mild ultrasonic  treatment is used  to
disperse the particles.  Different known volumes  of suspension are filtered
through a 0.1-ura (pore size), 25-mm-diaraeter  Nucleporet membrane  filter.   The
dried Nuclepore^filter is  then carbon-coated  and  transferred  to an EM grid
using the refined Jaffe wick technique described  previously.

     Bulk-solid samples are gently and slowly ground  to a powder  for  EM
analysis to minimize  localized heating;  the powder  is  then prepared  for  the EM
grid by the method described for bulk-air samples.
*"*                                             *                _.- • AI.I- «•!.  •• ' '
      For  EM analysis  of  bulk-air  samples,  a weighed portion is suspended in
 filtered  distilled  water,  deagglomerated  in an ultrasonic bath, transferred to
 a  volumetric flask, and  brought  to 'volume with filtered distilled water.  An
 aliquot is  then  filtered onto  a  0.1-um (pore size)  polycarbonate filter using
 a  5.0-ym  (pore size)  cellulose ester filter as a back-up filter on the
 filtration  apparatus.  The dried  polycarbonate filter is then carbon-coated.
 A  3-tmn x  3-mm portion of the  carbon-coated filter is then directly transferred
 to a  200-mesh carbon-coated copper EM grid using the refined Jaffe wick washer
 technique.   EM analysis  based  on  Level I, II, or III effort is then performed.
 * Fisher Scientific Co.  (Cat.  no. 50-A-292), 711 Forbes Ave., Pittsburgh, Pa.

 t Nuclepore Corporation, 7035  Commerce Circle, Pleasanton, Calif.

                                       66

-------
                                   SECTION 9

                  NUMERICAL RELATIONSHIPS AND ANALYTICAL AIDS
     The fibrous structures (fibers, bundles, clusters,  and  matrices)  in  an
air sample are to be counted, sized, and  identified  as  asbestos  or  non-
asbestos.  An air sample ranging from  1 to 5 m3,  depending on  its  total
suspended particulates  (TSP) content (in'ambient  air,  the average  TSP  is
between 30 and 300 ug/m3) is drawn  through a 37-mra  filter (effective
filtration area of 8.6  cm2) or a 47-mm filter (effective filtration area  of
9.6 cm2).  The asbestos content, unlike a prepared  laboratory  standard or
sample, is a very small percentage  (less  than 1%) of the particulate loading
(TSP content) collected on the filter  surface.

     Two small circular sections of the filter  of approximately  3-mm diameter
are transferred to EM grids for transmission electron  microscopy.   Either one
or both EM grids are examined for asbestos content;  10 random  grid  openings
are examined for each EM grid.  Each grid opening measures approximately  85 urn
x 85 urn.  At 20.000X magnification, a  field of  view of 4.5 gm  x  5.0 ym is used
in the examination.  This approximates to about  300 fields per grid opening or
3000 fields of view per grid examined  (6000 fields  for two EM  grids) if less
than 100 asbestos structures had been  found.  Unlike a field blank or
laboratory blank, the statistical significance  of obtaining  a  low asbestos
count in the midst of atmospheric clutter needs  to  be  recognized.

LIMITS OF DETECTION                              '                              •••"
                                                                               "
     The minimum detection limit of the EM method for counting airborne
asbestos fibers varies  depending on the amount  of total extraneous, particulate,.,!
.matter.'.In • the sample, and on the contamination  level in the  laboratory   -   '  'v-'
environment.  This limit also depends  on  the air sampling parameters,  loading
level, and EM parameters used.  For^example, assuming that a fiber count  has
an accuracy of ±1 fiber, when  10 full-grid openings are scanned, each grid
opening having an average area of 0.72 x  10~u cm2,  the detection limit is
determined from the equation

      n         , .  .        1   Area of filter  (cm2)           1
      Detection limit   =  — x 	-	 x
                           10    0.72 x 10'*  (cm2)     Volume of air (m3)

      The minimum detection limit,  then,  is  lower for very dilute samples.
 Examining full-grid openings leads to a  lower value of the minimum detection
 limit because of the large area scanned, as  compared with the field of view
 method.   With a given sample, the  detection  limit can be lowered considerably,
 but the  required experimental effort increases.  The guideline of using  10

                                      . 67

-------
"full-grid openings represents a judicious compromise between a  reasonable
experimental effort and a fairly low value of the detection limit.   However,
using two or more TEM grids reduces the detection limit  further and  improves
che precision of the estimates.

STATISTICAL METHODOLOGY

     Several statistical strategies have been used  to characterize airborne
asbestos distributions and estimate the abundance of asbestos fibers in  a
given sample.  These methods range from simple  tabulation  of observed frequen-
cies of fibers across grid opening samples to the fitting  of statistical prob-
ability distribution such as the Poisson.  This statistical section  outlines  a
general methodology for fitting observed data to a  statistical  probability
distribution (either Poisson or normal depending on the  fit of  the  former).
The mean and 95X confidence interval are- then estimated  and used  for the
purpose of sample description and drawing inference regarding  the abundance  of
airborne asbestos fibers in the environment  in  which  the samples  were
obtained.

     As an illustration, consider the hypothetical  data  in Table  5.

