EPA/540/2-90/005a
                                  OERR Directive 9285.5-02-1
                                        February 1990
Environmental Asbestos Assessment Manual
  Superfund Method for the Determination
        of Asbestos in Ambient Air
              Part 1: Method
              INTERIM VERSION

-------
                            DISCLAIMER
!™ "7/J& report wa&'prepaYGd under contract tqtbg &£ Environmental
^(otecf/o/t !4g&noy, T/3e,(?7efj$d/jub? ?ffi!cfe or vjotptnefatef products \
fnot constitute U& Environmental Protection Agency endorsement or


-------
                                         Contents

Figures   	   vj
Tables  	'.'.'.'.'.'.'.'.','.'.'.   vii
Acknowledgements	   vj(-j

1. Introduction   	                      1
2. Background  	.....'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.	   3
      2.1. Sensitivity	   	   5
      2.2. Sampled Air Volumes	           	'	' '  ' '   R
      2.3.  Filter Selection  	',',',	   7
      2.4. Precision	   9
      2.5. Asbestos Characteristics  	'.'.'.'.'.''.'.'.'.'.	  10
      2.6. Cost Considerations	   10

3. Overview of Method  	            -12
      3.1. Sample Collection  	           12
      3.2. Sample Preparation  	        12
        3.2.1 Indirect TEM Specimen Preparation  	' '   12
        3.2.2 Direct TEM Specimen Preparation  	   13
      3.3. Analysis	   13

4. Scope and Field of Application   	        15
      4.1. Substance Determined  	    15
      4.2. Range	    15
        4.2.1 Upper Limit of Range  	    15
        4.2.2 Lower Limit of Detection 	    1 g
      4.3. Analytical Sensitivities   	  _     17
      4.4. Dimensional Detection Limits  	    17

5. Definitions  	           1Q

6. Symbols and Abbreviations  	     22
      6.1. Symbols	      22
      6.2. Abbreviations   	       22

7. Equipment and Apparatus  	     24
      7.1. Air Sampling - Equipment and Consumable Supplies	    24
        7.1.1. Filter Cassette   	     24
        7.1.2. Sampling  Pump   	    24
        7.1.3 Stand	    24
        7.1.4 Rotameter  	    24
      7.2 Specimen Preparation Laboratory 	    25
      7.3 Laboratory Equipment	    25
        7.3.1. Transmission Electron Microscope  	    25
        7.3.2. Energy Dispersive X-ray Analyzer  	    27
        7.3.3. Computer  	   28
        7.3.4. Plasma Asher   	    28
        7.3.5. Filtration Apparatus    	   28
        7.3.6. Filtration Manifold  	   28
        7.3.7. Vacuum Pump  	   29

-------
        7.3.8.  Vacuum Coating Unit   	   29
        7.3.9.   Sputter Coater	   29
        7.3.10. Solvent Washer (Jaffe Washer)   	   29
        7.3.11. Condensation Washer   	   31
        7.3.12. Slide Warmer or Oven   	   31
        7.3.13. Ultrasonic Bath  	   31
        7.3.14. Carbon Grating Replica   	   31
        7.3.15. Calibration Grids for EDXA 	   31
        7.3.16. Carbon Rod Sharpener   	   31
        7.3.17. Disposable-Tip  Micropipettes   	   31
      7.4. Consumable Laboratory Supplies   	   33
        7.4.1. Glass Beakers  	   33
        7.4.2. Membrane Filters  	   33
        7.4.3. Copper Electron Microscope Grids  	   33
        7.4.4. Gold Electron  Microscope Grids  	   33
        7.4.5. Carbon Rod Electrodes   	   33
        7.4.6. Disposable Tips  for Micropipette  	   33
        7.4.7. Routine Electron Microscopy Tools and Supplies  	   34
        7.4.8. Reference Asbestos Samples	   34

8 Reagents                                                                         35
      8.1. Freshly-Distilled Water  	   35
      8.2. Dimethyl Formamide, Analytical Grade	   35
      8.3 Glacial Acetic Acid, Analytical Grade	   35
      8.4. Acetone, Analytical Grade   	   35
      8.5. Hydrochloric Acid,  Analytical Grade  	   35

9. Air Sample Collection  	   38
      9.1. Required Sensitivity  	   38
      9.2. Air Volume  	   38
      9.3. Row Rate  	   39
      9.4. Sampling Procedures 	   39

10. Procedure for Analysis  	   41
      10.1. Introduction  	   41
         10.1.1. Indirect TEM Specimen Preparation Method	   41
         10.1.2. Direct TEM Specimen Preparation Method  	   42
      10.2. Preparation of TEM Specimen Grids  by the Indirect Method 	   42
         10.2.1. Cleaning of Sample Cassettes   	   42
         10.2.2. Ashing of MCE Filter  	   42
         10.2.3. Re-dispersal  of Ashed Residues  	   42
         10.2.4. Rltration of the Aqueous Suspension  	   43
         10.2.5 Selection of Area of Filter for Preparation  	   45
         10.2.6. Preparation of Solution for Collapsing MCE  Filters  	   45
         10.2.7. Filter Collapsing Procedure 	   45
         10.2.8. Carbon Coating of Filter Sectors	   45
         10.2.9 Preparation of the Jaffe Washer	.•	   46
         10.2.10. Placing of Specimens Into the Jaffe Washer	   46
         10.2.11. Rapid Preparation of TEM Specimens from MCE Filters   	   46
                                               IV

-------
       10.3. Preparation of TEM Specimen Grids by the Direct Method	   46
         10.3.1. Cleaning of Sample Cassettes   	   46
         i 0.3.2. Selection of Area of Filter for Preparation   	   47
         10.3.3. Filter Collapsing Procedure  	   47
         10.3.4. Plasma Etching of the Filter Surface   	   47
         10.3.5. Carbon Coating of Filters   	   47
         10.3.6. Preparation of the Jaffe Washer  	   47
         10.3.7. Placing of Specimens into the Jaffe Washer  	   47
         10.3.8. Rapid Preparation of TEM Speciimens from MCE Filters	   47

       10.4. Criteria for Acceptable TEM Specimen Grids   	   47

       10.5. Procedure for Structure Counting by TEM  	   48
         10.5.1 Introduction   	   48
         10.5.2. Measurement of Mean Grid  Opening Area  	   49
         10.5.3. TEM Alignment and Calibration Procedures  	   49
         10.5.4. Determination of Stopping Point  	   49
         10.5.5. General Procedure for Structure Counting and Size Analysis 	   50
         10.5.6. Measurement of Concentration for Asbestos Structures
         Longer than 5 pm   	   54

       10.6. Blank.and Quality Control  Determinations  	   57
       10.7. Calculation of Results	   58

11. Performance Characteristics  . .	    59
       11.1. Interferences and Limitations of Structure Identification  	   59
       11.2. Precision and Accuracy	    59
         11.2.1. Precision	    59
         11.2.2. Accuracy  	    59
       11.3. Analytical Sensitivity 	    60
       11.4. Limit of Detection  	    61

12. Reporting Requirements   	    62
       12.1. Sample Analysis Report  	    62
       12.2. Sample Batch Report  	' '  .    53
       12.3. Data Review Report 	    64

Appendix A - Determination of Operating Conditions for Plasma Asher   	    71
Appendix B - Calibration Procedures  	    72
Appendix C - Structure Counting Criteria  	    75
Appendix D - Fiber Identification Procedure   	    84
Appendix E - Calculation of Results	    99
Appendix F - Bibliography  	    107

-------
7.1            Calibration Markings on the TEM Viewing Screen



7.2            Design of a Solvent Washer (Jaffe Washer).



73            Design of a Condensation Washer



10.1            Structure Counting Form        ;•




10.2            Scanning Procedure for a TEM Specimen Examination




12.1A   Format for Reporting of Structure Counting Data, Page 1



12. IB   Format for Reporting of Structure Counting Data, Page 2



12.1C          Format for Reporting of Structure Counting Data, Page 3



12.2            Format for Summary Laboratory Report



12.3            Format for Data Review Summary Report




C.1            Fundamental Morphological Structure Types



C.2            Examples of Recording of Complex Asbestos Clusters



C.3            Examples of Recording of Complex Asbestos Matrices



C.4            Counting of Structures that Intersect Grid Bars



C.5            Counting of Structures that Extend Outside the field of View



D.I            Measurement of Zone Axis SAED Patterns




D.2            Classification Chart for Fiber With Tubular Morphology



D.3            Chrysotile SAED Pattern



D.4            Classification Chart for Fibers Without Tubular Morphology
                                                VI

-------
9.1


10.1

D.1

P.?
Examples of the Minimum Number of Grid Openings Required to Achieve a Particular
Analytical Sensitivity and Detection Limit

Specifications for Phase 1 and 2 Sampling and Analysis

Classification of Fibers With Tubular Morphology

Classification of Fibers Without Tubular Morphology
                                                 VII

-------
                                   ACKNOWLEDGEMENTS
       We would like to acknowledge the timely and critical support for this effort provided by Jean
Chesson, Chesson Consulting, Washington D.C. and Kenny Crump, ICF Clement, Huston Louisiana.
Kenny Crump also assisted directly with the preparation of the Sections of this document addressing
data manipulation. Acknowledgements are also extended to David Suder, ENVIRON Inc., Emeryville
California for his assistance during the preparation of the sampling sections of this document.
                                             VIII

-------
1.  INTRODUCTION

      This is a sampling and analysis method for the determination of asbestos
in air.  Samples are analyzed by transmission electron microscopy (TEM).
Although a small subset of samples are to be prepared using a direct
procedure, the majority of samples analyzed using this method will be prepared
using an indirect technique.  The method allows for the determination of the
mineralogical type(s) of asbestos present and for distinguishing asbestos from
non-asbestos minerals.  In the method, asbestos structures are characterized
as fibers, bundles,  clusters,  or matrices and the length and width of each
asbestos structure are measured.  Although the method Is designed specifically
to provide results suitable for supporting risk assessments at Superfund
sites, it is applicable to a wide range of ambient air situations.

      To support a risk assessment,  this method addresses two objectives:

      (a)   to provide increased precision at the low concentrations of
            asbestos typically found in the environment;

      (b)   to provide measurements  that can be compared with risk factors
            derived from existing epidemiology studies.

An additional consideration addressed in this method is the need to control
sampling and analysis costs.

      The method focuses on sampling requirements for individual sampling
stations and the analysis of sample  filters collected at such stations.
During a site investigation, sampling stations would be arranged in an array
designed to distinguish spatial trends in airborne asbestos concentrations.
Sampling schedules would be fashioned to establish temporal trends.   Thus
proper design of a comprehensive sampling strategy, detailing the design of
the array of sampling locations and  the schedule for sample collection, is
also critical to the success of an investigation.  However, design of a
sampling strategy is necessarily site specific and site-specific
considerations are beyond the scope  of this document.

      Satisfying the two method objectives listed above requires innovations
that tax the limits of available technology.  Consequently, several variations
were considered during development and this method represents a workable
compromise among several technical constraints.  Although the method has  not
been validated as a whole and the feasibility of a few procedures needs to be
better documented, many of the component procedures of this method have been
performed in the laboratory.  Thus,  the principle features of the method  are
well enough established that the method can be profitably employed in current
field investigations.  There is no better way to acquire the necessary
experience and data for completing method validation.  At the same time,  until
a validation study and appropriate pilot studies are completed, this should be
considered an interim method.

-------
      Details of the Considerations addressed during the development of this
method are provided in a companion Technical Background Document, Part 2 of
this report, under separate cover.  A summary is presented in Chapter 2
(Background) of this document.

      NOTE

      This document is intended to serve several audiences including site
      proj ect managers, field sampling teams,  data reviewers,  and laboratory
      analysts.  The document may be separated into segments s6 that
      individuals may focus on the sections of most interest to their
      particular roles in a project.

-------
2. , BACKGROUND
      During the development of this method, existing sampling and analysis
technologies were considered to determine an appropriate approach for
achieving defined method objectives while remaining within the known
constraints associated with the measurement of airborne asbestos.   The method
is designed to provide the analytical sensitivity and precision necessary to
distinguish background from asbestos concentrations typical of environmental
contamination.  The method is also designed to provide results recorded in a
format that allows comparison with published risk factors or other proposed
representations of the biological activity .of asbestos.   :

      Recent developments toward understanding the relationship between
asbestos exposure and risk indicate that, to assess risks, modern field
measurements should be tailored to quantify asbestos characteristics that best
relate to biological activity.  This is true despite the fact that existing
risk factors are derived from epidemiology studies in which asbestos
concentrations were derived primarily by phase contrast microscopy (PCM).   A
broader range of asbestos structures potentially contribute to biological
activity than can be detected by PCM.

      To provide the needed capabilities for distinguishing the broadest range
of asbestos characteristics, transmission electron microscopy (TEM)  was
selected as the analytical technique employed in this method.  Although
considered, both PCM and scanning electron microscopy (SEM) were rejected for
use in this method due to their inherent limitations.

      The method is designed to provide the best description possible of the
nature, numerical concentration, and sizes of asbestos-containing particles
found in an air sample.  The method of data recording specified in the method
is designed to allow re-evaluation of the structure characterization and
counting data without the necessity for re-examination of the specimens.

      Currently-recognized methods require collection of airborne  particulate
on a membrane filter, followed by preparation of TEM specimens from the
filter.  Airborne asbestos is collected either on mixed cellulose  ester (MCE)
filters or polycarbonate filters.  TEM specimens can be prepared from such
membrane filters either by a direct-transfer technique or by an indirect
method.

      Direct-transfer TEM specimen preparation methods have the following
advantages:

      (a)   the particulate and fiber size distributions are undisturbed
            during specimen preparation;

      (b)   the limited amount of specimen manipulation reduces the
            possibility of fiber loss or introduction of extraneous
            contamination.

-------
However, the direct-transfer TEM preparation methods also have some
significant disadvantages:

      (a)   the achievable detection limit is restricted by the particulate
            density on the filter, which in turn is controlled by the sampled
            air volume and the total suspended particulate concentration in
            the atmosphere being sampled;

      (b)   the precision of the result is dependent on the uniformity of the
            deposit of asbestos structures on the sample collection filter;

      (c)   air samples must be collected that have particulate and fiber
            loadings within narrow ranges.  If too high a particulate loading
            occurs on the filter, it is not possible to prepare satisfactory
            TEM specimens by a direct-transfer method.  If too high a fiber
            loading occurs on the filter, even if satisfactory TEM specimens
            can be prepared, accurate fiber counting will not be possible.

      Indirect TEM specimen preparation techniques permit some of the
disadvantages of direct-preparation to be overcome:

      (a)   air samples can be collected without regard to the amount of
            deposit on the filter surface.  The filter loading can be adjusted
            in the laboratory to provide satisfactory TEM specimens;

      (b)   interfering particulate material can be completely or partially
            removed through a combination of dissolution and oxidation
            (ashing);

      (c)   the uniformity of distribution of asbestos structures on the
            filters to be analyzed is improved.

Indirect TEM specimen preparation methods have the following disadvantages:

      (a)   the size distribution of asbestos structures is modified;

      (b)   there is increased opportunity for fiber loss or introduction of
            extraneous contamination;

      (c)   when sample collection filters are ashed, any fiber contamination
            in the filter medium is concentrated on the TEM specimen grid.

      The question as to whether direct or indirect specimen preparation
yields the "correct" result (in terms of representing biological activity)  is
currently unresolved.  It can be argued that the direct methods yield an
under-estimate of the asbestos structure concentration because many of the
asbestos fibers present are concealed by other particulate material with which
they are associated.  Conversely, the indirect methods can be considered to
yield an over-estimate because some types of complex asbestos structures
disintegrate during the preparation, resulting in an increase in the numbers
of structures counted.

-------
 2.1.   SENSITIVITY

       To  provide  the  required sensitivity at asbestos  levels  typically
 found in  the  environment,  it  is necessary either  to selectively
 concentrate asbestos,  or  to count  structures over a much greater area of
 a TEM specimen grid than  has  traditionally been required.  To the extent
 that  it can be employed,  selective concentration  is the least costly of
 the alternatives  for  increasing sensitivity.  The optimal asbestos
 concentration on  a  filter requires filter loadings of  total particulate
 matter in excess  of what  can  generally be tolerated for analysis using
 direct-transfer methods for TEM specimen preparation.  Consequently, an
 indirect  technique  for TEM specimen preparation is employed in this
 method, which incorporates steps allowing for the selective
 concentration of  asbestos.

       In  most ambient  environments, a proportion  of the suspended
 particulate is organic, consisting of soot, spores, pollens and other
 debris from vegetation.   Organic materials such as these can  be removed
 from  the  analysis by low-temperature ashing.

       It  is common  to  find substantial numbers of calcium sulfate fibers
 (gypsum)  in airborne particulate collected in urban environments.  These
 can arise from various sources, and they can also be generated either in
 the atmosphere or on the  sample collection filter by reaction of
 airborne  calcite or dolomite particles with atmospheric sulfur dioxide.
 Gypsum can be removed by  dissolution in water or  dilute acids.

       Carbonates are another major component of most exterior
 atmospheres.   Calcite and  dolomite are commonly found in urban
 atmospheres;  these  originate from various sources, including  erosion of
 concrete and  cement, local geology, and in some cases from industrial
 operations.   Such carbonates can be removed by extraction with
hydrochloric  acid.  If acid extraction procedures are carried out
 correctly, no  chemical or  crystallographic degradation of asbestos can
be detected by routine methods of TEM analysis.

      Removal  of a  large proportion of the suspended particulate by use
of these techniques permits the asbestos present in the sample to be
concentrated  on to  a smaller area of the TEM specimen.   Consequently,
 the area of the TEM specimen that must be examined to achieve a
particular analytical sensitivity is proportionately reduced.   Also,
many of the non-asbestos fibrous structures normally found in a
directly-prepared TEM specimen, each of which must be identified and
rejected from  the asbestos structure count,  do not appear on the
indirectly-prepared TEM specimen.

      For many environmental samples,  a selective concentration of the
asbestos structures, incorporating low temperature ashing,  re-suspension
in water,  and acidification with HC1 is capable  of removing substantial
amounts of interfering particulate from the analysis,  and,  for a

-------
particular analytical sensitivity, reducing the area of the TEM
specimens that must be examined.

      The concentration steps included in this method, together with the
requirement to measure very low airborne asbestos concentrations, make
control of asbestos contamination more critical than for specimens
prepared by direct-transfer techniques.  Accordingly, the procedures in
this method have been designed to minimize the effects of contamination.