     The expected number of grid openings with  no fibers is:


                        Ne~" = (98)  (e-2""^*) = 4.7806
 The  expected member of  grid  openings with 1,  2,  3,  ...  fibers are found by
 multiplying Neu by u/r  ,  i.e.,  4.7806 x 3.0204,  4.7806  x 3.0204/2, 4.7806 x
 3.0204/3,  successively.
                «*•
      To test the fit of the  Poisson distribution to the observed data we
 compute a  chi-square statistic, x2 = £ (observed -  expected)2/expected = 8.26.

      Since there are nine different observed  frequencies (i.e., numbers of
 fibers) there are 9-2=7  degrees of freedom (since we estimate one
 parameter).  The probability of xa2 = 8.26 is p = 0.4;  therefore, we conclude -  j
 that the Poisson distribution fits the observed data.

      In certain cases,  the observed1frequencies will not have a Poisson
 distribution (as determined  by the 2rev*-°usly described chi-square statistic).
 In this case we estimate the mean (X), variance S2  , and 95% confidence limits
                                       68

-------
                         .TABLE 5.  HYPOTHETICAL DATA

No. of fibers
on grid 1
opening Observed* Expected Observed
(r) frequency (f) frequency - expected
0









1
2
3
4
5
6
7
8
9
10
1
1 or more
Total
3 4.78
17 14.44
26 21.81
16 21.96
18 16.58
9 10.03
3 5.04
5 2.18
0 0.82
It 0.27
0 0.08
_0 0.03,
98 98.0
-1.78
+2.56
+4.19
-5.96
+ 1.42
-1.02
-2.04
+2.84

1.20 -0.20



Probability
(r)
.049
.149
.224
.224
.168
.101
.050
.022

.003




* The number of grid openings showing that number of fibers.
T We combine adjacent frequencies to get a minimum of 1 fiber per group.
  Assuming a Poisson distribution, the mean is u = E fr/Ef = 296/98 =  3.0204.
  That is, the sum of the product of the observed frequencies and number  of
  fibers divided by the sum of the frequencies.
                                       69

-------
assuming normality.  The 95% confidence limits are computed as a  function  of
the sample variance and t distribution.
                                n I  X2 -  [E   X  ]2
                                     n(n -  1)


where k is the number of grid openings, n is the total number of  fibers  found,
and X^ is the number of fibers found in grid opening i.

     The 952 confidence limits are given by
where t is the value of the two-tailed t distribution  for  probability  p  <.025
and n - 1 degrees of freedom.

1.  95Z Confidence Limits for a Poisson Varlate
     To generate a level of confidence  regarding our  estimate  of  the number of
asbestos fibers per grid opening, a  95% confidence  limit  can be derived.   For
counts of 0 through 20, Table 6 may  be  used.

     For example, if out of 20 grid  openings  15 fibers  are  found,  the 95%
confidence limits are obtained by taking  the  lower  and  upper bounds from  Table
6 as 8.40 and  24.74.  Then per grid  opening,  the 95%  confidence limit is
8.40/20 to 24.74/20, or 0.42 tn  1.237  fibers  per grid opening.

     For counts, greater than 20, a simple normal approximation is  computa-
tionally convenient.  The normal approximation is  X - L ± x *  (S^) where  L is
the observed count, SL is /L  , and x*  *s  1.96 or 2.58 for the  95% or 99%
confidence limits, respectively.

     For example, if  35 fibers were  observed  from  inspection of 20 grid
openings, L =  35,-Sy =  /35 = 5.91,  A  = 35 ±1.96 (5.91) = 23.4 to 46.6 or   .
23.4/20 = 1.38 to 46.6/20 = 2.74 fibers per grid opening.

2.  Comparison of Two Poisson Variates .

     In certain cases, a new  test sample  is compared  to a "blank" or "control"
sample.  Table 7 gives those differences  between control  and  test samples that
are significant at  the 5% level.

      Inspection of  Table  7  reveals  that the minimal detectable difference
between  test and control  samples  is  5  fibers  in  the test  sample and 0 fibers
in  the  blanks. Typically,  inspection  of  20 grid openings for  a  blank control
reveals  between 0 and 5  fibers.  At  the upper bound (i.e.,  5  fibers in a
                                       70

-------
TABLE 6.  95 PERCENT CONFIDENCE LIMITS

No. of Fibers
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
95%
Lower
0.000
0.0253
0.242
0.619
1.09
1.62
2. -20
2.81
3.45
4.12
4.80
5.49
6.20
6.92
7.65
8.40
9.15
9.90
10.67
11.44
12.22
Limits
Upper
3.69
5.57
7.22
8.77
10.24
11.67
13.06
14.42
15.76
17.08
18.39
19.68
20.96
22.23
23.49
24.74
25.98
27.22
28.45
29.67
30.89
                   71

-------
TABLE 7.  CONTROL AND TEST SAMPLE DIFFERENCES

Fiber
Control




0
0
0 -
I
I
1
2
2
3
3
3
4
4
5
5
5
Count
Test sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Fiber
Count
Control Test sample
6
6
7
7
7
8
8
9
9
10
10
10
11
11
12
12
13
13
13 •
14
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
                      72

-------
control sample), 18 fibers in the test sample are required for statistical
significance.  This estimate of- 18 fibers may prove useful for establishing an
asbestos detection limit criterion.

MAGNIFICATION CALIBRATION

     The following steps should be performed to calibrate the magnification of
the EM:

     (1)  Align the EM using the manufacturer's instructions.

     (2)  Insert mag-calibration grating replica* (with 54,864 lines per
          inch, or 2160 lines per mm) in the specimen holder.