2.2.  SAMPLED AIR VOLUMES

      This method incorporates an indirect preparation procedure to
provide flexibility in the amount of deposit that can be tolerated on
the sample filter and to allow for the selective concentration of
asbestos prior to analysis.  To minimize contributions to background
contamination from asbestos present in the plastic matrices
(polycarbonate or mixed cellulose ester) of membrane filters while
allowing for sufficient quantities of asbestos to be collected,  this
method also requires the collection,of a larger volume of air per unit
area of filter than has traditionally been collected for 'asbestos
analysis.

      Due to the need to collect large volumes of air, higher sampling
flow rates are recommended in this method than have generally been
employed for asbestos sampling in the past.  As an alternative,  samples
may be collected over longer time intervals but this restricts the
flexibility required to allow samples to be collected while uniform
meteorological conditions prevail.  Higher flow rates through a  25 mm
filter result in increased face velocities at the filter surface.
Potential problems associated with the higher flow rates and increased
face velocities include:

      (a)   destruction of the filter due to failure of its physical
            support under force from the increased pressure drop;

      (b)   leakage of air around the filter mount so that the filter is
            bypassed;

      (c)   damage to the asbestos structures due to  increased impact
            velocities.

      The recommended flow rates and face velocities have been employed
in air sampling using membrane filters  in the past.  Based on such
studies,  (c) is not likely to be a problem.  Using the filters and flow-
rates specified, the first two concerns  (a and b) are also unlikely to
impose limitations.  However, documented experience  at the increased
flow rates with the types of filter cassettes to be  used in this method
is very limited.  Consequently, a pilot study is recommended to  confirm
that the structural integrity of a 25 mm filter cassette is not  altered
at the increased flow rates recommended.  There is also some question

-------
 whether available pumps can maintain such flow rates throughout the
 collection of large volumes of air while the filter loading and,
 correspondingly,  the resistance to flow steadily increase.

 2.3.   FILTER SELECTION

       Selection of the type of filter to be used in this method was
 determined by considering the performance of the filter during sample
 collection,  sample handling,  and TEM specimen preparation (by an
 indirect technique).   Concerns include the degree of asbestos loss  that
 may occur during sample transport or handling,  the efficiency with  which
 asbestos is transferred from the filter to the TEM specimen,  and the
 level of asbestos contamination potentially introduced to the TEM
 specimen from the filter material itself.   Both mixed cellulose ester
 (MCE)  and polycarbonate filters were evaluated.

       It is generally accepted that deposited particles may move  on the
 surface of polycarbonate filters when these filters  are heavily loaded.
 Such  movement may occur either during sample transport or handling  in
 the laboratory.   Such movement may potentially lead  to asbestos losses
 between the time  samples are  collected and the  time  they are  prepared
 for analysis.

       Unlike the  smooth surface of a polycarbonate filter,  MCE filters
 exhibit sponge-like surfaces  that trap filters within layers  of porous
 material.   The potential for  movement of deposits  on the surface  of this
 type  of filter is  generally considered minimal.  Correspondingly, the
 potential  for loss  of asbestos  during transport  or handling is  also
 reduced.   Thus, MCE filters are preferred  over polycarbonate  filters  in
 terms  of ease of handling.

       Two  methods  are available for  removal  of sampled particulate  from
 the sample collection filter  in an indirect  preparation:  complete ashing
 of  the  filter and  the particulate  deposit,  or washing  of the  particulate
 deposit from the filter  surface followed by  ashing of  the wash-suspended
 deposit only.  A number  of  problems have been experienced with  the
 ashing  of  filters during indirect-transfer procedures.   The efficiency
 with which deposited  asbestos is  transferred by washing has not been
 thoroughly investigated.

      Complete ashing  of a  filter  causes any asbestos contamination that
 may be present within  the polymer matrix of the filter to be  transferred
 to the  final TEM preparation.   Direct-transfer preparations from
 polycarbonate  filters  are known to yield a measurable background of
 asbestos contamination and current knowledge indicates that ashing of
polycarbonate  filters will yield an even higher value.  Direct-transfer
preparations of MCE filters generally yield low background asbestos
contamination.  However, ashing of either filter may yield higher levels
of background asbestos contamination due to sources of contamination
associated with the preparation itself.

-------
      Problems that have been experienced in association with low
temperature ashing include both potential sources of additional
contamination and filter behavior potentially leading to asbestos
losses.  During filter ashing, background contamination can arise from
within the asher chamber, the containers in which the filters are ashed,
the distilled water supply used for re-dispersing the ash,  the pipettes
used for transfer of dispersion for filtration, and the filtration
funnel.

      During ashing, both MCE and polycarbonate filters have
occasionally been observed to curl up and give the appearance of being
completely ashed, yet asbestos is trapped in a few very large,
incompletely ashed particles that do not disperse.  Consequently, such
asbestos is lost and does not appear on the final TEM preparation.
Procedures for avoiding this problem need to be refined.

      Due to the problems associated with ashing the filters, washing of
the particulate from the filter surface into ethanol is considered a
superior approach.  It is considered that washing of particulate from
the surface of a polycarbonate filter can be accomplished with high
removal efficiency.  However, contamination of polycarbonate filters by
asbestos has been measured by various laboratories, and currently
appears to be within the range of 50-200 asbestos structures/sq.mm. on
the active filter surface.1

      During the washing of the filter surface, asbestos contamination
on the surface of a polycarbonate filter (observed on preparations from
direct-transfer procedures) may be detached along with the collected
particulate and contribute to the background against which the
concentrations of any collected asbestos must be measured.   If the
filter is removed from the cassette and the washing of the filter is
performed ultrasonically in a beaker, contamination from both sides of
the filter may be detached, leading to higher background counts.
Therefore, potential but unquantified problems may be associated with
the washing of polycarbonate filters.

      Currently, it is not known whether collected particulate can be
efficiently removed by washing the surfaces of 0.45 /zm or 0.8 /zm pore
size MCE filters.  The MCE filter has a sponge-like texture, and many of
the fibers collected are embedded or entrapped within this structure.
It has been shown that any attempt to use ultrasonic techniques results
in disintegration of the filter surface and release of fragments of
      Certain laboratories have consistently reported lower values for
      polycarbonate filter contamination than other laboratories.
      Although this is well known and the matter has been addressed by
      numerous investigators, the source of the apparent discrepancy has
      not been elucidated.

                                 8

-------
filter material.   It is possible that this disintegration of the filter
surface may also release most of the collected particulate,  but this has
not been demonstrated.   On the basis of current knowledge, it is
unlikely that the small amount of polymer removed from the surface of
the filter would contribute a significant level of contamination to the
analysis, because contamination levels on the surface of MCE filters
have been demonstrated to be very low.

      Based on the above considerations, MCE filters were selected for
use in this method.  Although available data suggest that the optimum
combination of filter and preparation technique would be for samples to
be collected on MCE filters and to re-disperse the particulate in water
by washing of the filters,  little is known concerning the efficiency
with which asbestos can be removed from the surface of an MCE filter by
washing.  It is known that some of the filter polymer is detached from
the filter surface during washing and particles of this polymer can
interfere with the analysis.  Thus, ashing of the wash suspension would
be necessary in order to remove the filter polymer fragments.  Because
the efficacy of the filter washing procedure has not been adequately
substantiated, the preparation procedure incorporated into this method
is ashing of the filter followed by re-dispersion of the ash in water.

2.4.  PRECISION

      To maximize precision, potential sources of variation within the
method must be controlled.   To minimize systematic variation during
sampling and analysis,  the method specifies detailed procedures.  Random
variation is minimized by maximizing the number of structures to be
identified and counted, which includes increasing the probability of
encountering asbestos structures during analysis, as discussed in
Section 2.1.  In addition,  detailed and unambiguous counting rules and
rules for identification are specified in the method to minimize
variation due to subjective interpretation.

      Airborne asbestos is often found, not as single fibers, but as
very complex structures that may or may not be aggregated with equant
(non-asbestos) particles.  The mineralogy of structures found suspended
in an ambient atmosphere must be identified to confirm a structure as
asbestos.  Such structures can often be identified unequivocally, if
sufficient measurement effort is expended.  However, if each structure
were to be identified in this way, the analysis becomes prohibitively
expensive.  Because of instrumental deficiencies or because of the
nature of the particulate,  some structures cannot be positively
identified as asbestos, even though the measurements all indicate that
they could be asbestos.  Subjective factors therefore contribute to this
measurement, and consequently a very precise procedure for
identification and enumeration of asbestos is incorporated in this
method.

-------
      Despite the imposition of systematic counting and identification
criteria, there are large potential errors in characterizing asbestos in
ambient atmospheres, associated with the variability between filter
samples and the performance of individual microscopists.  For this
reason, a replicate sampling and analysis scheme is also incorporated in
order to determine the accuracy and precision of the method.

2.5.  ASBESTOS CHARACTERISTICS

      To analyze asbestos samples so that results can be related to
potential risks, it is necessary to quantify the characteristics of
asbestos that relate to biological activity.  This involves enumeration
of individual structures within certain size categories with particular
emphasis on the longest and thinnest structures.   Although the range of
dimensions over which asbestos structures contribute to biological
activity has yet to be precisely defined, this method is designed to
provide a detailed characterization of structures encompassing the
entire range of potential importance.

      When evaluating detailed asbestos size characterizations, it is
important to consider the effects of sample preparation.  The appearance
of the size distribution representing a sample of asbestos  may vary
depending on whether the sample was prepared by a direct or an indirect
technique.  Existing risk factors are based largely on studies
incorporating direct transfer techniques.2  Because indirect preparation
is preferred to achieve the desired analytical sensitivity for
environmental analysis (see Section 2.1), a procedure for evaluating the
relationship between structure counts on samples  prepared by an indirect
technique and by a direct transfer technique is included in the method.

2.6.  COST CONSIDERATIONS

      In environmental samples, the amount of airborne asbestos is
usually low relative to the total suspended particulate concentration.
Therefore, when using a direct-transfer preparation, large areas of the
TEM specimen grids must be examined to obtain a statistically valid
measurement for asbestos.  Consequently,  the analysis is expensive.   It
is therefore not feasible to perform this type of analysis on all of the
filters collected as part of a site investigation.

      A cost-effective compromise has been incorporated into this
method.  A two phase sampling and analysis scheme has been adopted.   In
Phase 1,  the majority of samples collected at a site will be prepared by
      PCM analyses are performed directly on sample filters where the
      filter material has been rendered transparent by a suitable
      solvent.  Since the fibers are observed as originally deposited,
      this corresponds closely to a direct transfer technique for TEM
      analysis.

                                10

-------
the indirect-transfer technique (to maximize precision) and analysis
will be performed using an.efficient evaluation procedure.  Air volumes
collected on Phase 1 samples may be maximized subject to the limitations
of sampling equipment.

      In Phase 2 a small subset of samples (5 to 10%) will be collected
and analyzed using an extended TEM examination to derive a relationship
between counts from samples prepared by a direct technique and counts
from samples prepared by an indirect technique for each of the various
structure size fractions of interest that are observed in Phase 1 and
Phase 2 samples.  Filters from Phase 2 samples will be split so that
half may be prepared by the direct-transfer procedure and half by the
indirect-transfer procedure.  Because half of each filter to be analyzed
under Phase 2 will be prepared by the direct-transfer technique,  the
volume of air collected for each sample is limited by the loading that
can be tolerated during analysis of a specimen prepared using a direct-
transfer technique.  Phase 1 and 2 samples shall be collected simul-
taneously .
                                11

-------
3.  OVERVIEW OF METHOD

      Samples are collected and prepared for TEM examination by one of two
techniques depending on the purpose for the sample.  The majority of air
samples (denoted Phase 1 samples) will be analyzed using an indirect procedure
for preparation of TEM specimens, optimized to provide low detection limits
and high precision.  A small number of samples (denoted Phase 2 samples) will
be collected in such a way that they can be analyzed using both indirect and
direct procedures for preparation of TEM specimens, to allow comparisons to be
made between the results from the two specimen preparation procedures.   TEM
examination procedures used for the two sets of samples also differ.

      3.1.  SAMPLE COLLECTION

            A sample of airborne particulate is collected by drawing a
      measured volume of air through a 25 mm diameter, 0.45 p,m pore size MCE
      membrane filter by means of a pump. Air volumes collected on Phase 1
      samples will be maximized.  Air volumes collected on Phase 2 samples
      will be limited to provide optimum loadings for filters to be prepared
      by a direct-transfer procedure.

      3.2.  SAMPLE PREPARATION

      TEM grids will be prepared according to either 3.2.1 or 3.2.2 below.

            3.2.1.  Indirect TEM Specimen Preparation

                  This preparation will be applied to all Phase 1 and Phase 2
            samples.   The filter is split and half of the filter is stored
            for future use.  The remaining half of the filter is ashed in a
            low-temperature plasma asher.  The residual ash is ultrasonically
            dispersed in freshly-distilled water.  The suspension is acidified
            using hydrochloric acid, and immediately filtered through a 25 mm
            diameter, 0.1 /jm pore size MCE filter.  The filter is dried and
            the filter structure is collapsed using a mixture of dimethyl
            formamide, acetic acid and water.  A thin film of carbon is
            evaporated onto the collapsed filter surface and small areas are
            cut from the filter.  These areas of filter are supported on TEM
            specimen grids and the filter medium is dissolved away by a
            solvent extraction procedure.

            NOTE

            An alternate procedure for indirect preparation in which filter
            deposits are washed off of MCE filters followed by the ashing of
            the wash-suspended deposit may be substituted into this method
            upon completion of validation in a pilot study.
                                      12

-------
      3.2.2.  Direct TEM Specimen Preparation

            The remaining one-half of the filters from all Phase 2
      samples will be prepared by this procedure.  One quarter of the
      filter is collapsed using a mixture of dimethyl formamide, acetic
      acid and water.  The collapsed filter is etched for a short time
      in a low temperature plasma asher to remove the surface layer of
     . filter polymer which may have encapsulated asbestos structures
      during the collapsing procedure.  A thin film of carbon is
      evaporated onto the collapsed filter surface and small areas are
      cut from the filter.  These areas of filter are supported on TEM
      specimen grids and the filter medium is dissolved away by a
      solvent extraction procedure.

3.3.  ANALYSIS

      The TEM specimen grids are examined at both low and high
magnifications to check that they are suitable for analysis before
carrying out a quantitative examination on randomly-selected grid
openings.  In addition to isolated fibers, ambient air samples often
contain more complex aggregates of fibers, with or without equant
particles.  Some particles are composites of asbestos fibers with other
materials.  Individual fibers and these more complex structures are
referred to as "asbestos structures".  A coding system is used to record
the type of fibrous structure, and to provide the optimum description of
each of these complex structures.

      The method requires that separate examinations be made for
asbestos structures of all sizes (incorporating asbestos fibers with
lengths greater than 0.5 /im) and for long asbestos structures
(incorporating structures longer than 5 /*m).  In both cases, asbestos
structures are defined as those exhibiting mean aspect ratios equal to
or greater than 5:1.  This TEM examination procedure allows for
specification of a lower analytical sensitivity for the measurement of
the concentration of asbestos structures longer than 5 jum.

      In the TEM analysis, electron diffraction (ED) is used to examine
the crystal structure of a fiber, or fibrous components of complex
structures, and the elemental composition is determined by energy
dispersive X-ray analysis (EDXA).  For a number of reasons, it is not
possible to identify (determine the mineralogy of) each structure
unequivocally, and structures are classified according to the techniques
that have been used to identify them.  A simple code is used to record
the manner in which each structure is classified.

      The classification procedure is based on successive inspection of
the morphology, the electron diffraction pattern, and the energy
dispersive X-ray spectrum.  Confirmation of the identification of
chrysotile is only by quantitative ED, and confirmation of amphibole is
only by quantitative EDXA and quantitative zone-axis ED.

                                13

-------
      Several levels of analysis to confirm mineralogy are specified,
the higher levels providing a more rigorous approach to the
identification of structures.  The procedure permits a minimum required
asbestos identification procedure to be defined on the basis of previous
knowledge, or lack of it, about the particular sample.  Attempts are
then made to achieve this defined minimum procedure for each asbestos
structure, and the degree of success is recorded for each.  The two
codes remove from the microscopist the requirement to interpret
observations made during the TEM examination, and allow this evaluation
to be made later without the requirement for re-examination of the TEM
specimens.

      The lengths and widths of all classified asbestos structures are
recorded.  The number of asbestos structures found on a known area of
the TEM specimen grids, together with the equivalent volume of air
filtered through this area, are used to calculate the airborne
concentration in asbestos structures/liter of air.

      This method requires that the desired analytical sensitivities for
the measurements of:

      (a)   asbestos structures of all sizes (incorporating structures
            longer than 0.5 /*m) ;  and,

      (b)   asbestos structures longer than 5 pin,

shall be specified before air samples are collected.  In both cases,
structures to be counted are defined as those exhibiting aspect ratios
equal to or greater than 5:1.  It will not always be possible to achieve
these analytical sensitivities, because the volume of air that can be
sampled is dictated by the nature and concentration of the suspended
particulate in the atmosphere being sampled.  To some degree, this
limitation can be overcome by selective concentration of asbestos
structures during the specimen preparation procedures and by examination
of a larger area of the TEM specimens.  However, the ease and cost of
achieving a specific value for the analytical sensitivity will vary from
sample to sample.
                                14

-------
4.  SCOPE AND FIELD OF APPLICATION

      This method defines procedures for analyzing of airborne asbestos
collected by drawing a known volume of air through mixed cellulose ester (MCE)
filters.  The method is suitable for the determination of asbestos in ambient
air.

      4.1. SUBSTANCE DETERMINED

            The method defined uses transmission electron microscopy for
      determination of the concentration of asbestos in ambient atmospheres.
      This method includes measurement of the lengths and widths of asbestos
      structures and a calculation of their aspect ratios.  In the method,
      asbestos structures are characterized as fibers,  bundles, clusters,  or
      matrices.  The method allows for the determination of the mineralogical
      type(s) of asbestos present.  The method,  however,  cannot discriminate
      between individual fibers of the asbestos  and non-asbestos analogues  of
      the same amphibole mineral.

      4.2.  RANGE

            The range of concentrations that can be determined on a TEM
      specimen grid is approximately 50 to 7000  asbestos  structures/mm2.  The
      lower limit is generally defined by the requirement that the
      concentration of asbestos collected from the air  be distinguishable  from
      background contamination contributed by the method and the filter.  The
      upper limit is defined by the concentration at which asbestos structures
      will significantly overlap and interfere with proper detection.

            4.2.1.   Upper Limit of Range

                  The upper limit of the range for detecting asbestos  on a
            filter may be limited both by excessive overlap of asbestos
            structures and by the presence of other interfering particles.  In
            practice, if obscuration of asbestos structures is not to  be a
            significant factor in a method's precision,  the final TEM  specimen
            grid should not exhibit more than approximately a 10% coverage  by
            particulate.   For the indirect TEM specimen preparation method,
            this maximum permissible particulate loading  is determined by the
            concentration of that portion of the total  suspended particulate
            concentration that cannot be removed by procedures such as
            oxidation or chemical dissolution.

                  Using the direct-transfer preparation method,  it has been
            found that if the total particulate  loading of the sample
            collection filter exceeds approximately 10  jig/cm2  of  filter
            surface,  corresponding to approximately 10% coverage of the
            collection filter by particulate,  the TEM specimens are loaded  to
            a point that obscuration of asbestos structures by particulate
            becomes a significant  source of error.   If  the particulate loading

                                      15

-------
exceeds approximately 25 /ig/cm2 of filter surface,  the  specimens
are seriously overloaded and it is often not possible to prepare
TEM specimens that have grid openings with intact carbon film.

      The degree that non-asbestos particulates may interfere with
asbestos analysis is a function of the concentration of total
suspended particulate (TSP) present in the atmosphere in addition
to the asbestos being sampled.  Concentrations of the TSP vary
over several orders of magnitude depending on location, time of
day, and weather.  Urban and agricultural sites tend to have
significantly higher concentrations of the TSP than rural
locations.  Consequently, higher volumes of air may be collected
at rural locations before interference by the TSP becomes a
limiting factor.

      In addition to variation in overall concentration, the
composition of the TSP varies significantly as a function of
location.  At urban sites and specific rural locations, the TSP
tends to be composed principally of organic matter that can be
ashed or inorganic substances that are soluble in acidified media.
Agricultural locations and other rural locations frequently
exhibit higher concentrations of refractory silicate particles.

      Due to the wide spatial and temporal variation in the TSP
concentrations, a general rule for estimating levels can not be
provided.   However, a review of available environmental studies
indicates that air volumes of 1 m3 per cm2 of  filter  area may
reasonably be collected for direct preparation samples in most
urban and rural environments.  Subject to the types of the TSP
present, samples collected for indirect preparation can be loaded
to levels up to an order of magnitude higher.   Data on the range
of local levels of TSP should be obtained and evaluated prior to
finalizing a sampling plan for a site.

4.2.2.  Lower Limit of Detection

      The detection limit theoretically can be lowered
indefinitely by:

      (a)   filtration of progressively larger volumes of air;

      (b)   use of intermediate specimen preparation procedures to
            concentrate the collected particulate on to smaller
            areas of the TEM specimen grids; and,

      (c)   examination of greater areas of the TEM specimens in
            the electron microscope.

      Although it might appear that the detection limit can be
lowered indefinitely by examination of a greater area of the TEM

                          16

-------
      specimens,  this is not actually the case if unused filters
      (blanks),  or the analytical procedure itself,  contribute
      background asbestos.   Although it is considered that the back-
      ground asbestos structure counts of TEM specimens prepared by the
      collapsed MCE filter  technique (direct preparation) are very low,
      there are few data currently available concerning the asbestos
      content of the MCE filter polymers themselves,  which might
      contribute to background during an indirect preparation.

            In a technique  that incorporates complete ashing of MCE
      filter membranes,  any asbestos contained in the filter polymer
      will be concentrated  on the final TEM specimens.   Accordingly,
      this effect should be minimized by collection of the maximum
      volumes of air possible,  and a continuous program of blank
      measurements must form part of the quality assurance procedures.

4.3.  ANALYTICAL SENSITIVITIES

      Based on a review of  data from past environmental asbestos
investigations and published studies of rural and urban background, the
lower end of the ranges of  concentrations typical of these types of
studies center on 5 structures/liter (s/L) for asbestos structures of
all sizes and 0.2 s/L for structures longer than 5 /un.   The significance
of the asbestos structures  longer than 5 /im is discussed in Section 2.5.
To distinguish among such concentrations at acceptable levels of
statistical significance, it is assumed that a minimum of 10 asbestos
structures should be identified and counted (see Section 4.2 of the
Technical Background Document,  Part 2 of this report).   Thus, analytical
sensitivities (defined as the concentration represented by the detection
of 1 structure)  required by this method are 0.5 s/L and 0.02 s/L for
asbestos structures of all  sizes and asbestos structures longer than
5/im, respectively.

4.4.  DIMENSIONAL DETECTION LIMITS

      Microscopists vary in their ability to detect very small asbestos
structures.  Therefore,  a minimum length of 0.5 ^m has been defined as
the shortest fiber to be incorporated in the reported results.

      The minimum detectable width for asbestos structures counted in
this method is determined by the ability of an operator to detect them
in a routine examination at the specified magnification of the method.
The minimum magnification specified in this method (10,000) is more than
sufficient to ensure visibility of the thinnest chrysotile asbestos
fibrils.
                                17

-------
5.  DEFINITIONS

Acicular:  The shape shown by an extremely slender crystal with cross-
sectional dimensions which are small relative to its length.

Amphibole:  A group of rock-forming ferromagnesium silicate minerals, closely
related in crystal form and composition, and having the nominal formula:
where:

            A - K, Na;

            B - Fe2*,  Mn,  Mg,  Ca,  Na;

            C - Al, Cr, Ti, Fe3+,  Mg,  Fe2*;

            T - Si, Al, Cr, Fe3+,  Ti.

In some varieties of  amphibole, these elements can be partially substituted by
Li, Pb, or Zn.  Amphibole is characterized by a cross-linked double chain of
Si-0 tetrahedra with  a silicon:oxygen ratio of 4:11, by columnar or fibrous
prismatic crystals and by good prismatic cleavage in two directions parallel
to the crystal faces  and  intersecting at angles of about 56° and 124°.

Amphibole asbestos:   Amphibole in an asbestiform habit.

Analytical sensitivity:  The calculated airborne concentration, in asbestos
structures/liter, equivalent to counting of one asbestos structure in the
analysis.

Asbestiform:  A specific  type of  mineral fibrosity in which the fibers and
fibrils possess high  tensile strength and flexibility.

Asbestos:  A term applied to a group of fibrous silicate minerals that readily
separate into thin, strong  fibers that are flexible, heat  resistant and
chemically inert.

Asbestos component:   A term applied to any individually identifiable  asbestos
sub-structure that is part  of a larger asbestos structure.

Asbestos structure:   A term applied to any contiguous grouping of asbestos
fibers, with or without equant particles.
 Aspect ratio:   The ratio  of length to  width of a particle.
                                       18

-------
Blank:  A fiber count made on TEM specimens prepared from an unused filter, to
determine the background measurement.   Blanks consist of filter blanks, field
blanks and laboratory blanks.

Bragg Angle:  The angle between a crystal surface and an incident ray of
incoming x-radiation.

Bundle:  A fiber composed of parallel, smaller diameter fibers attached along
their lengths.

Camera length:  The equivalent projection length between the specimen and its
electron diffraction pattern, in the absence of lens action.

Chrysotile:  A mineral of the serpentine group which has the nominal composi-
tion:
                   Mg3Si205(OH)4

In some varieties of chrysotile, the silicon may be partially substituted by
Al or less commonly by Fe.  The magnesium may be partially substituted by  Fe,
Ni., Mn or Co.  Some varieties contain Na, Cl or both.  Chrysotile is a highly
fibrous, silky variety, and constitutes the most prevalent type of asbestos.

Cleavage:  The breaking of a mineral along one of its crystallographic
directions.

Cleavage fragment:  A fragment of a crystal that is bounded by cleavage faces.

Cluster:  An assembly of randomly oriented fibers.

Component count: For any sample, a tally that includes the individually
identified components of complex asbestos structures and each single asbestos
structure with no identifiable components.

d-spacing:  The distance between identical adjacent and parallel planes of
atoms in a crystal.

Detection limit:  The calculated airborne fiber concentration in fibers/liter,
equivalent to counting of 3.69 fibers in the analysis.

Electron diffraction:  A technique in electron microscopy by which the crystal
structure of a specimen is examined.

Electron scattering power:  The extent to which a thin layer of substance
scatters electrons from their original directions.

Energy dispersive X-ray analysis:  Measurement of the energies and intensities
of X-rays by use of a solid state detector and multi-channel analyzer system.

Eucentric:  The condition when an object is placed with its center on a
rotation or tilting axis.

                                      19

-------
Fiber:  An elongated particle which has parallel or stepped sides.  In this'
method, a fiber is defined to have an aspect ratio equal to or greater than
5:1.

Fibril:  A single fiber of asbestos, which cannot be further separated
longitudinally into smaller components without losing its fibrous properties
or appearances.

Fibrous structure:  A contiguous grouping of fibers, with or without equant
particles.

Field blank:  A filter cassette which has been taken to the sampling site,
opened, and then closed.  Such a filter is analyzed to determine the back-
ground asbestos structure count for the measurement.

Filter Blank:  An unused filter which is analyzed to determine the background
asbestos structure count.

Habit:  The characteristic crystal form or combination of forms of a mineral,
including characteristic irregularities.

Identify:  During analysis, the use of a sequential set of procedures to
determine and confirm the mineralogy of a structure.

Laboratory Blank:  An unused filter which is analyzed along with sample
filters to determine the background asbestos structure count.

Matrix:  A connected assembly of asbestos fibers with particles of another
species (non-asbestos).

Miller index:  A set of either three or four integer numbers used to specify
the orientation of a crystallographic plane in relation to the crystal axes.

PCM equivalent structure:  A structure of aspect ratio greater than or equal
to 3:1, longer than 5 pm, and which has a mean diameter between 0.2 pm and
3.0 pm for a part of its length greater than 5 pm.  In this method, PCME
structures also must contain at least one asbestos component.

Replication:  A procedure in electron microscopy specimen preparation in which
a thin copy, or replica, of a surface is made.

Selected area electron diffraction:  A technique in electron microscopy in
which the crystal structure of a small area of a sample is examined.

Serpentine:  A group of common rock-forming minerals having the nominal
formula:

                 Mg3Si205(OH)A
                                      20

-------
Structure Count:  For any sample, a tally of each individually identified
asbestos structure regardless of whether the structure contains identifiable
components.   This is equivalent to a count of the total number of separate
asbestos entities encountered on the sample.

Twinning:  The occurrence of crystals of the same species joined together at a
particular mutual orientation, and such that the relative orientations are
related by a definite law.

Unopened fiber:  A large diameter asbestos fiber bundle which has not been
separated into its constituent fibrils or fibers.

Zone-axis:  The line or crystallographic direction through the center of a
crystal which is parallel to the intersection edges of the crystal faces
defining the crystal zone.
                                      21

-------
6.  SYMBOLS AND ABBREVIATIONS




      6.1.  SYMBOLS




      A



      eV



      kV
      L/min
      run




      Pa




      s/L




      W






      6.2.




      DMF




      ED




      EDXA




      FWHM




      HEPA




      MCE




      PCM




      PCME




      SAED




      SEM
              Angstrom unit (10~10 meter)




              electron volt




              kilovolt




              liters per minute




              microgram (10~6  grams)




              micrometer (10~6 meter)




              nanometer (10"9  meter)




              Pascal




          - structures per liter




              watt
ABBREVIATIONS
              Dimethyl formamide




              Electron diffraction




              Energy dispersive X-ray analysis




              Full width, half maximum




              High efficiency particle absolute




              Mixed cellulose ester




              Phase contrast optical microscopy




              Phase contrast microscopy equivalent




              Selected area electron diffraction




              Scanning electron microscope
                                      22

-------
STEM




TEM




TSP




UICC
Scanning transmission electron microscope




Transmission electron microscope




Total suspended particulate




Union Internationale Centre le Cancer
                               23

-------
7.  EQUIPMENT AND APPARATUS
      7.1.  AIR SAMPLING - EQUIPMENT AND CONSUMABLE SUPPLIES

            7.1.1.  Filter Cassette

                  Commercially available field monitors,  comprising 25 mm
            diameter three-piece cassettes, with conductive extension cowls
            shall be used for sample collection.  The cassette must be new and
            not previously used.  The cassette shall be loaded with an MCE
            filter of pore size 0.45 fjaa., and supplied from a lot number which
            has been qualified as low background for asbestos determination.
            The filter shall be backed by a 5 pm pore size MCE filter, and
            supported adequately so that distortion of the filter by the
            differential pressure across it does not occur during sampling.
            The cassettes shall be purchased with the required filters in
            position, or shall be assembled in a laminar flow hood or clean
            area such that the filter blank determinations meet the criteria
            of Section 10.6.  When the filters are in position, a shrink
            cellulose band or adhesive tape should be applied to cassette
            joints to prevent air leakage.  Suitable precautions shall be
            taken to ensure that the filters are tightly clamped in the
            assembly so that significant air leakage around the filter cannot
            occur.

            7.1.2.  Sampling Pump

                  The sampling pump shall be capable of a flow-rate and
            pumping time sufficient to achieve the desired volume of air
            sampled.  The face velocity through the filter shall be between
            4.0 cm/s and 65.0 cm/s.  The sampling pump used shall provide a
            non-fluctuating air-flow through the filter,  and shall maintain
            the initial volume flow-rate to within ±10% throughout the
            sampling period.  A constant flow or critical orifice controlled
            pump meets these requirements.  Flexible tubing shall be used to
            connect the filter cassette to the sampling pump.  A means for
            calibration of the flow-rate of each pump is also required.

            7.1.3.  Stand

                  A stand shall be used to hold the filter cassette at the
            desired height for sampling and the filter cassette shall be
            isolated from the vibrations of the pump.

            7.1.4.  Rotameter

                  A rotameter or other flow measuring device calibrated such
            that the operator can measure  flow rates to ± 5% accuracy at the
            expected sampling flow rate.

                                      24

-------
7.2.  SPECIMEN PREPARATION LABORATORY

      Asbestos, particularly chrysotile, is present in varying
quantities in many laboratory reagents.  Many building materials also
contain significant amounts of asbestos or other mineral fibers that may
interfere with the analysis if they are inadvertently introduced during
preparation of specimens.  It is most important to ensure that during
preparation, contamination of TEM specimens by any extraneous asbestos
fibers is minimized.  All specimen preparation steps shall therefore be
performed in an environment where contamination of the sample is
minimized.  The primary requirement of the sample preparation laboratory
is that a blank determination shall yield a result which will meet the
specifications in 10.6.  A minimum facility considered suitable for
preparation of TEM specimens is a laminar flow hood.  However, it has
been established that work practices in specimen preparation appear to
be more important than the type of clean handling facilities in use.
Preparation of samples shall be carried out only after acceptable blank
values have been demonstrated.

NOTE

Activities involving manipulation of bulk asbestos shall not be
performed in the same area as TEM specimen preparation,  because of the
possibilities of contaminating the TEM specimens.

7.3.  LABORATORY EQUIPMENT

      7.3.1.  Transmission Electron Microscope

            A TEM operating at an accelerating potential of 80-120 kV,
      with a resolution better than 1.0 nm,  and a magnification range of
      approximately 300 to 100,000 shall be used.   The ability to obtain
      a direct screen magnification of about 100,000 is  necessary for
      inspection of fiber morphology;  this magnification may be obtained
      by supplementary optical enlargement of the screen image by use of
      a binocular if it cannot be obtained directly.  It is also requir-
      ed that the viewing screen be calibrated (as shown in Figure 7.1)
      with concentric circles and a millimeter scale such that the
      lengths and widths of fiber images down to 1 mm width can be
      measured in increments of 1 mm.

            For Bragg angles less than 0.01 radians, the TEM shall be
      capable of performing ED from an area of 0.6 /zm2 or  less,  selected
      from an in-focus image at a screen magnification of 20,000.   This
      performance requirement defines  the minimum separation between
      particles at which independent ED patterns can be  obtained from
                                25

-------
Figure 7.1:    Calibration Markings on the TEM Viewing Screen
                                 26

-------
 each particle.   If SAED is used,  the  performance of a particular
 instrument may  normally be calculated using the following
 relationship:

            A - 0.7854  x (D/M + 2000  x  CS93)2

 Where:  A    -  the  effective SAED area in fan2
        D    -  the  diameter of  the  SAED aperture  in fun
        M    -  the  magnification of the objective lens
        C,    -  the  spherical aberration coefficient of the
                    objective lens in mm
        9    -  maximum required Bragg  angle in radians

      It  is not possible to reduce the effective SAED area
 indefinitely by the use of progressively smaller SAED apertures,
 because there is a fundamental limitation imposed by the spherical
 aberration coefficient  of the objective  lens.

      If  zone-axis ED analyses are to  be performed,  the  TEM shall
 incorporate a goniometer stage which permits  the TEM specimen to
 be either:

      (a)   rotated through 360°, combined with  tilting  through at
            least +30"  to -30° about an  axis  in  the  plane of the
            specimen; or,

      (b)   tilted through at least +30"  to  -30°  about two
            perpendicular axes in the  plane of the specimen.

      The analysis  is greatly facilitated if  the  goniometer
 permits eucentric  tilting, although this  is not  essential.  If
 EDXA and zone-axis  ED are required on  the  same fiber, the
 goniometer shall be of  a type which permits tilting  of the
 specimen and acquisition of EDXA  spectra without  change  of
 specimen holder.

      The TEM shall have  an illumination and  condenser lens system
 capable of forming an electron probe smaller  than 250 nm in
 diameter.

      Use of an anti-contamination trap around the specimen is
 recommended if the required instrumental performance is to be
maintained.

 7.3.2.   Energy Dispersive X-ray Analyzer

      The TEM shall be equipped with an energy dispersive X-ray
analyzer capable of achieving a resolution better than 175 eV
 (FWHM)  on the Mnl^ peak.  Since the performance of individual
combinations of TEM and EDXA equipment is dependent on a number of

                          27

-------
geometrical factors, the required performance of the combination
of the TEM and X-ray analyzer is specified in terms of the
measured X-ray intensity obtained from a fibre of small diameter,
using a known electron beam diameter.  Solid state X-ray detectors
are least sensitive in the low energy region, and so measurement
of sodium in crocidolite shall be the performance criterion.  The
combination of electron microscope and X-ray analyzer shall yield,
under routine analytical conditions, a background-subtracted NaKa
integrated peak count rate of more than 1 count per second (cps)
from a 50 nm diameter fiber of UICC crocidolite irradiated by a
250 nm diameter electron probe at an accelerating potential of
80 kV.  The peak/background ratio for this performance test shall
exceed 1.0.