     (3)  Switch on the beam, obtain the image of the replica grating at
          20,OOOX magnification (or at .the magnification at which  the
          asbestos samples will be analyzed), and focus.

     (4)  If the fluorescent screen has scribed circles of known
          diameters, align one line tangentially to the circumference of
          one circle using stage control.  Count the number of lines in
          a diameter perpendicular to the lines.  In most cases, the
          other end of the diameter will be between the n" and nc^ + 1
          line.  The fractional spacing can be estimated by eye.
          Alternatively, the separation between lines can be estimated
          using the scribed circles.

     (5)  If X line spacings span Y mm on the fluorescent screen using
          this grating replica, the true magnification, M, is given by


                              M   Y x 2160
                              M =	
          The readings should be repeated at different  locations  on  the
          replica, and the average of about six  readings should be  taken
          as the representative or true magnification for  that setting
          of the EM, as in the following example:
*  For  example,  Cat. no.  1002,  E.  F.  Fullam  Co.,  Schenectady,  N.Y.

                                       73

-------
          Line Spacings,         mm on Screen,           Magnification,
                X                    Y                        M
               9.5                    83

               9.3                    80
               7.0          .          60
               8.8                    80
               9.0                    80
               9.0                    80
                                                  Average:  19000
     On most EM's with large (18-cm diameter) fluorescent screens, the
magnification is substantially constant only with the central 8- to  10-cm-
diameter region.  Therefore, calibration measurements should be made within
this small region and not over the entire screen.

PREPARATION OF BLANKS

     Even after taking the utmost precautions to avoid asbestos contamination,
the possibility of some contamination cannot be ruled out.  Contamination
should be checked periodically by running field blank samples in addition  to
laboratory blanks.  Field blanks should be analyzed prior to laboratory
blanks.  A blank sample may consist of a clean filter subjected to all the
processing conducted for an actual air sample.  This processing may  include
ashing, resuspension, redeposition, carbon-coating, transfer to a TEM grid,
and TEM examination.,

     When analyses of blank samples show significant background levels of
asbestos, these should be subtracted from the values obtained for field
samples.  Also, the minimum detection limit may be calculated as twice or   "';.
three times the standard deviation of the blank or background value.           '

USE OP COMPUTERS  -                                                           '  j'
                                                                            i
     Data reduction is facilitated by computers.  Computer  printouts can be
used in reports.  Each laboratory. s,hould develop software suitable for its
needs as well as  to maintain basic information, such as  fiber, areas examined,
volume/mass of sample, and  size  distribution, for possible  interlaboratory
comparison.

     Appendixes B and C  present  sample  printouts from Level I and Level  II
analyses, respectively.
                                       74

-------
                                  REFERENCES
Anderson, C. H., and J. M. Long.  1980.  Interim Method for  Determining
     Asbestos in Water.  EPA-600/4-80-005, U.S. Environmental  Protection
     Agency, Athens, Georgia.  44 pp.

Hileman, B.  1981.  Participate Matter:  The  Inhalable Variety.   Environ. Sci.
     Technol., 15(9):983-986.

John, Wi, and G. Reischl.  1980.  A Cyclone for Size-Selective Sampling of
     Ambient Air.  APCA Journal, 30(8):872-876.

Jones, D. R., S. C. Agarwal, and J. D.  Stockham.   1981.   Asbestos Analysis of
     Iron Ore Beneficiation  Plant Samples.  Final  Report,  Work Assignments No.
     3 and  No. 9, Contract No. 68-02-2617.  U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina.   228 pp.
                              -,
Leidel,  N.  A., S. G. Bayer,  R. D. Zurawalde, and K.  A.  Bursch.   1979.
     USPHS/NIOSH Membrane Filter Method for Evaluating Airborne Asbestos
     Fibers.  U.S. Department of Health, Education, and Welfare.  National
     Institute for Occupational Safety  and Health,  Cincinnati, Ohio.   97 pp.

Mueller, P.  K., A. E. Alcover, R. L. Stanley, and  G.  R.  Smith.  1975.
     Asbestos Fiber Atlas.   EPA-650/2-75-036, U.S.  Environmental Protection
     Agency, Research Triangle Park, North Carolina.   58  pp.

Ortiz, L. W., and B. L.  Isom.   1974.   Transfer Technique  for Electron
 !•'• Microscopy of Membrane  Filter  Samples.   Amer.  Ind.  Hyg. Assoc. J.
     35(7):423-425.
  ••'. •,  ••<.••     • •  ' -
Samudra, A.  V., F. C. Bock,  C.  F. Harwood, and J.  D.  Stockham.  1977.
 - •   Evaluating and Optimizing  Electron Microscope Methods for Characterizing
     Airborne Asbestos.   EPA-600/2-.78-038, U.S.  Environmental Protection
     Agency, Research Triangle  Park, North Carolina.   194 pp.

Samudra, A. V., C.  F. Harwood,  and  J.  D. Stockham.   1978.  Electron Microscope
     Measurement  of Airborne Asbestos  Concentrations:   A Provisional
     Methodology  Manual.   EPA-600/2-77-178,  U.S.  Environmental Protection
     Agency, Research  Triangle  Park,  North Carolina.   57  pp.

Zurawalde,  R. D.,  and J.  M.  Dement.   1977.   Review and Evaluation of Analytical
     Methods for  Environmental  Studies of  Fibrous  Particulate Exposures.
      DHEW(NIOSH)  Publication No.  77-204.  National Institute  for Occupational
      Safety and Health,  Cincinnati, Ohio.   71 pp.