      The EDXA unit shall provide the means for subtraction of the
background, identification of elemental peaks, and calculation of
background-subtracted peak areas.

7.3.3.  Computer

      Many repetitive numerical calculations are necessary, and
these may be performed conveniently by relatively simple computer
programs.  For analyses of zone-axis ED pattern measurements, a
computer with adequate memory is required to accommodate the more
complex programs involved.

7.3.4.  Plasma Asher

      For ashing of filters or particulate deposits washed from
filters, a plasma asher, with a radio frequency power rating of
100 W or higher, shall be used.  The asher shall be supplied with
a controlled oxygen flow, and shall be modified, if necessary, to
provide a valve to control the speed of air admission so that
rapid air admission does not disturb the ashed particulate.  It is
recommended that filters be fitted  to the oxygen supply and the
air admission line.

7.3.5.  Filtration Apparatus

      After re-dispersal of the ashed particulate in water,
the particulate suspension is filtered through a membrane  filter
of 25 mm diameter.  A glass frit support is required in order to
obtain a uniform deposit on the filter.  The reservoir must be
easily cleaned in order to prevent  sample cross-contamination.  A
25 mm analytical filter holder  (Millipore Corporation, Cat. No.
XX10 025 00) or equivalent has been found to be suitable.

7.3.6.  Filtration Manifold

      When a number of samples are  to be filtered, several

                          28

-------
 filtration units can be operated simultaneously from a single
 vacuum source by using a multiple port filtration manifold
 (Millipore Corporation,  Cat.  No.  XX26  047  35  or equivalent has
 been found to be suitable).   The manifold  should include valves  to
 permit each port to be opened or closed independently.

 7.3.7.   Vacuum Pump

       A pump is required to provide  a  vacuum  of approximately 20
 kPa for the filtration of water  samples.   A water jet pump
 (Edwards High Vacuum Inc., Grand Island, NY 14072,  Cat. No.
 01-C046-01-000-female connection or  01-C039-01-000-male connection
 or equivalent) has  been found to provide sufficient vacuum for a
 3-port filtration mani-fold and  also incorporates a non-return
 valve to prevent back-streaming.

 7.3.8.   Vacuum Coating  Unit

       A vacuum coating  unit capable  of producing a vacuum  better
 than 0.013 Pa shall be  used for  vacuum deposition of carbon on the
 membrane filters.   A sample holder is  required  which will  allow  a
 glass microscope slide  to be  continuously  rotated during the
 coating procedure.   A mechanism  which  allows  the rotating  slide
 also to be tilted through an  angle of  approximately 45° during the
 coating procedure is recommended.  A liquid nitrogen cold  trap
 above the diffusion pump may  be  used to minimize the possibility
 of contamination of the  filter surfaces by oil  from the pumping
 system.   The  vacuum coating unit may also  be  used for deposition
 of the  thin film of gold, or  other calibration  material, when it
 is required on TEM  specimens  as  an internal calibration of ED
 patterns.

 7.3.9.   Sputter Coater

      A sputter coater with a gold target may be used for
 deposition of gold  on to TEM  specimens as  an  internal calibration
 of ED patterns.   Experience has shown  that a  sputter  coater allows
 better  control  of the thickness of the calibration material.

 7.3.10.   Solvent  Washer  (Jaffe Washer)

      The  purpose of the Jaffe washer is to allow dissolution of
 the  filter polymer while leaving an  intact evaporated carbon film
 supporting  the  fibers and other particles from the filter surface.
 One  design of a washer that has been found satisfactory for
various solvents and filter media is shown in Figure 7.2.   Several
 acceptable variations have been described in the scientific
 literature.  When specimens are not being inserted or removed, the
petri dish lid should be in place.  The washer should be cleaned
before  it  is used for each batch of specimens.

                          29

-------
          GLASS PETRI DISH
          (  lOOmm x 15mm )
ELECTRON MICROSCOPE
   SPECIMENS
                                                      ST. STEEL MESH
                                                      BRIDGE ( SOmesh)
         LENS
         TISSUE
Figure 7.2:     Design of a Solvent Washer (Jaffe Washer).
                Solvent is added until the meniscus contacts the underside of
                the stainless steel mesh
                                      30

-------
 7.3.11.   Condensation Washer

      For more  rapid dissolution of the filter polymer, a
 condensation washer, consisting of a flask, condenser and cold
 finger assembly, with a heating mantle and means for controlling
 the  temperature may be used.  A suitable assembly is shown in
 Figure 7.3, and acetone should be used as the solvent for MCE
 filters.

 7.3.12.   Slide  Warmer or Oven

      Either a  slide warmer or an oven shall be used for heating
 slides during the preparation of TEM specimens.  It is required to
 maintain  a temperature of 65-70°C.

 7.3.13.   Ultrasonic Bath

      An  ultrasonic bath is required for cleaning of apparatus
 used for  TEM specimen preparation, and for re-dispersal of ashed
 residues  from MCE filters.

 7.3.14.   Carbon Grating Replica

      A carbon  grating replica with about 2000 parallel lines
      per mm shall be used to calibrate the magnification of the
      TEM.

 7.3.15.   Calibration Grids for EDXA

      For calibration of the efficiency of the EDXA system,
 reference silicate standard material must be used.   The National
 Institute for Standards and Technology (NIST) Standard Reference
 Material  (SRM)  2063 is a sputtered thin film containing known
 concentrations  of magnesium, silicon,  calcium and iron.   Other
 suitable  calibration materials include riebeckite,  chrysotile,
 halloysite, phlogopite, wollastonite and bustamite.   The mineral
 used for  calibration of the EDXA system for sodium must be
 prepared  using  a gold TEM grid.

 7.3.16.   Carbon Rod Sharpener

      The use of necked carbon rods,  or equivalent,  allows the
 carbon to be evaporated on to the filters without creating
 sufficient heat to damage the filters.

 7.3.17.   Disposable-Tip Micropipettes

      A disposable tip micropipette,  capable of transferring a
volume of approximately 30 /^L,  is required for the  preparation of
TEM specimen grids.

                          31

-------
                              ADAPTER
  CONDENSER
  SPECIMEN
        •4
  COLD WATER
   SOURCE
                                   COLD FINGER
        a         c
WATER
DRAIN
                                 THERMOSTATICALLY
                                 CONTROLLED
                                 HEATING MANTLE
Figure 7.3:    Design of a Condensation Washer
                  .  32

-------
7.4.  CONSUMABLE LABORATORY SUPPLIES

      7.4.1.   Glass Beakers

            Borosilicate glass beakers of 100 mL capacity are required
      for ashing of filters and re-dispersal of ashed residues.

      7.4.2.   Membrane Filters

            For filtration of the particulate suspensions,  two types of
      filter are required:

            (a)   MCE filters of 0.1 pm pore size (Millipore Corporation
                  VCWP 025  00 or equivalent) are used to collect the
                  particles from the liquid suspension;

            (b)   MCE filters of 5 pm pore size (Millipore  Corporation
                  SMWP 025  00 of equivalent).   One of these filters is
                  placed between the glass frit of the filtration
                  apparatus and the 0.1 /tm pore size filter,  in order to
                  ensure a  uniform particulate deposit.

      7.4.3.   Copper Electron Microscope Grids

           Two-hundred mesh TEM grids are required.   Grids which have
      grid openings of uniform size such that they meet  the requirement
      of 10.4.2 should be chosen.

      7.4.4.   Gold Electron Microscope Grids

           Two-hundred mesh gold TEM grids are required to mount TEM
      specimens when sodium measurements are required in the  fiber
      identification procedure.   Grids that have grid openings of
      uniform size such that they meet the requirement of 10.4.2 should
      be chosen.

      7.4.5.   Carbon Rod Electrodes

           Spectrochemically pure carbon rods are required for  use in
      the vacuum evaporator during carbon coating of filters.

      7.4.6.   Disposable Tips  for  Micropipette

           Disposable  tips  for  a  micropipette are required for
      measurement of 30  /zL volumes  of the filter collapsing solution.
                               33

-------
7.4.7.  Routine Electron Microscopy Tools and Supplies

      Fine-point tweezers,  scalpel holders and blades, microscope
slides, double-coated adhesive tape, lens tissue, gold wire,
tungsten filaments, liquid nitrogen, and other routine supplies
are required.

7.4.8.  Reference Asbestos Samples

      Asbestos samples are required for preparation of reference
TEH specimens of the primary asbestos minerals.  The NIST SRM 1866
contains well-characterized samples of chryostile, crocidolite,
and amosite, although compositional data are not available for
these standards.  The UICC set of minerals may be used as
additional reference materials.
                          34

-------
8.  REAGENTS

      8.1.  FRESHLY-DISTILLED WATER

            A supply of freshly-distilled water must be available.  The water
      must be produced using an all-glass still.  Any water used for
      re-dispersal of ash, or for final rinsing of glassware must have been
      distilled no longer than 24 hours prior to use.

      8.2.  DIMETHYL FORMAMIDE, ANALYTICAL GRADE

      8.3.  GLACIAL ACETIC ACID, ANALYTICAL GRADE

      8.4.  ACETONE, ANALYTICAL GRADE

      8.5.  HYDROCHLORIC ACID, ANALYTICAL GRADE
WARNING - Use the reagents in accordance with the
             appropriate health and safety regulations.
                                      35

-------
9.  AIR SAMPLE COLLECTION

      Sampling procedures for this method were developed in consideration of
and designed to be compatible with the analytical requirements of this method
including, primarily, the need to achieve the desired analytical sensitivity.

      Analytical sensitivity is a function of the volume of air sampled, the
active area of the collection filter, the dilution or concentration factor
introduced during specimen preparation, and the area of the TEM specimen over
which asbestos structures are counted.  If total airborne dust levels are
high, it may be necessary to terminate sampling before the required volume has
been sampled.  If this happens, the analytical sensitivity required can be
achieved only by counting structures over a greater area of the TEM specimen,
or by use of a specimen preparation technique that incorporates some selective
concentration of asbestos structures relative to other particulate species.
For example, depending on the nature of the suspended particulate, a con-
centration factor up to about 2 may be possible for 25 mm filters, if the
entire filter is used in the analysis.  Larger concentration factors are
possible for larger sample filters.  In either case, the number of grid
openings to be examined will be proportionately reduced.

      The number of grid openings that must be examined in order to achieve a
particular analytical sensitivity, without any selective concentration of the
asbestos structures, is shown in Table 9.1.  An average grid opening size of
0.0081 mm2 is assumed in this table.  However, grid opening sizes vary
depending on the manufacturer of the grid.  When using this table to derive
the number of grid openings to be counted to achieve a defined analytical
sensitivity, the number of grid openings to be counted must be adjusted to
preserve the total areas represented.  For example, if a volume of 5000 liters
of air is collected, the table indicates that 96 grid openings must be counted
to achieve an analytical sensitivity of 0.1 s/L.  This corresponds to an area
of 0.78 mm2 (96 x 0.0081).   If the average grid opening size on another grid
is 80 pm square (resulting in an average grid opening area of 0.0064 mm2),
than a total of 122 grid openings will have to be counted, to cover the same
area and achieve the same analytical sensitivity.  Under no circumstances
should fewer than 4 grid openings be counted.

      The volume in liters required to achieve a specified analytical
sensitivity is calculated from the formula:
where:
   V  -  Af/(N x Ag x  S x  F)

Af  -  Active area of sample collection filter in mm2

N   -  Number of grid openings to be examined.  N shall be
       rounded upwards to the next highest integer

Ag  —  Mean area of grid openings in mm2

S   -  Analytical sensitivity in structures/liter

F   -  Concentration factor.
                                       36

-------
Table 9.1 - Examples of the Minimum Number of Grid Openings Required to Achieve
            a Particular Analytical Sensitivity
  Analytical
  sensitivity

     s/L
2000
             Volume of air sampled (liters)
5000
7000
10000
15000
20000
     0.01
     0.02
     0.03
     0.04
     0.05
     0.07
     0,
     0,
     0,
     0.4
     0.5
     0.7
     1.0
     2.0
     3.0
     4.0
     5.0
     7.0
    10.0
 2377
 1189
  793
  595
  476
  340

  238
  119
   80
   60
   48
   34

   24
   12
    8
    6
    5
    4
    4
 951
 476
 317
 238
 191
 136

  96
  48
  32
  24
  20
  14

  10
   5
   4
   4
   4
   4
   4
 680
 340
 227
 170
 136
  98

  68
  34
  23
  17
  14
  10

   7
   4
   4
   4
   4
   4
   4
  476
  238
  159
  119
   96
   68

   48
   24
   16
   12
   10
    7

    5
    4
    4
    4
    4
    4
    4
  317
  159
  106
   80
   64
   46

   32
   16
   11
    8
    7
    5

    4
    4
    4
    4
    4
    4
    4
  238
  119
   80
   60
   48
   34

   24
   12
    8
    6
    5
    4

    4
    4
    4
    4
    4
    4
    4
NOTES

In this table, it is assumed that the collection filter area is 385 mm2, there
is no concentration or dilution introduced during specimen preparation, and the
TEM grid openings are 90 jum square.

For grid openings with dimensions other than 90/jm square, the minimum number of
openings to be counted should be adjusted by the ratios of the squares of the
grid opening dimensions, such that the area of the TEM specimen examined remains
the same.
                                       37

-------
    The concentration factor F arises from the fact that,  when using an
indirect specimen preparation technique, the fibers on a particular area of
the TEM specimen have usually been transferred from a different area of the
original sample collection filter.  The concentration factor may correspond to
a concentration or a dilution of the original fiber density on the collection
filter.  It is calculated from the ratio between the areas of the original and
analytical filters, the volume of water used to re-disperse the ashed particu-
late, and the volume of the final dispersion used to prepare the analytical
filter.  For the purposes of calculating the required volume of air to be
sampled in order to achieve the specified analytical sensitivity when a direct
transfer preparation is to be used, the concentration factor (F) is assumed to
be 1.

    9.1.   REQUIRED SENSITIVITY

          The required analytical sensitivities  for characterizing environ-
    mental samples is established in section 4.3 as 0.5  s/L and 0.02 s/L for
    asbestos  structures of all sizes and asbestos structures longer than 5 /jm,
    respectively.   For other applications of this method,  different analytical
    sensitivities  should be established prior to sample  collection.   Air
    volumes to be  collected in this method are selected  to provide the  desired
    analytical sensitivity when coupled with the counting  of a manageable
    number of grid openings.

    9.2.   AIR VOLUME

          The sampling rate and the period of sampling shall be selected to
    yield as  high  a sampled volume  as possible,  which will minimize  the
    influence of filter contamination.   Wherever possible,  a volume  of  15
    cubic meters shall be  sampled for those samples intended for analysis only
    by the indirect TEM preparation method (Phase 1 samples).   For those
    samples to be  prepared by both  the  indirect  and the  direct  specimen
    preparation methods (Phase 2  samples),  the volumes must be  adjusted  so as
    to provide a suitably-loaded  filter for the  direct TEM preparation method
    (See  Section 4.2.1).   It  has  been found that the volume cannot normally
    exceed 5  cubic meters  in an urban or agricultural area,  and 10 cubic
    meters  in a rural  area for samples  collected on a 25 mm filter and
    prepared by a  direct transfer technique.

          Prior to  planning a sampling  program,  air volumes  to  be  collected
    should be  adjusted based  on available  information characterizing  typical
    local  TSP  concentrations  (see Section 4.2.1).   It has  not proven possible
    to judge  the loading of filters  in  real time  during  sampling so  that
    filters may easily become  overloaded to the point were  analysis by the
    direct method becomes  impossible.   To minimize  the potential that filters
   will have  to be rejected  for  analysis due to  overloading, conservative
    estimates  of the loading  rate based on  typical  TSP concentrations have
   been developed  so  that an  "optimum"  volume of air per unit  area of filter
   may be collected.   However, because  such estimates tend  to be  conservative
    to avoid overloading, a fair percentage of the  filters  analyzed from such

                                     38

-------
a sampling effort tend  to be underloaded forcing a corresponding increase
in analysis costs.

      NOTE

      One option is to  collect filters at several loadings to bracket the
      estimated optimum loading for a particular site.  Such filters can
      be screened in the laboratory so that only those filters closest to
      optimal loading are analyzed.

9.3.  FLOW RATE

      To remain within  known constraints on filter face velocities as
reported in Section 2.2, an upper limit to the range of acceptable flow
rates for this method is 15 L/min.  In practice, available pumps that can
be used for environmental sampling at remote locations operate under load
at a maximum of approximately 12 L/min.  Consequently, the recommended
15 m3 of air for Phase 1 samples requires approximately 20 hours to
collect.  Such pumps typically draw 6 amps at full power so that 2
lead/acid batteries should provide sufficient power to collect a full
sample.  The lower volumes required for Phase 2 samples can be collected
in shorter periods with a corresponding reduction in stored energy
requirements.  The use  of line voltage, where available, eliminates the
difficulties associated with transporting stored electrical energy.

      At many locations, wind patterns exhibit strong diurnal variations.
Therefore, intermittent sampling (sampling over a fixed time interval
repeated over several days) may be necessary to accumulate 20 hours of
sampling time over constant wind conditions.   Other sampling objectives
also may necessitate intermittent sampling.   The objective is to design a
sampling schedule so that samples are collected under uniform conditions
throughout the sampling interval.  This method provides for such options.
9.4.  SAMPLING PROCEDURES

      Before air samples are collected, unused filters shall be analyzed
as described in Section 10.6 to determine the mean background asbestos
structure count for the analytical procedure.

      Air samples shall be collected using cassettes as described in
7.1.1.  The cassettes used for sampling shall be new and unused.  Sampling
shall be conducted with the cassette open-face.  During sampling, the
filter cassette shall be supported on a stand so that it is isolated from
the vibrations of the pump.  The cassette shall be held facing vertically
downwards at a height of approximately 1.5-2.0 m above ground/floor level
and connected to the pump with a flexible tube.  In many cases it will
likely be sufficient to collect samples with a short cassette.  If
conditions dictate the need for additional protection, however,  extension
cowls may be affixed to the front of the cassette.  Such cowls must be

                                  39

-------
constructed of electrically conducting material to minimize electrostatic
effects.