                                        75

-------
        APPENDIX A




          FIGURES
Figure Al.   Vacuum evaporator,
            76

-------

                                            '
(a)  37~mm diameter.
(b)  kj-ntm diameter,
     Figure A2.   Multiple coating arrangement in evaporator.

-------
 I
1



                                             3B$Bj^^
                             (a) Modified ^7~mm-diameter Petri-slides

                                     (b) 37~mro~diameter  cassette.
                        Figure A3,  Close-up of multiple coating  arrangement.
                                                 78


-------
                 (a)  Plan view.
Carbon-Coated
   EM  Grid
                                                Wedge
                   Ground
                   Glass  Dish
                    Edge  Cover
                                                                                            Nuclepore
                                                                                              Grid
                                                                                              Stainless mesh
— T*. \ Stainless Mesh ^V * ' *' — —

\
o i
— i /
Filter
^ Petrl 	 »•
_ JLft
L
Dish 1
— ' i / Foam /, \ ^,
t
\
i
1
l
" i 1 ter Paper
                                                                      (b)  Elevation  view.
        Nuclepore.
     Carbon Coat
Carbon Substrate
      Grid
                                                                               Asbestos
                                                                                Fibers
                                                       Ch loroform
                                                         Level
 Asbestos
  F i bers

~Gr id
                                                               Chloroform
                                                                  Wash
                                                              (d)  Principle of  the Jaffe  method.
          Nuclepore C-coated
          Particle Side Down
            Stainless Mesh
                                                         Polyurethane  Foam
           (cj Details of placing the  specimen  for washing.
                      Figure A4.  Modified Jaffe wick washer method  (sketch)

-------

                                                                                                             -J
00
o
             Figure A5.   Modified Jaffe wick washer,
Figure A6.   Transmission electron microscope.




-------
Count as one fiber:
Count as two_fibers (space between fibers greater than the width of one  fiber)
Count as three fibers:
Count as bundles:
Count as cluster/clump:
Count as matrix/debris:
               Figure A7.   Morphology and counting guidelines used
                       in  determining asbestos structures.
                                    81

-------
               No.
                             - 3"
Filter Type.
Filter Area_
Grid Openi/iq Area
                    Nueto
Acc. Volta
11
Ift
19
2Q
£1
2Z
23
24
25
Struct.
P
P
F
P
M
P>
F
F
F
U
IU
M
kA
M
IU
P
IU
M
kJ
p
M
U
M
F
M
Dimension
Width
1
1
1
1
1
2
4

















3
Icnqin
33
IQ
20
^
Q
il
?r\
»o
M
«.
%
2*?
23
7
12
H
U
15
19
17
3A
15
|5
7
13
SAED Observation
Cdyrs
v/
tX
•

*s
I/

x/
tX
•
I/
s
lX

v'
i/
S


St
•'•

:•
i
iS
Ai-gli

























Ambiq



*>


»/






^




W>

•
»x
•


•10-ID
.
















v^





,/

GO
4





5







to




7





Struct.
¥
2k
77
ZB
20
^
31
.^2^
33
54
JK
3fa
^7
.•^ft
.^0
40
41
42
4*
44
4.^
4h
47
48
49
50
Struct.
U
F
F
M
u

M
R
U
^A
U
U
F
M
P
R
F
P
P
LI
F
F
F
F
P
Dimension
Uirlth







-






1
4
5
i







Length
•7
»L
7
Zt
10
ii
4|
^Z
(,
r,,
1.^
17
2Q
in
24
2ft«7
IfT
U
17
0
?rt
Iff
13
9
4|

Cnrys
I/

*/
t/


iX
*/
t/

^




w/

K/
lX
v/

*/
S

S
SAE3 Obs-'-.-jv :-
Anph
















iS








A»b:; -.T- :

,/


S
t/



S







'


S


v/

00
                                             .        .
                                           Figure'-A8.  Level I  data sheet  (example)

-------
CO
                                                                    Tilt Area of  Fluorescent  Screen
                                     .Figure A9.  Scanning of full-grid opening.

-------

1
                           Figure ALO... Jransmission, electron microscope
                                with energy dispersive spectrometer.
                                                   84

-------
1
                                                   .


                                                 •
                         ~ IF 48 LD2
                —r.ii RD&veieKEv/CH    46. SE;;"
                EuG= 2.eeeKEv,   IB.
               ; :2811688 CHRVSO fl4
               ?Hflii  flDD .016KEV/CH     40. SEC;
               BUG= 2. fllKEYJ    27.     •'
                                95  6. 64  8. 51
                             0-4-10-3-<1
                                95  8. 94  8. 91
                            0-7-10-0-0
                TREHOLITE  T 79  988
                PHftli HDD .818KEV/CH
                BUG= 1. 758KEV,   334.
48. SECS
.  ::511680  RNTHOP C2
"":'Sll fiDD '. 816KEV/CH
5JG» 2. 90BKEV,   172.
                                          18.238
                             0-4-10-3-<1
                CROCIDOLITECR 37  11
                PHflll flDD .818KEV/CH    48. SECS
                                                                    0-3-10-0-1
                    ?1?076 fiflOSITE 1
                    : flDD . 810KEV/CH
                     1 946<'E-',

                                95  8.04  8.91
                             1-1-10-0-6
                                 95   8.84  8.51
                            0-2-10-0-7
                         Figure All.   Spectra  profiles of asbestos standards,
                                                  85

-------
Sample  No.Pto4-7n-llQ-
  /.>••?."/?'-  *".;«.;.
 .••'•''•'i ''::.  •>.->">
ExlO     •"""
                                                            Voltage_
                                          Oate
           Filter Type Mur Igno
                               re
           Filter Area   Q.fo  Cmz
                                  i-'''':!;-:'.yi' -'Beam Current  inn I1.A	
                                 ' ;':/r':' |;''--Mag"'« eat i on_2DJQOOx - Fi'lm
                                          Grid Box
           Grid Opening Area 72.25 X
                                          Grid Location   D-7	
.• CommentsQond f)j.
7



e

9
10








Struct.
IT
1
2.
0
3
4
S
ft>
7
8
9
10
II
IZ
13
14
15
Ik.
n







Struct.
F
n
n

n
n
it
u
II
II
II
II
II
II
II
II
II
ll







Dimension -r
Width
2.
9

10
b
4
I
3
2 '
Z
IS'
10
5
b
3
7
4
4
•
•

.

.

Length..,
2(6 -I.
30 -:
• '
85 :
.fo5 ;
23
' 8
40
30, '
!l .-:••
80 :,
80J
.45 --.••'
43. SJ
25, -X;
zi*-:
z5- •-';;
-•*3fl':?

- ^.;!*-
• V." " "> "

. . •-' ''•
•^r- .
• ••• •
SAE Obserwatinn
Chyrs
r
-
."
' '.
*-' ' '."'' '
~
•
>i
( '• :-'•-.•
' :'>r .
: !" '
» ' '»
• - •'
• :
1


•• '
•<#.-.',

.

.•• •

- '
'
Amph
S
^

^
S
•

S
"





S


>.'•
•




' .

Ambig



•


',
<&>


.


<^>
(f^>
•









Non-A






/

/
^
t/
l/
y
,/

v/
l/
v^







No-P
























•
FD/Imaqe
0°




. •

•

•
















EOS
Ha
—

•
___
~
^~
i
— :
.•_

•



—



•






Hg
9


35
>9
fob

37






1*7










Si
19


HO
59
15?

I7O






5 1










Ca
T


3^5
2.2


4J?






'8




,





Fe
•7


IV
Z)
23

»fo






U>





•





10
r^>
TV^Ii 1

(^

^j









.

oo
                                           :<•'•!.
                                            Figure  A12!. Level I.I  data sheet  (example)
                                           :     •    f -iJ>,4   J;   ^                            ,:
                                                   i i  J-! ^ i

-------
                                EM  DATA REPORT
  Sample Number:    R09-2865      _
  IITRI Sample No.: C010-1B59
  Sample Type:  Bulk, Air, Water.  Misc.
  Filter Type:    Nuclepnre	
  Volume of Fluid Sampled:   NA
                                                                3/26/81
                                                                1/30/81
  t.o Anal yzoH:
Date Sample Received:_
(circle one)
Area of Filter  Deposit  (cm2):_
Mass Deposited:        NA
                                                                       8.6
r
Total Number of Structures:
Total Number of Asbestos Structures
2.1 Chrysotile 88
2.« Amphibole 5
Crocidolite 5
Tremolite
Amosite
101 |
: 93
Anthophyllite
Actinolite
Non- Identity
 6.
    2.3  Non-Identity	8_
3.  Asbestos Structure Description
3.1
3.1
3.2
3.2
3.3
3.4
Area
Total Number of Fibers:
.1 Chrysotile
Fiber Length; Range (urn)
Fiber Diameter; Range (um)
Aspect Ratio; Range (um)
.2 Amphibole
•
Fiber Length; Range (um)
Fiber Diameter; Range (um)
Aspect Ratio; Range (um)
Total Number of Bundles:
85
80
.31 -
.06 -
3.5 -
5
.31 -
.06 -
5.0 -

Total Number of Clusters/Clumps:
Total Number of Matrix/Debris:
of Filter Sample Analyzed,
(cm2):


2.25
.25
31.0

2.19
.31
10.0
5
2
1
Mass
Mass
Mean
Mean
Mean
Mass
Mean
Mean
Mean
V
»
•
(ng)
(ng)
(um)
(um)
(um)
(ng)
(um)
(um)
(um)
Mass (ng)
Mass (ng)
Mass (ng)


0
0
11

1
0
6


.0016
.0010
.78
.07
.00
.0006
.00
.15
.40
.0007
.0002
negligible
.0007225
A.
5.  Total Mass of Asbestos  Analyzed  (ng):
    Number of Pictures Attached:	2
                                               .0025
 7.   Qualitative Description of Non-Asbestos Particles  Few, small particles--
       non-descriptive	..-
 8.   Comments:    Particulate loading OK.       	_________
                Figure  A13.  EM data  report (example)
                                          87

-------
             SAMPLE SUMMAKY KLI'OIM
Sample Number: R09-2865 Dale of Ruport:
IITRI Sample No.: C010-1859 Date Sample Received:
Sample Type: Bulk. ffiTj) Water, Misc. (circle one)
Filter Type: Nuclepore Area of Filter Deposit
4/1/81
1/30/81 •
(cmz): 8.6
Volume of Fluid Sampled: NA Mass Deposited: NA
1. Total Number of Structures: 1,202,215
2. Total Number of Asbestos Structures: 1,106,990
2.1 Chrysotile 1,047,474
2.2 Amphibole 59?516
Crocidolite 59,516 Ant hophyl lite