      The sampling flow-rate should be measured at the cassette at the
beginning and at the end of the sampling period.  Sampling equipment shall
be selected such that either the sampling flow rate can be maintained to
within ± 10%, using a flow controller or critical orifice, or flow.rates
shall be measured at least every 4 hours.  If at any time the measurement
indicates that the flow-rate has decreased by more than 30%, the sampling
shall be terminated.  The mean value of these flow-rate measurements shall
be used to calculate the total air volume sampled.  After sampling,  a cap
shall be placed over the open end of the cassette, and the cassette packed
in a clean plastic bag for return to the laboratory.  Field blank filters
shall also be included, as described in Section 10.6, and processed
through the remaining analytical procedures along with the samples.

      Should intermittent sampling be required, sampling cassettes must be
covered between active periods of sampling.  To cover the sampler:  turn
the cassette to face upward, attach a rotameter to measure the flow rate,
turn off the sampling pump, and place the cap.  To resume sampling:  remove
the cap, turn on the sampling pump, attach a rotameter to measure the flow
rate, and invert the cassette to face downward.
                                  40

-------
10.  PROCEDURE FOR ANALYSIS

    10.1.   INTRODUCTION

   '       Fi.lters  representing Phase  1  samples will be prepared by  the
    indirect  method.  Filters representing Phase  2 samples will be  split.
    Half of the  Phase 2  filters will  be prepared  by the  indirect method and  a
    quarter sector of the filter will be prepared by  the direct method.

          10.1.1.   Indirect TEM Specimen Preparation  Method

                The indirect method requires collection  of particulate  from  a
          large  volume of air on a 0.45 /HB pore size  MCE filter.  In this
          procedure, half of the filter is ashed  by treatment in a  low
          temperature plasma asher, and the ashed residue is re-dispersed in
          distilled water.  Immediately before filtration, the  suspension of
          particles is acidified using  hydrochloric acid to remove  any
          acid-soluble species.  The  suspension is filtered through a 0.1 pra.
          pore'size MCE  filter.  The  filter is collapsed using  a mixture of
          dimethyl  formamide, acetic  acid and water.  A  sector  of the
          collapsed filter is carbon  coated, arid  TEM  specimen grids are
          prepared  from  the carbon coated filter by the  Jaffe washer
          technique, using dimethyl formamide as  the  solvent.

               Use of the 0.1 pm pore  size membrane  filter eliminates the
          requirement for the plasma  etching step incorporated  into both NIOSH
          Method 7402 and the method  of Burdett and Rood.  In previous work by
          Chatfield, Dillon and Stott,  no statistically-significant fiber
          losses could be detected when TEM grids were prepared from 0.1 /zm
          pore size MCE  filters without including the etching step.

               Indirect preparation of TEM specimens for determination of
          asbestos  is particularly sensitive to problems of contamination by
          extraneous asbestos.   This method requires that very low detection
          limits be achieved so that control of extraneous asbestos
          contamination becomes even more important.   The ability to meet the
         blank sample criteria is critically dependent on the cleanliness of
          equipment and supplies.   All supplies such as microscope slides and
          glassware should be considered as potential sources of asbestos
         contamination.   All solvents used should be filtered through a
         0.2 ^m pore size filter or they should be distilled in a glass
         still.   It is  necessary to wash all glassware before it is used and
         the final washing of glassware should be performed using
         freshly-distilled water.   Any tools or glassware that come into
         contact with the air sampling filters or TEM specimen preparations
         should be washed both before use and between handling of individual
         samples.   Where possible,  disposable supplies should be used.
                                     41

-------
      10.1.2.   Direct TEM Specimen Preparation Method

            The direct TEM specimen preparation method requires that the
      air volume sampled must not result in a filter that is over-loaded
      by particulate.  Particulate is  collected on a 0.45 /zm pore size MCE
      filter.   A quarter sector of the MCE filter is collapsed using a
      mixture of dimethyl formamide, acetic acid and water.   The collapsed
      filter is etched for a short time in a low temperature plasma asher.
      The collapsed and etched filter  is carbon coated,  and TEM specimen
      grids are prepared from the carbon coated filter by the Jaffe washer
      technique, using dimethyl formamide as the solvent.

10.2.  PREPARATION OF TEM SPECIMEN GRIDS BY THE INDIRECT METHOD

      10.2.1.   Cleaning of Sample Cassettes

            Asbestos can adhere to the exterior surfaces of air sampling
      cassettes and this asbestos can  be inadvertently transferred to the
      sample during handling.  To prevent this possibility of
      contamination, and after ensuring that the cassette is tightly
      sealed,  wipe the exterior surfaces of each sampling cassette with a
      clean, wet paper towel before it is taken into the clean facility or
      laminar flow hood.

      10.2.2.   Ashing of MCE Filter

            Open the filter cassette and remove the MCE filter.  Place the
      filter on a clean microscope slide and cut it in half, using a
      curved scalpel blade.  Return one half of the filter to the
      cassette.  Place the other half filter in the bottom of a 100 mL
      glass beaker and cover the top of the beaker with a piece of
      aluminum foil.  Make approximately 8 1-2 mm perforations in the
      aluminum foil.  Completely ash the half filter in the low
      temperature plasma asher, using operating conditions as defined in
      Appendix A.

      10.2.3.  Re-dispersal of Ashed Residues

            After the filter has been completely ashed, remove the beaker
      from the asher, add 40 mL of distilled water, and place the beaker
      in the ultrasonic bath for a period of 15 minutes.  Using a
      disposable plastic pipette, add 0.2 mL of concentrated hydrochloric
      acid and stir  the suspension.  After addition of the acid, place the
      beaker in the ultrasonic bath for a period of 5 minutes.  Filtration
      of the suspension must be performed immediately after the ultrasonic
      treatment.  Pov,er in the ultrasonic bath should be maintained below
      0.1 watt/mL.
                                  42

-------
 10.2.4.   Filtration of  the Aqueous  Suspension

      The separation of suspended particulate by  filtration of the
 aqueous  suspension through a membrane  filter is a critical  step in
 the  analytical  procedure.  The  objective  is to produce  an analytical
 filter on which the suspended solids from the sample  are distributed
 uniformly,  and  at  an acceptable loading for the microscopical
 examination.

      The volume to be  filtered generally depends  on  either the
 total suspended solids  content  or the  asbestos structure
 concentration of the sample, and the intention is  to  prepare a
 filter that has a  loading not exceeding about 50  asbestos structures
 per  grid opening and with a particulate coverage not  exceeding 10%
 of the filter area.   Some judgment  is  required to  achieve the
 optimum  loading and,  if the asbestos concentration is very  low,  it
 will be  found that the  suspended solids concentration will  limit the
 volume that can be filtered.  The maximum particulate loading  on the
 filter that can be tolerated is  about  10  fig/cm2.

      In practice,  the  best procedure  for optimizing  the loading on
 the  final filter is  to  prepare  several filters using  different
 volumes  of the  suspension.  Any fraction  of the 40 mL suspension
 (prepared as per Section 10.2.3) may be selected as a unit  for
 filtration.  For example, aliquots  containing 1 mL, 4 mL, 8 mL, and
 20 mL fractions  from the original suspension may be used to prepare
 a series  of increasingly loaded filters.  However, if the con-
 centration of suspended solids  dictates that a very small volume of
 the suspension be  filtered, small volume  aliquots need to be further
 diluted prior to filtering.  Do not attempt to filter a volume of
 less than 10 mL.

      If  aliquots  smaller than  10 mL from the original suspension
 are to be  filtered,  it  is difficult to ensure that a uniform deposit
 of particulate  is  obtained on the filter  unless the aliquot is first
 diluted.   Samples  of high solids content,  or of high asbestos
 concentration, may require filtration of volumes less than 10 mL.
 Such aliquots shall be  diluted with freshly-distilled water so that
 the filtered volumes exceed the minimum of 10 mL.   These dilutions
 shall be made by transferring a known volume of the sample  (such as
 an aliquot  from  the original suspension)  to a disposable plastic
beaker and making  the volume up to a known volume with
 freshly-distilled water.  The mixture shall be stirred vigorously
before sub-samples are  taken for filtration.

      The following instructions for filtration must be followed
precisely:

      (a)   Assemble the filtration base and turn on the vacuum.
            The upper surface of the filtration base (both the glass

                            43

-------
(b)
(c)
frit and the ground mating surface) must be dry before
the membrane filters are installed.  Place a 5 pm pore
size MCE filter on the glass frit.  If the filter
appears to become wet by capillary action on residual
water in the glass frit it must be discarded and
replaced by another filter.  Place a 0.1 pm pore size
MCE filter, smooth side facing up, on top of the 5 pm
filter.  The mating'surface of the reservoir component
of the filtration apparatus should be dried by shaking
off any surplus water and draining on paper towel or
tissue.  Position the reservoir on the filters and
firmly clamp it, taking care not to disturb the filters.
The vacuum should not be released until the filtration
has been completed.

It is necessary to comment- on the 'use of filtration
equipment that is still wet after washing, -since
improper procedures at this -point can seriously
compromise the results.  If the glass frit is wet when
the first filter is applied to it, capillary action will
result in some areas of the filter structure being
filled by water.  When the second, smaller pore size
filter is added and the vacuum is applied, the
differential pressure across the 5 yum pore size filter
will be insufficient to overcome the surface tension of
the water in the filled areas.  Thus no filtration will
take place through the corresponding areas of the 0.1 pm
pore size'filter, and a grossly non-uniform deposit of
particulate will be obtained.

Add the required volume of sample to the filtration
funnel.  Disposable plastic beakers and pipettes provide
a means of measuring the sample volumes without
introducing problems of sample'cross-contamination.  The
reservoir may not be sufficiently large to accommodate
the total volume of liquid to be filtered.  In this
case, more of the sample may be added during the
filtration, but this should be done carefully and only
when the reservoir is more than half full.  In this way
the addition of'liquid will riot disturb or affect the
uniformity of particulate already deposited on the
filter.  Do not rinse the sides of the reservoir, and
avoid other manipulations which may disturb the
particulate deposit on the filter.

Disassemble the filtration apparatus, and transfer the
0.1 pro pore size filter to" a labelled, clean plastic
petri dish.  Place the cover loosely over the top of the
dish to limit any deposition of material on the filter,
and allow the filter to dry.  Discard the 5 pm pore size
filter.      -....-•...•
                      44

-------
10.2.5.  Selection of Area of Filter for Preparation

      Using clean tweezers, and a clean, curved scalpel blade, cut
out a 90 degree sector of the filter while it is still in the
plastic petri dish.

10.2.6.  Preparation of Solution for Collapsing MCE Filters

      Mix 35 mL of dimethyl formamide, 15 mL of glacial acetic acid
with 50 mL of freshly-distilled water.  Store this mixture in a
clean bottle.

10.2.7.  Filter Collapsing Procedure

      Using a micropipette with a disposable tip, place
approximately 30 pL of the collapsing solution on a cleaned
microscope slide, and using the end of the pipette tip spread the
liquid over the area to be occupied by the filter sector.  Place the
filter sector, active surface upwards, on top of the solution,
lowering the edge of the filter at an angle of about 20 degrees so
that air bubbles are not created.  Remove any solution not absorbed
by the filter by allowing a paper tissue to contact the liquid at
the edge of the filter.  More than one filter sector may be placed
on one slide, and a laboratory blank filter sector shall also be
prepared on each slide.  Place the slide either on a
thermostatically-controlled slide warmer at a temperature of
approximately 65-70°C, or in an oven at the same temperature, for
about 10 minutes.  The filter collapses slowly to about 15% of its
original thickness.  The procedure leaves a thin, transparent
plastic film adhering to the microscope slide, with particles and
fibers embedded in the upper surface.

10.2.8.  Carbon Coating of Filter Sectors

      Place the glass slide holding the filter sectors on the
rotation-tilting device, approximately 10-12 cm from the evaporation
source, and evacuate the evaporator chamber to a vacuum better than
0.013 Pa.  The evaporation of carbon shall be performed in very
short bursts, separated by some seconds to allow the electrodes to
cool.  If evaporation of carbon is too rapid, the surface of the
filter may be damaged.  The thickness of carbon required is
dependent on the size of particles on the filter, and approximately
30-50 run has been found to be satisfactory.   If the carbon film is
too thin, large particles will break out of the film during the
later stages of preparation, and there will be few complete and
undamaged grid openings.on the specimen.   Too thick a carbon film
will lead to a TEM image which is lacking in contrast,  and the
ability to obtain SAED patterns will be compromised.   The carbon
film thickness, should be the minimum possible, while retaining most
of the grid openings of the TEM specimen intact.
                            45

-------
      10.2.9.   Preparation of the  Jaffe Washer

           Place several pieces of lens  tissue  on the  stainless  steel
      bridge,  and fill the washer  with dimethyl  formamide  to  a level where
      the meniscus contacts the underside of the mesh,  resulting  in
      saturation of the lens tissue.  If  it is  intended to complete the
      washing of the specimens in  the condensation washer,  the pieces of
      lens tissue should be of such a size that  they will  fit on  to the
      stainless steel mesh of the  condensation washer cold finger.  Three
      TEM specimen grids shall be  prepared from  each of the coated
      analytical filters, and for  each of the analytical filters, three
      new specimen grids should be placed on to  one of  the pieces of lens
      tissue.

      10.2.10.  Placing of Specimens Into the Jaffe Washer

           Using a curved scalpel blade, cut along the radius of the
      coated filter sector, approximately 1 mm  from the edge.  Discard  the
      thin strip of coated filter, which  usually exhibits  signs of edge
      effects introduced during the collapsing  procedure.   Starting from
      the freshly-cut edge, cut three approximately 3 mm square pieces  of
      the coated filter.  Select  three  squares  to represent the center  and
      the outer periphery of the  active  surface  of the  filter.  Place each
      square  of filter, carbon side up,  on one  of the TEM  specimen grids
      already on the lens tissue  in the Jaffe washer.  Cover  the  Jaffe
      washer with the lid, and allow the  washer to stand.   Specimens are
      normally cleared after approximately 4 hours.

      10.2.11.  Rapid Preparation of TEM Specimens From MCE Filters

            An alternative washing procedure may be used to prepare TEM
      specimens from MCE filters  more rapidly than can  be  achieved by the
      Jaffe washing procedure.  After  the specimens have been washed in a
      Jaffe washer for approximately 1  hour, transfer the  piece of lens
      tissue supporting the specimens to  the cold finger of a condensation
      washer operating with acetone as  the solvent.  Operate  the
      condensation washer for approximately 30  minutes.  This treatment
      removes all remaining filter polymer.

10.3.  PREPARATION OF TEM SPECIMEN GRIDS  BY THE DIRECT  METHOD
      10.3.1.
Cleaning of Sample Cassettes
            After ensuring that the cassette is tightly sealed,  wipe the
      exterior surfaces of each sampling cassette with a clean,  wet paper
      towel as described in 10.2.1 before it is taken into the clean
      facility or laminar flow hood.
                                  46

-------
       10.3.2.   Selection of Area of Filter  for  Preparation

            Using  clean tweezers,  remove  the MCE  filter  from  the  filter
       cassette,  and cut a sector as described in  10.2.5.

       10.3.3.   Filter  Collapsing Procedure

            Collapse the filter sector as described  in 10.2.7.

       10.3.4.   Plasma  Etching of the Filter Surface

            The  conditions required in a particular  plasma asher  shall
       first be  established using the procedure  defined in Appendix A.
       Place the  microscope slide holding the collapsed filter sectors in
       the plasma asher,  and etch for the time and under  the conditions
       determined.   Care should be  taken to  ensure that the correct
       conditions are used.  After  etching,  admit  air slowly to the chamber
       and remove the microscope slide.

       10.3.5.  Carbon  Coating of Filters

            Carbon coat the microscope slide holding the collapsed filter
       portions as  described in 10.2.8.

       10.3.6.  Preparation of the Jaffe Washer

            Prepare  the Jaffe washer as described in 10.2.9.

       10.3.7.  Placing  of Specimens into the Jaffe Washer

            Place  specimens in the Jaffe washer as described in 10.2.10.
       Specimens  are  normally cleared after approximately 4 hours.

       10.3.8 .  Rapid Preparation of TEM Specimens from MCE Filters

            The  alternative rapid washing procedure,  described in 10.2.11,
      may be used  to prepare TEM specimens from MCE filters more rapidly
       than can be  achieved by the Jaffe washing procedure.

10.4.  CRITERIA FOR ACCEPTABLE TEM SPECIMEN GRIDS

      Examine the TEM specimen grid in the TEM at a magnification
sufficiently low (300-1000) so that complete grid openings  can be
inspected.   Reject the grid if:

       (a)   the TEM specimen has  not been cleared of filter medium by the
            filter dissolution step.   If the TEM specimen exhibits areas
            of undissolved filter medium and,  if at least two  of the three
            specimen grids are not cleared,  either additional  solvent
            washing shall be carried out or  new specimens shall be
            prepared from the filter;
                                  47

-------
      (b)    the sample is over-loaded with particulate.   If the specimen
            grid exhibits more than approximately 10%  obscuration on the
            majority of the grid openings,  the specimen will be designated
            as over-loaded.  This filter cannot be analyzed satisfactorily
            because the grid is too heavily loaded with debris  to allow
            separate examination of individual particles by ED  and EDXA
            and obscuration of structures by other particulate  would lead
            to under-estimation of the structure count;

      (c)    the particulate deposits on the specimen are not uniformly
            distributed from one grid opening to the next.   If  the
            particulate deposits on the specimen are obviously  not uniform
            from one grid opening to the next, the specimen will be
            designated as non-uniform.  This condition is caused either by
            use of an unsatisfactory procedure for the water filtration or
            it is a consequence of the fundamental nature of the airborne
            particulate.  Such a filter cannot be analyzed satisfactorily;

      (d)    the TEM grid is too heavily loaded with asbestos structures to
            make an accurate count.  Accurate counts cannot be  made if the
            grid has more than approximately 7000 asbestos structures/mm2;
            or,

      (e)    less than approximately 75% of the grid openings have unbroken
            carbon film over the whole grid opening.  Since the breakage
            of carbon film is usually more frequent in areas of heavy
            deposit, counting of the intact openings can lead to an
            underestimate of the structure count.  An  additional carbon
            coating may be applied to the carbon coated filter and new
            specimen grids prepared.  The larger particles can often be
            supported by using a thicker carbon film.   If this  action does
            not produce acceptable specimen grids, this filter cannot be
            analyzed and a filter prepared with a lighter loading of
            particulate should be selected.

      If any one or more of the conditions described in (b), (c), (d) or
(e) exists, the specimen grids cannot be analyzed.

10.5.  PROCEDURE FOR STRUCTURE COUNTING BY TEM

      10.5.1.   Introduction

            The TEM examination consists of a count of asbestos
      structures, which are present on a specified number of grid
      openings.  Asbestos structures are classified into groups on the
      basis of morphological observations, ED patterns and EDXA spectra.
      The total number of asbestos structures to be counted depends on the
      statistical precision desired.  In the absence of asbestos
      structures, the area of the TEM specimen grids which must be
      examined depends on the analytical sensitivity required.   The
                                  48

-------
 precision of the asbestos structure count depends  not only on the
 total number of asbestos structures counted,  but also on their
 uniformity from one grid opening to the next.   Additional asbestos
 structure counting will be necessary if greater precision is
 required.

       In order  that the estimate of the asbestos structure density
 on the sampling filter  shall  not be based on  the asbestos structure
 deposits found  within the small  area represented by one  specimen
 grid,  grid openings shall be  examined on at least  two of the  three
 specimen grids  prepared.   The results shall then be combined  in the
 calculation of  the asbestos structure density.   Structure counts
 shall  be made at a magnification of approximately  20000,,  and  will be
 terminated at the number of asbestos structures  as  defined below,
 except that the count shall be continued until  a minimum of 4 grid
 openings have been examined.   Otherwise,  the asbestos structure
 count  shall continue  to that  number of grid openings  at  which the
 specified analytical  sensitivity has been achieved.

 10.5.2.   Measurement  of Mean  Grid Opening Area

       The grid  opening  area shall be measured for the type  of TEM
 specimen grids  in use.  This  may be  performed either  on  an  optical
 microscope  at a calibrated magnification of about 400, or at  a
 calibrated  magnification  in the  TEM.  A mean value  for the  grid
 opening  dimensions may be  used if this  value can be shown to  be
 sufficiently precise.   If  a mean value  is used,  the standard
 deviation of the mean of 10 openings  selected randomly from each  of
 10 grids  shall be  less  than 5%.  Alternately,  the dimensions  of one
 grid opening on each of the grids examined shall be measured  and  the
 mean of  these used in the  calculations  for the particular analysis.


      Values for the number of grid openings to be counted to
 achieve a desired level of sensitivity  (Table 9.1)  will be adjusted
 based on  the average grid opening size derived in this section.  The
 procedure for performing the conversion is presented  in a note at
 the bottom of Table 9.1.

 10.5.3.  TEM Alignment and Calibration Procedures

      Before structure counting is performed,  align the TEM
 according to instrumental specifications.  Calibrate the TEM and
 EDXA system according to the procedures described in Appendix B.

 10.5.4.  Determination of Stopping Point

      Before structure counting is commenced,  the area of specimen
to be examined in order  to achieve the required analytical
                            49

-------
sensitivity shall be calculated.  The maximum number of grid
openings to be examined shall be calculated from the formula:
                  N
Af/(Ag x V x  S x  F)
where:
          N  - Number of grid openings to be examined.  N shall
               be rounded upwards to the next highest integer

          Af - Area of collection filter in mm2
          A. — Area of TEM specimen grid opening in mm

          V  — Volume of air sampled in liters

          S  = Required analytical sensitivity in
               asbestos structures/liter

          F  — Concentration factor.

      The concentration factor F arises from the fact that the
asbestos structures on a particular area of the TEM specimen have
usually been transferred from a different area of the original
sample collection filter.  The concentration factor may correspond
to a concentration or a dilution of the original asbestos structure
density on the collection filter.  It is calculated from the ratio
between the areas of the original and analytical filters, the volume
of water used to re-disperse the ash, and the volume of the final
dispersion used to prepare the analytical filter.  F is calculated
from the following formula:
                    F - A. x Vf/(Af x Vr)
where:
  Aa - Area of filter ashed
  Af - Area of analytical filter
  VE — Volume of water used to re-disperse ashed particulate
  Vf - Volume of dispersion filtered through analytical filter

 10.5.5.   General Procedure for Structure  Counting and  Size Analysis

      Use at  least two  specimen  grids  prepared  from  the filter  in
 the structure count.  Select  at  random several  grid  openings  from
 each grid,  and  combine  the data  in the calculation of  the results.

      Use a form similar to that shown in Figure 10.1  to record the
 structure counting data.  Insert the first specimen  grid into the
 TEM.  Select  a  typical  grid opening and set the screen magnification
 to  the  calibrated  value (approximately 20,000).  Adjust the  sample
                             50

-------
                      TEH ASBESTOS COUNT
                                                           Page   of
Preparation Date:
Analysis Date:
Report Number:
Filename:  	
                         By:_
                         By:_
Sample Description:
Sample Number:
                  Sample  Filter Area:	sq.mm.
                  Area of Filter Taken:	sq.mm.
                  Analytical Filter Area:	sq.mm.
                  Redispersal Volume:	mL
                  Liquid  Volume Filtered:	mL
                  Magnification:	
                  Air Volume:	
                                             Grid Opening Dimension:
                                                               ..Liters
                                                                  urn
 Grid
Opening
Structure
 Number
           Class
Structure
  Type
Length
 (mm)
Width
(mm)
Comments
                 Figure 10.1:     Counting Form
                                  51

-------
height until the features in the center of the TEM viewing screen
are at the eucentric point.  Set the goniometer tilt angle to zero.

      In column 1 of the structure counting form, record the
sequential number of the grid opening.  Position the specimen so
that the grid opening is positioned with one corner visible on the
screen.  Move the image by adjustment of only one translation
control, carefully examining the sample for asbestos structures,
until the opposite side of the grid opening is encountered.  Move
the image by a pre-determined distance less than one screen
diameter, using the other translation control, and scan the image in
the reverse direction.  Continue the procedure in this manner until
the entire grid opening has been inspected in a pattern similar to
that shown in Figure 10.2.

      When a fibrous structure is detected, assign a sequential
number in column 2, perform the identification procedures required
as detailed in Appendix D, and enter the appropriate compositional
classification on the fiber counting form in column 3.  Assign a
morphological classification to the structure according to the
procedures in Appendix C, and record this in column 4.  Measure on
the TEM viewing screen the length and width of the image of each of
the components of the fibrous structure in mm and record these
measurements in columns 5 and 6 of the structure counting form.

      If fibrous structures are present that are of obvious
biological origin or that are determined to be non-asbestos, record
data for a minimum of the first 10 such structures; further
recording of data from non-asbestos structures is optional.  After a
fibrous structure has been examined and measured, re-locate the
original field of view accurately before continuing scanning of the
specimen.  Failure to do this may cause asbestos structures to be
overlooked or counted twice.  The point at which the examination is
to be terminated depends on the specimen, and is defined by either
(a), (b) or (c) below:

(a)   for Phase 1 samples, which have been prepared by the indirect
      TEM specimen preparation method only, complete the examination
      at the end of the grid opening on which the 50th asbestos
      structure is counted;

(b)   for Phase 2 samples, which have been prepared both by the
      direct and indirect TEM specimen preparation methods, complete
      the examination at the end of the grid opening on which the
      100th asbestos structure is counted; or

(c)   for either of the sample types, until the number of grid
      openings required to achieve the specified analytical
      sensitivity for asbestos structures of all lengths, calculated
      according to 10.5.4, have been inspected, whichever occurs
                             52

-------
      TEM field of viei
Grid opening
                                             First pass




                                             Second pass
                                              \
                      XX"
   Figure 10.2:   Scanning Procedure for TEM Specimen Examination
                                53

-------
      first.   The data shall be drawn approximately equally from the
      two grids.   Regardless of the  value calculated according to
      10.5.4,  asbestos structures shall be counted on a minimum of  4
      grid openings.   Table 10.1 presents a summary of the counting
      requirements for Phase 1 and Phase 2 samples.

(d)    Only those structures that are identified as, or are suspected
      to be,  either chrysotile or one of the amphibole minerals,
      will be included in either the original or the extended
      structure count.  Other materials, such as gypsum, cellulose
      fibers,  and filter artifacts such as undissolyed filter
      strands, will not be included in the structure count.  This
      restriction is intended to ensure that the best statistical
      validity is obtained for the materials of interest.

10.5.6.     Measurement of Concentration for Asbestos Structures
            Longer than 5 fj.ro.

      To increase the statistical validity of the measurement, this
method requires that an extended count be made of those asbestos
structures with aspect ratios equal to or greater than 5:1 and
longer than 5 pm.  This separate structure count takes account only
of the asbestos structures longer than 5 /zm, and scanning of the TEM
specimen shall be performed using a magnification of approximately
10,000.

      Measurement of the asbestos structure lengths shall be
performed at a magnification of 10,000, but asbestos structure
diameters shall be measured at a magnification of  a minimum of
20,000.  For Phase 1 samples, which have been prepared by  indirect
transfer only, the count is continued until either 50 asbestos
structures have been recorded or an area of specimen has been
examined that achieves the specified analytical sensitivity for
asbestos structures longer than 5 /wn.   For Phase 2 samples, which
have been prepared both by indirect and direct transfer,  the count
is continued until either 100 asbestos  structures  have been counted
or an  area of specimen has been examined that achieves  the specified
analytical sensitivity for asbestos structures longer than 5 fan.
Table  10.1 presents a summary of the counting requirements for
Phase  1  and Phase 2 samples.

       Only those  structures  that are identified as,  or  are suspected
to be, either chrysotile or  one  of  the  amphibole minerals, will be
reported in either the original  or  the  extended structure count.
Other  materials,  such as gypsum, cellulose  fibers, and  filter
artifacts such  as undissolved  filter strands, will not  be included
in the structure  count.  This  restriction  is  intended to ensure that
the best statistical  validity  is obtained  for  the  materials of
interest.
                             54

-------
                                                   TABLE 10.1:  SPECIFICATIONS  FOR PHASE  1 AND 2  SAMPLING AND ANALYSIS
                                                                                                                         WORKING DRAFT  -   DO NOT COPY Oft QUOTE
         Setting Definitions
Ul
Ul
        Ambient:
        Rural:
        Urban
         and
        Agri-
        cult
        ural:
        Activity
        Specific:
 Structure sizes to be counted
 Number of structures to be counted

 Volume of air to be collected is 15 m3
 Maximum area to be scanned
 Maximum number of grid openings
 to be counted
 Minimum number of grid openings
 to be counted
 Volume of air to be collected is 10 m
 Maximum area to be scanned
 Maximum number of grid openings
 to be counted
 Minimum number of grid openings
 to be counted

 Volume of air to be collected is 5  m3
 Maximum area to be scanned
 Maximum number of grid openings
 to be counted
 Minimum number of grid openings
 to be counted
Volume of air to be determined in the field
Maximum area to be scanned
Maximum number of grid openings
to be counted
Minimum number of grid openings
to be counted
Phase 1 Samoles:
High Magnifi-
cation Scan
total
50
0.05 mm2
7
4
0.08 mm2
10
4
0.15mm2
20
4
TBD
TBD
4
Low Magnifi-
cation Scan
length > 5 im
50
1.3 mn2
160
4
2.0 mm2
240
4
3.9 mm2
480
4
TBD
TBD
4
Phase 2 Samoles:
High Magnifi-
cation Scan
total
100
0.11 mm2
13
4
0.16 rm2
20
4
0.32 mm2
40
4
TBD
TBD
4
Low Magnifi-
cation Scan
length > 5 jun
100
25 mm2
307
4
3.7mm2
460
4
7.5 irm2
920
4
TBD
TBD
4

-------
                                   TABLE 10.1
                                   (continued)
NOTES:

     Phase 1 samples  will be prepared by the  indirect  technique.   Filters from
Phase 2 samples will be split. Half will be prepared by the indirect technique
and half will be prepared by the direct technique.  Note that the required
sensitivity for the Phase 2 samples is higher than that necessary for the Phase
1 samples.  Consequently, it is assumed that 20 structures need to be counted
for phase 2 samples at a minimum.

    The information presented in this table is  derived from the  equation presen-
ted in the Section 10.5.4 assuming that Af - 385 mm2,  Ag - 0.0081 mm2, F - 1,  and
that 10 structures will have to be counted for statistical significance.  Based
on a review of reported background concentrations and typical concentrations
encountered in previous studies, it is also assumed that the critical
concentration for total structures is 5.0 structures/liter and the critical
concentration for structures longer than 5 micrometers is 0.2 structures/liter.

    During any site investigation,  it will likely be necessary to characterize
both ambient air concentrations and concentrations of asbestos associated with
specific human activities.  Although the method is designed for ambient
conditions, it will be possible to adapt it to activity specific sampling by
accounting with some adjustment for expected differences in conditions between
the two types of samples.   The principle difference likely is that a much higher
concentration of non-asbestos particulates will likely be contained  in air
sampled in  association with specific, dust-generating activities.  As a
consequence, the estimated  values for maximum  tolerable air volumes presented in
section 4.2.1 of the text can not be applied to activity specific sampling and
some measure of TSP will have to be provided to derive alternate  sampling
volumes.  Thus, "TBD" means:  to be determined.

    It must be emphasized that the volume of air recommended for collection in
each of the environmental settings listed  represents  a conservative  estimate
(based on published  data) of the maximum that  can be  collected and still allow
analysis  without overloading the filter.   If additional information  regarding
the  local average  concentration of ambient TSP, air volumes to be collected in
any of these settings should be optimized  accordingly.

    To achieve the desired  analytical sensitivity, the numbers of grid openings
required  for counting  (as presented  in  this  Table) may either be  distributed
over a set of specimen  grids prepared from a single  sample  or a  larger  set of
specimen  grids prepared from several samples collected in  a manner assuring that
the samples are appropriately homogeneous  (see Section 4.3  of the Technical
Background Document, Part 2 of  this  report).
                                        56

-------
10.6.  BLANK AND QUALITY CONTROL DETERMINATIONS

   Before  air samples are collected, a minimum  of  2 unused 0.45 jum pore size
MCE filters from each lot of 100 filters shall be sectioned and analyzed
both by the indirect and the direct transfer procedure to determine the mean
background asbestos structure concentration.  If the mean concentration for
all types of asbestos structures, expressed as the concentration per unit
area of the sample collection filter,  is found to be more than
10 structures/mm2,  the  reasons for the high blank values shall be determined
and the situation corrected before air samples are collected.

   To ensure that contamination by extraneous asbestos structures during
specimen preparation is insignificant compared with the results reported on
samples, it is necessary to establish a continuous program of blank measure-
ments .   The number of field blanks incorporated into the program shall be at
least 10% of the total number of samples, and all of these shall be
analyzed.  In addition, one unused 0.1 /im pore size filter shall be included
with every group of samples prepared on one microscope slide.

   It is  further recommended  that laboratory blanks be collected  intermit-
tently at all critical phases of the laboratory program.  For example, in
addition to the filter blanks and field blanks, suspension blanks should be
prepared by subjecting a clean ashing tube to the asher, rinsing the tube
and completing the standard indirect preparation on the "clean" rinse.
Other blanks may be constructed to test for the potential introduction of
contamination at other phases of the process.   Such blanks need not be
analyzed but may be stored to provide a chain for trouble-shooting should a
problem with contamination occur.

   Initially, and also  included at random points in the  sampling program, it
is necessary to ensure that samples of known asbestos structure
concentrations can be analyzed satisfactorily by this procedure.   There are
many opportunities in the procedure for a low recovery to be obtained, and
the recovery must be measured frequently, particularly if more than one
laboratory performs the analysis.  Reference filters of known concentration
will be incorporated,  and shall constitute a minimum of 5% of the total
analyses.  These reference filters shall not be identified to the analytical
laboratory prior to the analysis of any group of samples.   The results of
these reference analyses shall not differ at the 5% significance level from
the mean value obtained by a selected group of experienced laboratories.

   Since  there  is a subjective component  in  the structure counting procedure,
it is necessary that re-counts of some  specimens be made by different
microscopists,  in order to minimize the  subjective effects.   Such recounts
provide a means of maintaining comparability between counts made by
different microscopists.  Variability between and within microscopists and
between laboratories shall be characterized.  These quality assurance
                                     57

-------
measurements shall constitute a minimum of approximately 10% of the
analyses.  Repeat results should not differ at the 5% significance level.
10.7.  CALCULATION OF RESULTS
                                                                      Prior
   Calculate the results using the procedures detailed in Appendix E.
to the TEM examination of the specimens,  the level of analysis was
specified.  Before the results are calculated,  the compositional and
morphological classifications to be included in the result shall be
specified.  The chi-squared uniformity test shall be conducted using the
number of isolated asbestos structures (individual entities) found on each
grid opening,  prior to the application of the cluster and matrix counting
criteria.  The concentration result shall be calculated using the numbers of
asbestos structures reported after the application of the cluster and matrix
counting criteria.
                                    58

-------
11.  PERFORMANCE CHARACTERISTICS
 11.1.  INTERFERENCES AND LIMITATIONS OF STRUCTURE IDENTIFICATION

    Unequivocal identification of every chrysotile structure is not possible,
 due to both instrumental limitations and the nature of some of the component
 fibers.  The requirement for a calibrated ED pattern eliminates the pos-
 sibility of an incorrect identification of the structure selected.  However,
 there is a possibility of misidentification of other chrysotile structures
 for which both a morphology and ED pattern are reported on the basis of
 visual inspection only.  The only significant possibilities of
 misidentification occur with halloysite, vermiculite scrolls or
 palygorskite,  all of which can be discriminated from chrysotile by the use
 of EDXA and by observation of the 0.73 run (002) reflection of chrysotile in
 the ED pattern.

    As  in the case  of chrysotile  structures, complete  identification of every
 amphibole structure is not possible due to instrumental limitations and the
 nature of some of the component fibers.  Moreover, complete identification
 of every amphibole structure is not practical due to the limitations of both
 time and cost.  Particles of a number of other minerals having compositions
 similar to those of some amphiboles could be erroneously classified as
 amphibole when the classification criteria do not include zone-axis ED
 techniques.  However, the requirement for quantitative EDXA measurements on
 all structures as support for the random orientation ED technique makes
 misidentification very unlikely, particularly when other similar structures
 in the same sample have been identified as amphibole by zone-axis methods.
 The possibility of misidentification is further reduced with increasing
 aspect ratio,  since many of the minerals with which amphibole may be
 confused do not display its prominent cleavage parallel to the c-axis.