Tremolite Actinolite
Amosite Non-Identity
2.3 Non-Identity 95,225
3. Asbestos Structure Description
3.1 Total Number of Fibers: 1,011,765 Mass (ng)
3.1.1 Chrysotile 952.249 Mass {ng)
Fiber Length; Range (urn) .31 . 2.25 Mean f^1")
Fiber Diameter; Range (um).06. - ,25 Mean (urn)
Aspect Ratio; Range (um) 3.5 - 31, Q Mean (um)
3.2.2 Amphibole 59.516 Mass (ng)
Fiber Length; Range (um) ..31 - 2.19 Mean (um)
Fiber Diameter; Range (urn). 06 - .31 Mean (ym)
Aspect Ratio; Range (um) 5.0 - 10.0 Mean (um)
3.2.' Total Number of Bundles: 59,516 .Mass (ng)*
3.3 Total Number of Clusters/Clumps: 23,806 .Mass (ng)
3.4 Total Number of Matrix/Debris: 11,903 ..Mass (ng) ,
4. Area of Filter Sample Analyzed, (cm2): .0007225'"-'
5. Total Mass of Asbestos Analyzedt(ng): 29.5
6. Number of Pictures Attached: 2
19.00
11.90
0.78
0.07
11.00
7.14
1.00
0.15 -
6.40
8. 30
2.10
. , '0.10 .-
•
7. Qualitative Description^ Non-Asbestos Particles Few, small particles--
Non-descn'ptive
6. Comments: Particulate loading OK.



Figure A14.  Sample  summary report  (example).
                        88

-------
                                                                 xs,

      (a)  Effects  of  tiIting.
(b)  Fiber  alignment.
         Figure A15.   Effects of tilting and alignment of fiber.

Figure A16.  Method of measuring two perpendicular diameters for each ring,
                                    89

-------
     Draw 0,  1st, 2nd,  ... order of horizontal  rows--
     perpendicular separation between  horizontal  rows.
Figure A17.  Method of recognizing a horizontal row of spots.


-------
                                                        Horizontal  Row
Draw the Oth horizontal row.
Draw a perpendicular through the origin.

Join the origin to the first spot to the right
of the perpendicular in the 1st row and extend
the line to measure the acute angle 6 of this
line from the Oth row.
       A18.  Relationship of di, da, 61^2, and R.

-------
                        '.

                       • p
                                                                       ti*

      (a)  Zone axis [lOO].
(b)  Zone axis  [301].

      (c) Zone axis [101],
(d)  Zone axis [101].
Figure A19.  Typical Zone-axis SAED patterns from amosite standard specimen.
                            (Jones et al., 1981)
                                   92

-------
1
                    (a) Zone axis  [100]
(b)  Zone axis  [101].

                    (c) Zone axis [lio].
(d)  Zone axis  [301].
             Figure A20.  Typical zone-axis patterns from crocidolite  standard specimen.
                                         (Jones et al.,  1981)
                                                 93

-------
                (a) Zone axis [100].
(b)  Zone axis [101J.
                (c) Zone axis [201].
(d)  Zone axis [301].
\
           Figure A21.  Typical zone-axis patterns from  tremolite standard specimen.
                                      (Jones et al.,  1981)
                                              94

-------
       (a)  Zone axis [100],
(b) Zone axis [H»2].
                                        :£511680  ftNTHOP  C2
                                       •nil  fiDD .618KEV/CH
                                       
-------
                                                         APPENDIX B
                                    COMPUTER'PRiNTODT; OP  LEVEL  I ANALYSIS (EXAMPLE) '••
                                   •>.'.•=?: --£ Y::v:V~"  -   '                            •'?     '.•
                                     _  —•»* i ^   -   ^                                    *
vO
a*
                                           '   '
I IT RESEARCH INSTITUTE. STRUCT.URfcfANALYSIS  DATA
INDIVIDUAL OBJECT}. DATA :TA'DLE:>.".(F=F:IBER j . B=BlJNDL.E.f C
TABLE  PREPARATION" DATE:^21^fifeR-i|i;;;>- -,1. j«5;;;^-  /•.;'. -.
                                                                    =CLUSTERf  M=MATRIX)  -->.
::= ss ^
=:a = 3S
SAMPLE
(3rd
Opn
.1
1
2
•p
2
3
3
3
4
4
7
7
7
8
8
9
12
12
12
13
13
13
14
14
14
15
15
1A
JA

ObJ
1
2
3
4
S
A
7
0
V
10
11
12
13
14
15
1A
17
18
19
20
21
22
23
24
= = = =:
CODE

Str
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F-'
F
F
25' F
2A
2?
28
29
F
F
F
)•"
0.000  0. 3 75 V-;.-j;2V.l-9 ."
0.000" 0.7SO•-£*$'••
0.000
0.000  0.. 937  /-3-..00 \
0.000  0.125.r-:4l'V2E
0.000  0.250 . "V-'.
0.000  0.812 •-:j,'CiS;'COO
                                                                         .307
                                                                        0.04A
                                                                        1 . 300
                                                                        5.464
                        0.000 0
                        o.ooo
                        0. OQO-/0 . 3.12 >:t;'l.i 07. :•
                                                                                        Not
                                                                                        As be

                                                                                          X
                                                                                          X
                                                                               X
                                                                               .X
                                                                               •
                                                                               X
                                                                               X
                                                                                          X
                                                                                          x
                                                                              '".X