 11.2.   PRECISION AND ACCURACY

    11.2.1.  Precision

         The analytical precision that can be obtained is dependent upon the
   number of structures counted, and also on the uniformity of the
   particulate deposit on the original filter.  Assuming that the structures
   are randomly deposited on the filter, if 100 structures are counted and
   the loading is at least 3.5 fibers/grid opening, computer modeling of the
   counting procedure shows that a coefficient of variation of about 10% can
   be expected.  As the number of structures counted decreases, the precision
   will also decrease approximately as ,/N, where N is the number of
   structures counted.  In practice, particulate deposits obtained by
   filtration are rarely ideally distributed,  and it is found that the
   precision is correspondingly reduced from that predicted by the Poisson
   distribution.  This degradation is a consequence of:
                                     59

-------
         (a)    non-uniformity of the filtered particulate  deposit;

         (b)    distortion of the structure distribution by application of  the
               structure counting criteria;

         (c)    variation between microscopists in their interpretation of  the
               fibrous structures;  and,

         (d)    variation between microscopists and instruments  in their
               ability to detect and identify structures.

         The  95% confidence interval about the mean for a  single  structure
   concentration measurement using this  analytical method  should  be about
   ±25% when  about 100 structures are counted.

   11.2.2.  Accuracy

         There is no independent method  available to determine  accuracy.

11.3.  ANALYTICAL SENSITIVITY

   Analytical sensitivity is the concentration corresponding  to observation
of one asbestos  structure  in an analysis.  The analytical sensitivity of the
method can be varied by choice of the area of the collection filter,  the
volume of air sampled, the dilution or concentration factor used in the
specimen preparation and the area of the specimen examined in the TEM.  The
analytical sensitivity shall be quoted for each sample analysis.
  NOTE
         In practice,  the lowest achievable value  of  analytical sensitivity
         is frequently determined by  the  total  suspended particulate
         concentration in the  air sample,  since each  particle on the TEM
         specimen must be separated from  adjacent  ones by a sufficient
         distance that the particle can be identified without interference.
         Particulate  loadings  on filters  greater than about 25 jig/cm2 usually
         preclude preparation  of TEM  specimens.

         If the  analysis  is to be performed with an acceptable expenditure of
         time, the area of the specimen examined in the TEM must also be
         limited.  The analytical sensitivity that can be achieved for any
         particular sample depends  on the  air volume, the area of TEM
         specimen examined,  and  the dilution or concentration factor used in
         the analysis,  which in  turn  is controlled by the amount of
         particulate  that is not removed by the selective concentration
         procedures.
                                     60

-------
11.4,  LIMIT OF DETECTION

   Should asbestos be  observed during  analysis,  it  is  generally important  to
distinguish,whether such asbestos originated in the sampled medium or if it
was introduced as contamination during analysis.  Asbestos that can be
attributed to the sampled medium is generally considered to have been
"detected".  Thus, a detection limit is defined as the smallest measurement
that is unlikely (probability less than a specified value) to be due
entirely to contamination from sources other than the air being sampled.
Detection limits are generally quantified by considering the magnitude and
frequency of occurrence of the analytical background associated with a
particular method.  However, detection limits for asbestos methods are
difficult to quantify because the distribution of analytical background
associated with asbestos analysis tends to be poorly behaved.

   To avoid the problems typically associated with  defining a detection  limit
for asbestos analysis, an alternate procedure for distinguishing sampled
asbestos from contamination due to analytical background has been
incorporated into this method3.   A statistical  test is performed to  compare
structure concentrations observed on sample filters with structure
concentrations observed on appropriate blanks to determine whether
differences between the concentrations observed on the two sets of filters
are statistically significant.

   NOTE

   Due to  the difficulty in characterizing contributions to observed asbestos
   on individual samples from  contamination introduced as analytical
   background, tests to distinguish "detected" asbestos from  such
   contamination are better applied to groups of samples.  Adjustments to
   account  for analytical background are not recommended for  individual
   samples.
        A detection limit is more appropriate for analytical methods that
        employ wet-chemical or spectroscopic techniques rather than
        structure counts (as in the case of asbestos) because the former
        are invariably associated with a lower limit to the amount (number
        of molecules) that can be detected.  For asbestos, however,
        analysis by TEM is sufficiently sensitive to detect the smallest,
        single asbestos structure.  Thus the issues involved in setting
        detection limits for TEM analyses of asbestos structures are quite
        different than from those involved with detecting populations of
        molecules such as in gas chromatograph-mass spectrometry (GC-MS)
        analyses, for example.

                                    61

-------
12.  REPORTING REQUIREMENTS

 Reporting requirements  for  the  laboratory include the individual documen-
tation for each sample (sample analysis report) and the cumulative report for
any batch of samples from a particular study.  The structure counting form to
be used by the analyst was presented in Figure 10.1.   Required data reduction
by the site data reviewer is presented at the end of this chapter, Section
12.3.  It is critical that the following documentation requirements be
followed exactly to preserve the ability to interpret results without the need
to re-examine samples.

 12.1.  SAMPLE ANALYSIS  REPORT

    The sample analysis report for each sample shall  include at least the
 following information:

    (a)   reference to this analytical method;

    (b)   identification of the air sample and the batch number of the
          collection filter used;

    (c)   the dates and times of the air sample collection period;

    (d)   the volume of air passed through the sample collection filter, the
          area of filter used for the sample preparation, the volume of water
          used for re-dispersal, the batch number of the filters used for
          filtration of the aqueous suspension, and,  for the set of TEM
          specimen grids examined, the volume of the aliquot used in the
          aqueous suspension filtered;

    (e)   a complete listing of the structure counting data.  The dimensions
          of the TEM grid openings shall be specified, and for each asbestos
          structure (or structure component) the following data shall be
          included: grid opening number, structure number, identification
          category (morphological classification), structure type /composi-
          tional classification), length and width of the structure in /^m, and
          any comments concerning the structure;

    (f)   a statement of the minimum acceptable identification category and
          the maximum identification category attempted;

    (g)   a statement specifying which identification and structure categories
          have been used to calculate the concentration values;
                                       62

-------
    (h)


    (i)


    (j)


    (k)

    (1)


    (m)
 separate concentration values for chrysotile and amphibole struc-
 tures, expressed in asbestos structures/liter4;

 separate concentration values for chrysotile and amphibole structure
 components, expressed in asbestos structures/liter*;

 the 95% confidence limits for the concentration values, expressed in
 asbestos structures/liter;

 the analytical sensitivity,  expressed in asbestos structures/liter;

 compositional data for the principal varieties of amphibole  if
 present;                                                    '

 items d - 1 for the separate count of asbestos structures longer
 than 5 fim;
 _  An example of a suitable format for the structure counting data is shown
 •*•**• £ IgUjTG  J_^ . J_ *
   NOTE

   As improved definitions of the dimensions of asbestos structures critical
   to the determination of risk become available,  the reporting requirements
   for this method may be modified accordingly.   At the same time, sufficient
   information is preserved in these reporting requirements to allow later
   re-evaluation of existing analysis results without the need to re-examine
   the original sample.

12.2.  SAMPLE BATCH REPORT

   In addition to the  sample analysis report for each sample,  a summary page
should be provided for each batch of samples representing an entire project
The summary sheet should include the following information-
   (a)

   (b)

   (c)
project title;

date samples received and data results reported;

a summary listing of sample results with chrysotile and amphibole
reported separately including:

      the sample number;
        If asbestos is not detected during a particular analysis  the
        resulting mean concentration for that sample should be reported as
        "NF" for "not found".
                                    63

-------
              the analytical sensitivity for each size range characterized
              (structures of all sizes and structures longer than 5 mm)

              the total number of structures counted;

              the concentration of asbestos structures of all sizes in
              structures per liter with appropriate 95% confidence limits ;

              the concentration of asbestos structure components in struc-
              tures per liter with appropriate 95% confidence limits;

              the concentration of asbestos structures longer than 5 urn in
              structures per liter with appropriate 95% confidence limits;

              the concentration of asbestos structure components longer than
              5 (im in structures per liter with 95% confidence limits;

  A sample summary sheet is presented  in figure 12.2.

12.3.   DATA REVIEW REPORT

  Once  the Sample Batch Report is received from the laboratory, the
following reporting procedure will be performed by the project data
reviewer.  The data reviewer shall provide a summary sheet reflecting
specific steps in data reduction.   The summary data sheet shall include the
following information:
(a)

(b)


(c)


(d)

(a)


(f)


(g)


(h)


(i)
         the  sample  number;

         the  type  of sample  (lab blank,  field blank,  sampling  station
         identifier) ;

         the  analytical sensitivity for  each size  fraction reported
         (structures of all  sizes  and structures longer than 5 ^m) ;

         the  number  of total structures  counted;

         the  concentration of asbestos structures  of all sizes in structures
         per  liter with 95%  confidence limits;

         the  concentration of asbestos structure components in structures  per
         liter with  95% confidence limits;
the concentration of asbestos structures longer than 5
structures per liter with 95% confidence limits;
                                                                   in
         the concentration of asbestos structure components longer than 5  /«n
         in structures per liter with 95% confidence limits;

         the ratio of free asbestos fibers to total asbestos  structures;

                                     64

-------
                       SAMPLE ANALYSIS INFORMATION (Page 1)
 Laboratory name
Report number
Date
SAMPLE:  456 Queen Street
         San Diego
         Downwind sample 1988-09-25
         Collected:  8 a.m. to 4 p.m.
              ANALYSIS:   Name of method
                          Reference to method
Air volume:
Area of collection filter:
Area of filter ashed:
Volume of water used for re-dispersal:
Volume of suspension filtered:
Area of analytical filter:
Magnification used for fiber counting:

Mean dimension of grid openings:
                  12500.0 liters
                    385.0 mm2
                     96.3 mm2
                     40.0 mL
                     35.0 mL
                    199.0 mm2
                  20500 (high magnification)
                  10200  (low magnification)
                     90.0 /im
High Magnification Scan
Number of grid openings examined:
Analytical Sensitivity:
Number of chrysotile structures counted:
Number of chrysotile structure components counted:
Number of amphibole structures counted:
Number of amphibole structure components counted:
                     38
                      0.5 s/L
                  63 (including F, M, C and B)
                     91
                     12 (F only)
                     12
Low Magnification Scan

Number of grid openings examined:
Analytical sensitivity:
Number of long chrysotile structures counted:
Number of long chrysotile structure components counted:
Number of long amphibole structures counted:
Number of .long amphibole structure components counted:
                       357
                         0.02 s/L
                         3 (B and C)
                        14
                         0
      NOTES:    Only structures with components  longer  than Spin counted during
               the low magnification scan.
          Figure 12.1A:  Format for Reporting of Counting Data,  Page  1
                                        65

-------
                      SAMPLE ANALYSIS  INFORMATION (Page 2)
 Laboratory name
                Report number
                 Date
 SAMPLE:   456 Queen  Street
          San Diego
          Downwind sample 1988-09-25
          Collected:   8 a.m.  to 4 p.m.
                               ANALYSIS:
  Name of method
  Reference to method
Concentration of
Concentration of
Concentration of
Concentration of
Concentration of
Concentration of
Concentration of
Concentration of
chrysotile structures:
chrysotile structure components:
long chrysotile structures:
long chrysotile structure components;
amphibole structures:
amphibole structure components:
long amphibole structures:
long amphibole structure components:
Mean
(s/L)

 31.5
 45.5
  0.06
  0.28
  6
  6
  NF
  NF
                                                                         95%
                                                                     Confidence
                                                                      Limits

                                                                       Range
                                                                       (s/L)
  24
  37
 0 -
0.16
 3.1
 3.1
 0 -
 0 -
- 38
- 54
 0.12
 - 0.4
- 8.9
- 8.9
 0.07
 0.07
NOTES:   NF means Not Found

         Long structures and structure components are those greater than 5/*m in
         length.
            Figure 12.IB:   Format for Reporting of Counting Data, Page 2
                                      66

-------
               TEM ASBESTOS COUMT - RAW DATA (Page 3 and higher)
Grid "Structure Classif- Structure Length
Opening Number ication* Type pm

1
1 '
1
1
2
3
' 3
3
3
3
3
> 4
5
5
6
7 ,
8
8
y
1
2
:/ 3-
4
5 "
6
" 7
8
9
10
11
12
13
14
is
•v \ %
16
17
,
'CD
CD
CDQ
'AD
CDQ
, CMQ
CMQ
CMQ
CD
ADQ
ADQ
CD
CD
CD
AQZZ
v , •.
CMQ
CD

F
MS002
MF002
MMQ02
B
CS006
CF006
CBQ06
DF006
F
F
M
"MS013
, MB013
v ,f ;
'"No Fibers
F
F'

1,7
5.5
2.6
2,0
8.7
8.5
5.41
8.1
3.6
9.0
4.5
3,0
6.1
4.1
3.2

1.3
1.1
Width Comments
'•••
0.045
4.3
0.045
0.53
0.21
4.3 :
i.'o
1.2
0.75
0.045 Crocidolite
0.60 Amosite
2.0
1.0
0.50
0.10 Amosite

0.030
0.045
* Classification codes listed in Tables Dl and D2
Figure 12.1C:   Format for Reporting Counting Data,  Page 3 and Subsequent Pages
                                   67

-------
  Batch tf:
  Project:
  Lab:
  Date:
Sample Medium:
Analysis Type:
Preparation Technique:
Sample I.D.
Date Received:


Structures counted:
Structure concentration (s/L):
Mean
Range

Low
High
Analytical sensitivity:
Components counted:
Component concentration (s/L):
Mean
Range

Low
High
Analytical sensitivity:
Long structures
counted:
Long structure concentration (s/L):
Mean
Range

Low
High
Analytical sensitivity:
Long components counted:
Long component concentration (s/L):
Mean
Range

Low
High
Analytical sensitivity:







































































"~"~ ~""














••••-''• 	 "~-



„,,,,*„,,,,.,,









I
;: ' ' i ' ' . :
! f


•:
$ '$
\ 1
r^~rr



—









:
i '
;••
i 	 • -



—












	
,.,V.M..,V.'.V.'.'.V.V
	


	





NOTES:
   NF = not found
   Long structures or components are those longer than 5 fan.
   The range presented represents the 95% confidence limits constructed as discussed in Section E.4.
   Provide separate sheets for chrysotile and amphiboles or total asbestos, as necessary.

                    Figure 12.2:  Format for the Summary Batch Report
                                              68

-------
         the ratio of free asbestos fibers to total asbestos structures and
         components;


         the ratio of matrices to total asbestos structures;

         the ratio of bundles to total asbestos structures;

         the ratio of clusters to total asbestos structures;

         items  i-m for asbestos structures longer than 5 urn;


         a complete listing of the structure counting data for each sample
         (item  "e" of section 12.1).


An example of the summary form for data review  is provided  in Figure 12.3.
(j)


(k)


(1)

(m)

(n)

(o)
                                    69

-------
           Case No.:
           Site:
           Lab:
           Reviewer:
           Date:
                                                           ANALYTICAL RESULTS
Concentration of Asbestos (Chrysolite unless otherwise noted)
                                                        Analysis Type: Air for TEM asbestos
                                                                     Indirect Preparation
Sample ID. (number and type)
Analysis/Preparation
Parameter
Number of liters sampled
HIGH MAGNIFICATION
Number of grids examined
Structures/area viewed
Fibers/area viewed
Bundles/area viewed
Clusters/area viewed
Matrices/area viewed
Structures/L
Components/L
Long structures/L
Long components/L
Analytical sensitivity
LOW MAGNIFICATION
Number of grids examined
Structures/area viewed
Fibers/area viewed
Bundles/area viewed
Clusters/area viewed
Matrices/area viewed
Long structures/L
Long components/L
Analytical sensitivity

Mean























95%
Low






















,, .
CL
High
























Mean























95%
Low























CL
High















1








Mean























95%CL
Low 1 High
!














































Mean























95%
Low























CL
High
























Mean





















95%
Low





















|
1
CL
High























-J
o
           NOTES:
            NF = not found
            95% CL = 95% confidence limit
            The range presented represents the 95% upper or lower confidence limit constructed as discussed in Section E.4.
                                                  Figure 12.3: Format for Data Review Summary Report

-------
APPENDIX A - DETERMINATION OF OPERATING CONDITIONS FOR PLASMA ASHER

 A.I  INTRODUCTION

    It is important  to establish standard conditions  for operation of the
 plasma  asher, so that:

    (a)    ashing  conditions  for MCE  filters  and particulate  are  sufficient  and
          remain  constant; and,

    (b)    so  that a  known amount of  etching  of  MCE filters can be  achieved  on
          the surface  of  filters being prepared by the  direct-transfer method.


    Therefore,  the plasma asher  shall be  calibrated so  that  a  known material
 is completely ashed in a constant time.   This is carried out by adjusting
 the radio-frequency power output and the oxygen flow rate,  and measuring the
 time taken to completely oxidize an unused 25 mm diameter 0.45 um pore size
 MCE filter.

 A. 2  PROCEDURE

    Place an unused, 25 mm diameter, 0.45 pm pore  size MCE filter  in  the
 center of a glass microscope slide.  Position the slide approximately in the
 center of the asher chamber.   Close the  chamber and  evacuate to a pressure
 of approximately 40 Pa,  while admitting  oxygen to the  chamber at a rate of
 8-20 cc/min.   Adjust the tuning of the system so that  the  intensity of the
 plasma is maximized.   Measure the time required for  complete oxidation of
 the filter.   Determine the  power setting that  result in complete oxidation
 of the filter in a  period of approximately  15  minutes.


   NOTE

   Plasma oxidation at high radio-frequency powers will cause the filter to
   shrink and curl,  followed by sudden violent ignition.  At lower powers the
   filter will remain in position, and will become slowly thinner until it
   becomes nearly transparent.  It is recommended that  the  radio- frequency
   power be used is  selected such that violent ignition does not occur.   If
   powers are used such that the particulate ignites  violently,  fibers will
   possibly be disturbed and lost from the analysis.
                                     71

-------
APPENDIX B - CALIBRATION PROCEDURES

 B.I  CALIBRATION OF THE TEM

    B.I.I  Calibration of TEM Screen Magnification

          The electron microscope should be aligned according to  the
    specifications of the manufacturer.   Initially,  and at regular intervals,
    calibrate the magnifications used for the analysis using a diffraction
    grating replica.   Adjust the specimen height to the eucentric position
    before carrying out the calibration.  Measure the distance on the
    fluorescent viewing screen occupied by a convenient number of repeat
    distances of the grating image, and calculate the magnification.   Always
    repeat the calibration after any instrumental maintenance or  change  of
    operating conditions.   The magnification of the image on the  viewing
    screen is not the same as that obtained on photographic plates or  film.
    The ratio between these is a constant value for the particular model of
    TEM.