                                                                               X
                                                                               f
                                                                               X
                                                                               «.
                                                                               X
                                                                               X
                                                                               •
                                                                               x
                                                                               X
                                                                               x
                                                                              • x
                                                                                   : NO.
                                                                                   .;Patt -.-X-Ray
                                "_.         *>"      ^• * t .   ~ "    ••         •

-------
                                              APPENDIX B (Continued)
VO
            TIT  RESEARCH  INSTITUTE  STRUCTURE ANALYSIS  DATA
            SINGLE  SAMPLE  SUMMARY TABLETS
            SAMPLE  CODE:                             TABLE PREPARATION  BATE:  SI-
= = = = = = = = = =r= = as±:== = = = s
Aerosol Object Count
Object
Structure?
Tiber
rifHTlF'lC* Co]
Tape
Chrwsotile
Other
All Fiber
1 net ion find
fli r Vc>] ume -• 1 .
J.ieF-osit Ares -: 1.
A?ih*?d Are ft = I .
Kedoposit Aresi =: 1.
= K = = sssrsss = =: = ==: = =:=:=:=:si=i=: = =ri3:j5-:
And Calculated Object Mass C
Actual
Object
Count
0.
1 2 .
22.
34.
Prepnrat
00 Cu M
00 So Cm
('0 So Citi
00 So C,m
Number
Conc:r?n.
( Number
Per Cu M)
0.
SJ304.
15223.
23529.
inn Iiatu

haracterj sties
Mass
Concen. Avprjisle
(Pi cos! ram Width
Per i;u M) (Micron)
0
19155
Grjd H
1 nd j vi
U'jmber
I'llm ;-i
.0 0.00 :k 0.00
.4 0.38 .1. 0.21
0.46 J: 0.37
0.43 .1: 0,32

riual Hrid iipenin.«l
of Grid Or- en in. -.IG
sun i rirst i on

Average
Length
(Micron)
0.00 .1: 0.00
S.20 .1 4,50
3.V.9 1 3.24

= 0.000072
- 20
= '..;0000

Aver-aae
Length
To Width
Ft a t.i.o
0.00 :t
15.^5 1
7.81 .i
10.47 i

Go Cm


0.00
1.3.02
5.03
9.27


-------
                                    . "  --- , •'•--•  • •r.  .
                                     ;^,:;^;;:  f- ;£ APPENDIX B  (Continued)
                                   1 " ' !•*' r-~ * * •* ''  * '< T-
          .TIT RESEARCH  INSTITUTE STRUCTURE ANALYSIS  DATA
          INDIVIDUAL OBJECT DATA TABLED! (F=F.I:BERr  B=BUNDLE,  C=CLUSTER»  M=MATRIX)
          TABLE  PREPARATION DATE: 21-A:PR-B1  .j?-'
          SAMPLE  CODE:
vo
00
                       Size v•-•-	-.    ^""-Tj71 :'•   •.'•  •' ~ri    r
          OFT, ObJ Str Depth Width Lerjath .."Rat'io  Chrasptile  Amphibole   Ambid   Asbe   Patt   X-Rs«.
1A  30  F  0.000 0.
17  31  F  0.000 0
19  32  F  0.000 6.
                                                                         5.591
,!• f  «I^IM  I    ^v v ^f ^f v v, v A v* r  .-t^'viv*--   f f-v t m  -- P«B  v-   ->       w
19  33  F   0.000 .O.OA2-  .i:.'&'? -:  27V6 •.' j:''".':.-'.«  .  .
19  34  F   0.000 0.187-.;--.'^5|^. ,:; 10^V3  .'' ;-••.'•!:• '-' :i      0.1/0.

               Total  Ma'sis  .'/•jP.fcc6dpam!i)=   ~-':;    O.OOO'.^     27.680
               Total  Coun-t.-i'.^^.; .;:•  .w,|:.=    f-.:.:-9!\   :'"      12f
                                                                                           ..
                                                                                           X
                                                                                          . X
                                                                                           X
                                                                                         22.
                                                                                      0.

                                      /;Vt,-   f:'"1  ; :
                                       '--•: '.'« :      f .
                                         -
                                     •*?^-.??'---jr'.  .-••-I-
                                      "/v^---^.-'-; #•  '   .'•

                                     -F=:rt;.~ar = =.!rf!==.:==!-'r!=
                                     .  -.  •;-;;.<>.;?   :


-------
                                          APPENDIX C

                       COMPUTER PRINTOUT  OF LEVEL II ANALYSIS (EXAMPLE)


TIT RESEARCH INSTITUTE STRUCTURE  ANALYSIS DATA
INDIVIDUAL OBJECT DATA TABLE   (F=FIRER, B=E«UNDLF_. r:=t;Ll)STfTR, M = MATRIX)
TABLE PREPARATION PATE! 21-APR-81

SAMPLE CODE!
Size (Micron)
                                        Mass < Pi cosrsni)
lira
(Jpn
1

1
3

"S 3
\O

4

5
5

6
/
7
7
7
8
8

9
10
10
ObJ
1

2
3

4


5

A
7

tt
V
10
1 1
12
1 3
14

15
16
17
Str
F

F
f"

F


F

F
F

F
F
F"
F'
F
F
F

F
F
K
Depth
0

0
0

0


0

0
0

0
0
0
0
0
0
0

0
0
0
,000

.000
,000

.000


.000

.000
.000

.000
. 000
. 000
.000
.000
.000
.000

.000
.000
.000
Uicith t.ensth
0,

0,
0.

0.


0.

0.
0.

0.
0.
0.
0.
0.
0.
0.

0.
0.
0.
Total
125

562
625

375


250

062
187

125
125
937
625
312
375
187

437
250
250
Mass
1 . 63

1.87
5.31

4.06


1 .44

0.50
3.00

1.87
0.69
5.00
5 . 00
2.01
2 . 6 9
1 . 56

1 . 63
1,56
2.JH
— 	 - 	 - 	 not
Ratio Chrysotile Amphibole Anibia Asbe
13.0

3.3
0.5

10.8


5.8

8.0
16.0

1 5 . 0
S . 5
5. IS.
8.0
9.0
7.2
B..?

3.7
A.:.1
9 . t:;
(f-'icosl rstni)r-
lolaJ L'o'jnt •=
0.0598

1 . 3977
4.8892

1.3460

.
0.2117

• • t X
0.2485

• * • X
• t . • . .
X
• # « A
• • • '
• • • X
0.1294

• • « X
X
• • * *
0.0000 0.:.'82.?
0. 7, 0. .10.
no
rat*. X-Ray
MG(9)
r;A(7>
'I CGUI
MG(35)
CA( 35)
MG(19)

CA(22)
. MG(66)
FE(2B)
•
. MG(57)
CA<43>
•
•
•
t
•
•
MG
CA<18)
•
•
»

0.

s;i ( 19)
FE ( '/ )

Sl(llO)
F-T< IV)
r>I(59)

Ft: (21 >
!JI(152)


SI (170)
rr.( LA)






KK5U
r i: ( 6 >







-------
                                           A^APPENDIX c (continued)
         .TIT RESEARCH INSTITUTE STRUCTURE ANALYSIS DATA
         SINGLE SAMPLE SUMMARY TABLES'^Li '  ;.:.". f
         SAMPLE CODE:                '?^i   •'-• i   ™BLE PREPARATION DATE: 2i-APR-ei
         SSSSSESSSSSSSSISSISSSSSSSSrSSSSSSSSS SSSZSSSS=S=SS=X = S=SSS = SSSSSSSS SSSSSSSZSS^ 33 SSSiSS = = == = = SSSSIS! a =;=:=: = = =::
         Aerosol Object Count And Calculated Object Mass Characteristics
         Object
         Structure  Tape
                                            Mass                            .        Average
                        Actual'.'?: Concert.     Concern.     Average       Average       Length
                        Object.--!^(Number     (Pico^ram   Width         Length        To  Width
                        County •('
-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing/
 REPORT NO.
                            3.
                                                          3. RECIPIENT'S ACCESSION NO.
 TITLE AND SUBTITLE
 Methodology  for  the Measurement of Airborne Asbestos
 by Electron  Microscopy
                                    5. REPORT DATE
                                      July 1984
                                     i. PERFORMING ORGANIZATION CODE
 AUTHOFHS)
 George Yamate,  Satish C.  Agarwal, Robert D. Gibbons
                                    B. PERFORMING ORGANIZATION REPORT No
                                      C06470
 PERFORMING ORGANIZATION NAME AND ADDRESS
 IIT Research Institute
 10 West  35th Streeet
 Chicago, Illinois  60616
                                                          10. PROGRAM ELEMENT NO.
                                    11. CONTRACT/GRANT NO.
2. SPONSORING AGENCY NAME AND ADDRESS
 Environmental  Monitoring Systems Laboratory
 Office of Research and Development
 U.S. Environmental Protection Agency
 Research  Triangle Park, North Carolina   27711
                                                           13. TYPE OF REPORT AND PERIOD COVERED
                                     14. SPONSORING AGENCY CODE
5. SUPPLEMENTARY NOTES
6. ABSTRACT
      The  provisional electron microscope methodology for measuring the concentration
 of airborne asbestos fibers was refined.  The methodology is divided  into  separate
 protocols.   The step-by-step procuedures for each protocol are nearly identical,  so
 that cumulative data can be obtained and uncertainties, especially in asbestos  iden-
 tification,  can be clarified.  The operational  steps encompass (1) type of sample,
 (2) collection and transport, (3) sample preparation, (4) examination under  the
 transmission electron microscope (TEM) and  data collection, (5) data  reduction  and
 reporting of results, and (6) quality control-quality assurance.

      The  TEM analytical protocol is subdidvided into three levels of analysis:
 Level  I,  for screening many samples; Level  II,  for regulatory action; and  Level III,
 for confirmatory analysis of controversial  samples.   Because identification  of
 asbestos  structures is critical, the level  of analysis is directly related to the
 information sought:

         Level I—morphology and visual selected  area electron diffraction  (SAED)
                  pattern recognition.
        Level II—morphology; visual SAED; and elemental analysis.
      Level III—morpholgy; visual SAED, a selected number of SAED micrographs
                  of zone-axis patterns; and  elemental analysis.	
                               KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                              >. IDENTIFIERS 'OPEN ENDED TERMS
                                                                           COSATI held Croup
18. DISTRIBUTION STATEMENT

  RELEASE TO PUBLIC
                        19 SECURITY CLASS (Tins Report;
                          UNCLASSIFIED
21 NO OF FACES
     112
                                              20 SECURITY CLASS iTInspagci
                                                UNCLASSIFIED
                                                                         22. PRICE
Er-A Form 2220-1 
-------