    B.I.2  Calibration of ED Camera Length

          Calibrate the camera length of the TEM when used in ED  mode.  Use a
    specimen grid supporting a carbon film on which a thin film of gold  has
    been evaporated or sputtered.  Form an image of the gold film and  select
    ED conditions.  Adjust the objective lens current to optimize the  pattern
    obtained, and measure on the fluorescent viewing screen the diameters of
    the innermost two rings.  Calculate the camera constant, A x  L, from the
    relationship:
             a x D
  A x L  -
                + k2
 where:
  A x L
      a
      D
the wavelength of the incident electrons
the effective camera length in mm
the unit cell dimensions of gold in Angstr6m units
the diameter of the (hkl) diffraction ring in mm
    Using gold as the calibration material, the camera constant is given by;
  A x L  -  2.3548 x D (smallest ring)
  A x L  -  2.0393 x D (second ring)
                                      72

-------
 B.2   CALIBRATION  OF  THE EDXA SYSTEM

   Energy calibration of the EDXA system for a low energy and high energy
 peak  shall be performed regularly.  Calibration of the intensity  scale of
 the EDXA  system permits quantitative composition data, at an accuracy of
 about 10% of the  elemental  concentration, to be obtained from EDXA spectra
 of silicate minerals  involving the elements Na, Mg_, Al, Si, K, Ca, Mn, and
 Fe.   SRM  2063 should  also be used to calibrate the EDXA system for Mg[ Ca,
 and Fe relative to Si.  If  quantitative determinations are required for
 minerals  containing other elements, reference standards other than those
 referred  to below will be required.  Well-characterized mineral standards
 permit calibration of any TEM-EDXA combination that meets the instrumental
 specifications of 6.3.1 and 6.3.2, so that EDXA data from different
 instruments can be compared.  Reference minerals are required for the
 calibration; the  criteria for selection being that they should be silicate
 minerals with matrices as close as possible to those of the amphiboles or
 serpentine, and that  individual small fragments of the minerals are
homogeneous in composition within a few percent.

  Determine the compositions of  these standards by electron microprobe
analysis or chemical methods,  and prepare TEM grids from aqueous dispersions
of crushed fragments of the same selected mineral specimens.   These TEM
specimens can then be used to calibrate any TEM-EDXA system so that
comparable compositional results can be obtained from different instruments.


  NOTE

  The microprobe  analyses of the mineral standards are made by conventional
  techniques, which can be  found in the bibliography.  The mineral is first
  embedded in a mount of polymethyl methacrylate or epoxy resin.  The mount
  is  then ground  and polished to achieve a flat,  polished surface of the
  mineral fragment.  This surface is then analyzed, using suitable reference
  standards, preferably oxide standards of the individual elements wherever
  these are available.  It  is necessary to take account of the water
  concentration in the minerals, which in the case of chrysotile amounts to
  13% by weight.
  Express the results of the electron microprobe analyses as atomic ratios
  relative to silicon.  X-ray peak ratios of the same elements relative to
  silicon, obtained from the EDXA system, can then be used to calculate the
  relationship between peak area ratio and atomic ratio.   The technique was
  described by Cliff and Lorimer (See Bibliography).

  The X-rays generated in a thin specimen by an incident  electron beam have
  a low probability of interacting with the specimen.   Thus mass absorption
  and fluorescence effects are negligible.   In a specimen composed of
  elements i and j,  the following relationship can be  used to perform
  quantitative analyses in the TEM.

                                    73

-------
where :
        A£  -  the elemental integrated peak areas for element i
        Aj  -  the elemental integrated peak area for element j
        C±  —  the concentration or atomic proportion of element i
        Cj  -  the concentration or atomic proportion of element j
          k  —  a constant
   To incorporate  correction for  the  particle  size  effect  on peak area  ratios
(See Small et al. in Bibliography), extend the Cliff and Lorimer technique
by obtaining separate values of the constant k for different ranges of fiber
diameter.  It is recommended that about 20 EDXA measurements be made for
each range of fiber diameter.  Suitable ranges of fiber diameter are
<0.25 [an, 0.25-0.5 /jm, 0.5-1.0 /«n and >1.0 >m.
   Insert the TEM grid into the TEM,  obtain an image at the calibrated higher
magnification of about 20,000, and adjust the specimen height to the
eucentric point.  If the x-ray detector is a side-entry variety, tilt the
specimen towards the x-ray detector.  Select an isolated fiber or particle
less than 0.5 pm in width, and accumulate an EDXA spectrum using an electron
probe of suitable diameter.  When a well-defined spectrum has been obtained,
perform a background subtraction and calculate the background- corrected peak
areas for each element listed, using energy windows centered on the peaks.
Calculate the ratio of the peak area for each specified element relative to
the peak area for silicon.  All background- subtracted peak areas used for
calibration  shall exceed 400 counts.

         Repeat this  procedure for 20 particles of each mineral  standard.
   Reject analyses of any obviously foreign particles.   Calculate  the
   arithmetic mean peak area to concentration ratio,  k,  for each specified
   element of each mineral standard and for each of the fiber diameter
   ranges.   Routine checks to ensure that there has been no degradation of
   the detector performance shall be carried out.   These k-values  are  used to
   calculate the elemental concentrations of unknown fibers,  using the Cliff
   and Lorimer relationship.
                                     74

-------
APPENDIX C - STRUCTURE COUNTING CRITERIA

 G.1  INTRODUCTION

    In addition to isolated fibers,  other assemblages  of particles  and fibers
 frequently occur in air samples.  Groupings of asbestos fibers and
 particles, referred to as "asbestos structures", are defined as fiber
 bundles, clusters and matrices.  The numerical result of a structure count
 depends strongly on whether the analyst assigns each such assemblage of
 fibers as a single entity, or as the estimated number of individual fibers
 which form the assemblage.  It is therefore important that a logical system
 of counting criteria be defined, so that the interpretation of specific
 structures is the same for all analysts, and so that the numerical result is
 meaningful.  Imposition of specific structure counting, criteria generally
 requires^that some interpretation,  partially based on uncertain health
 effects information,  be made of each asbestos structure found.  It is not
 the intention of this method to make interpretations based on health
 effects.  Rather, it is intended that a clear separation shall be made
 between recording of structure counting data and later interpretation of
 those data.  The system of coding specified in this method permits a clear
 morphological description of the structures to be recorded in a concise
 manner suitable for later interpretation, if necessary, by a range of
 different criteria,  without the necessity for re-examination of the
 specimens.   The system requires each asbestos structure to be assigned a
 predominant classification as fiber, bundle,  cluster or matrix and in some
 cases individual components of a structure are separately enumerated.
 Examples of the various types of morphological structures and the  manner in
 which these shall be  recorded are shown in Figure Cl..


 C.2.  STRUCTURE DEFINITIONS AND TREATMENT

   C.2.1  Fiber

         Any particle with parallel or stepped sides, with an aspect ratio of
   5:1 or greater, shall be defined as a fiber.  For chrysotile asbestos, the
   single fibril shall be defined as a fiber.   A fiber with stepped sides
   shall be assigned a width equal to the average of the minimum and maximum
   widths.  This average shall be used as the width in determination of the
   aspect ratio.                    '

   C.2.2  Bundle

         A grouping composed of apparently attached parallel fibers shall be
   counted as a bundle of a width equal to an estimate of the mean bundle
   width, and a length equal to the  maximum length of the structure.   The
   overall aspect ratio of the bundle may be any value,  provided that the
   individual constituent fibers have aspect ratios equal to or greater than
   J • -L •
                                     75

-------
C.2.3  Cluster

      An aggregate of randomly oriented fibers, with or without bundles,
shall be defined as a cluster.  Clusters occur as two varieties:

Cluster type A;  a disperse and open network, in which both ends of
individual fibers and bundles can be separately identified and their
dimensions measured.  If the cluster consists of up to 5 such fibers or
bundles, the individual fibers and bundles comprising the cluster shall be
separately counted, measured, and recorded as individual fibers and
bundles that are components of the overall cluster.  If the cluster
consists of more than 5 fibers or bundles it shall be noted as such on the
count sheet.  In addition to characterizing and measuring individual
components, the overall outer dimensions of the cluster will be recorded
as prescribed for type B clusters.

Cluster type B:  a complex and tightly bound network, in which one or both
ends of each individual fiber or bundle are obscured, such that the
dimensions of individual fibers cannot be measured.  In this case the
cluster shall be recorded as a single cluster with no components, and the
overall dimensions in the two perpendicular directions defining the
maximum and minimum dimensions shall be recorded.

      The recording of clusters shall be based on the predominant charac-
teristics of the structures.  For example, a cluster consisting
predominantly of 4 fibers, but with smaller regions of attached material
containing fibers, shall be considered as a type A cluster of 4 fibers to
be recorded separately as fibers that are components of the overall
cluster.

      The procedure for treatment of clusters is illustrated by examples
in Figure C2.

C.2.4  Matrix

      A fiber, fibers, or bundles, may be attached to, or partially
concealed by, a single particle or group of overlapping non-fibrous
particles.  This structure shall be defined as a matrix.  The TEM image
does not discriminate between particles that are attached to fibers, and
those that, by chance, overlap in the TEM image.  It is not known,
therefore, whether such a structure is actually a complex particle, or
whether it has arisen by a simple overlapping of particles and fibers on
the filter.

      Since a matrix structure may involve more than one fiber, it is
important to define in detail how matrices shall be counted.  Matrices
exhibit different characteristics, and three types can be defined:

Matrix type A:  a disperse grouping of overlapping fibers and/or bundles
and associated equant particles or groups of particles in which some

                                  76

-------
 fibers have less than one third of their lengths obscured.   If the matrix
 consists of up to 5 such fibers or bundles,  the fibers or bundles shall be
 recorded as separate fibers or bundles that are components  of the overall
 matrix, and their lengths and widths shall also be recorded.   If the
 matrix consists of more than 5 fibers or bundles,  it shall  denoted as such
 on the count sheet.  In addition to characterizing and measuring
 individual components,  the overall outer dimensions of the  matrix will be
 recorded as prescribed for type B matrices.

 Matrix type B:  a structure consisting of an equant particle  or linked
 group of particles, in which the ends of fibers or bundles  project from
 the particles, but the other ends of the fibers or bundles  are obscured.
 Fibers and bundles shall be treated differently,  depending  on whether the
 obscured length could not possibly be more than one third of  the total
 length.  If the matrix exhibits up to 5 fibers or  bundles,  those fibers
 for which the obscured length could not be more than one  third of the
 total length shall be counted as separate fibers or bundles that are
 components of the overall matrix.   The assigned length for  each partially-
 obscured fiber or bundle shall be equal to the visible length plus the
 maximum possible contribution from the obscured portion.  Fibers or
 bundles which appear to cross the matrix,  and for  which both  ends  can be
 located approximately,  shall be included in  the maximum of  5  and counted
 as  separate fibers or bundles that are components  of the  overall matrix.
 If  more than 5 such fibers or bundles can be individually identified,  this
 will be denoted on the  count sheet.   The residual  matrix, if  it exhibits
 additional fiber terminations that cannot be separately counted because
 the unobscured lengths  are too short,  shall  be recorded as  one matrix.
 All other matrices of type B shall each be recorded as  one  matrix  with no
 components.   The overall dimensions  of each  matrix in the two
 perpendicular directions defining  the maximum and  minimum dimensions  shall
 be  recorded.

 Matrix type  C:   a structure  in which fibers  can be  seen and identified in
 the interior,  but which incorporates  no  fibers  which project from  the
 outside edges.   This  type  of matrix  can  originate  as  a  result  of partial
 dissolution  of binders  during specimen preparation,  when the original
 particle was  a composite material  containing asbestos.  This type  of
 matrix shall be  recorded as  a single  matrix  with no  components.  The
 overall dimensions  of the matrix in  the  two  perpendicular directions
 defining  the maximum and minimum dimensions  shall be  recorded.

 In  practice, structures can  occur  in which different  areas  exhibit
 features of the  three types  of matrix, and more than  one variety of
 asbestos may be  incorporated  in the same structure.   In this case, the
 predominant characteristic of the structure should be determined,'and  then
 a logical procedure should be followed, in which predominant fibers and
bundles are enumerated first up to a maximum of 5 followed by assignment
 of  the remaining asbestos structures according to the counting rules.
Complex structures shall be assigned no more than 6 separate components.
Examples of the procedure which shall be followed are shown in Figure C3.

                                   77

-------
0.3  OTHER STRUCTURE COUNTING CRITERIA

   C.3.1  Structures  that  Intersect Grid Bars

        Structures that intersect grid bar shall be counted only for two
   adjacent  sides  of  the grid opening, as illustrated in Figure C4.  The
   length of the structure shall be recorded as twice the unobscured length.
   Structures intersecting either of  the other two sides shall not be
   included  in the count.   This procedure ensures that the numerical count
   will be accurate,  and that the best average estimate of length has been
   made.

   C.3.2  Structures  That  Extend Outside the Field of View

        During scanning of a grid opening, systematically count structures
   that extend outside  of  the field of view, so as to avoid double-counting.
   In general,  a rule should be established so that structures extending
   outside the field  of view in only  two quadrants are counted.  Structures
   without terminations within the field of view shall not be counted.  The
   procedure is illustrated by Figure C5.  Measure the length of each such
   structure by moving  the specimen to locate the other end of the structure,
   and then  return to the  original field of view before continuing to scan
   the specimen.

C.4  PROCEDURE FOR DATA RECORDING

   When a fibrous  structure is detected during the structure count, the com-
ponents of the structure are first identified according to the procedures in
Appendix D.   The procedure for recording the morphological description
starts by classifying  the predominant characteristic of the structure as  a
fiber, bundle, cluster, or matrix.   The code F,  B,  C or M shall be the first
component of the morphological descriptor.   Depending on the nature of the
individual structure,  additional coded descriptions will be added to
classify separate components of the structure.

        C.4.1 Fibers

               On  the counting form,  isolated fibers as defined in C.2.1
        shall be  recorded by the simple designation "F".  If the fiber is a
        separately-counted part of a cluster or matrix structure, the
        sequential structure number  shall be attached as a suffix, to
        identify  all such constituent components of the particular cluster
        or  matrix structure.  For example, CF004 shall be used to denote a
        fiber forming  part of structure number 004, the overall and predomi-
        nant characteristic of which is a cluster.

        C.4.2 Bundles

               On  the counting form,  isolated bundles as defined in C.2.1
        shall be  recorded by the designation "B".  If the bundle is a

                                     78

-------
 separately-counted part of a cluster or matrix structure,  the
 sequential structure number shall be attached as a suffix,  to
 identify all such constituent components of the cluster or matrix
 structure.  For example,  MB004 shall be used to denote a bundle
 forming part of structure number 004,  the overall and predominant
 characteristic of which is a matrix.

 C.4.3  Clusters

       On the fiber counting form,  isolated clusters,  as defined in
 C.I.4,  shall be recorded by the designation "C".   The dimensions of
 the overall structure are recorded.   If some components of the
 cluster have been separately counted,  the cluster shall be  recorded
 by the designation CS,  followed by the sequential structure number
 "n".   For example,  the code CS004 indicates that the  primary
 structure number 004 was  a cluster,  for which the overall  dimensions
 are specified.   Thus if a localized cluster is attached to  a group
 of fibers or bundles,  and the fibers and bundles  have been  counted
 separately because  this procedure  was  defined by the  predominant
 characteristics of  the structure,  the  sequential  structure  number
 shall be attached as a suffix,  to  identify all constituent
 components of the structure.   For  example,  CF004  shall be used  to
 denote  a fiber  forming part of structure number 004,  the overall and
 predominant characteristic of which  is  a cluster.  When more  than
 5  components  can be  individually identified,  the  cluster is  still
 recorded by the designation CS  and the  overall dimensions of  the
 structure are recorded, but the sequential  structure  number  is
 excluded from the code  and components  are not recorded separately.

 C.4.4  Matrices

      On the  fiber counting form,  isolated  matrices of type B or  C,
 as  defined in C.I.5, shall  be recorded by the  designation "M".   The
 dimensions  of the overall  structure  are  recorded.  If the type B  or
 C matrix is attached to a  group  of fibers or bundles,  and the fibers
 and bundles have been counted separately  because  this  procedure was
 defined by  the predominant  characteristics  of  the structure, the
 sequential  structure number shall be attached  as a suffix, to
 identify all  constituent components of the  structure.  For example,
 the primary matrix structure shall be designated as MS004, and the'
 overall  structure dimensions recorded.  All separately-counted
 components of the structure shall carry the suffix 004.  Thus the
 code MB004 shall be used to denote a bundle forming part of
 structure number 004, the overall and predominant characteristic of
which is a matrix.  When more than 5 components can be individually
 identified, the matrix is still recorded by the designation MS and
 the overall dimensions of the structure are recorded,  but the
 sequential counting number is excluded from the code and components
are not recorded separately.
                            79

-------
                                                                FIBERS
          \
                                                                BUNDLES
      CLUSTER TYPE A
                     CLUSTER TYPE B
MATRIX TYPE A
MATRIX TYPE B
MATRIX TYPE C
      Figure C.I:    Fundamental Morphological Structure Types
                                 80

-------
                                        Record primary structure as one cluster
                                        designated C.
                                        Record primary structure as 1 cluster
                                        designated as CSn.  Record 5 fibers  as
                                        components.   Use code CFn for the component
                                        fibers.
                                        Record primary structure as 1 cluster
                                        designated CSn.  Record 3 fibers and 2 bundles
                                        as components.  Use code CFn for component
                                        fibers and code CBn for component bundles.
                                        Record primary structure as 1 cluster
                                        designated as CSn.  Record 4 fibers and 2 sub-
                                        clusters as components.  Use code CFn for
                                        component fibers and code CCn for component
                                        clusters.
     Notes:

     "n" is  a three digit number identifying the structure number of the
     primary structure that each component belongs to.

     Dimensions assigned to each component represent the best estimate of the
     maximum length and mean width of that particular component.   Dimensions
     recorded for a primary structure represent the best estimate of the
     overall outside dimensions of that structure.
Figure C.2:    Examples of Recording of Complex Asbestos Clusters

                                   81

-------
                                        Record primary structure as 1 matrix
                                        designated as MSn.  Record 1 fiber as a
                                        component.  Use code MFn for component fiber.
                                            Record primary structure as 1 matrix
                                            designated as MSn.  Record 5 fibers as
                                            components.  Use code MFn for component
                                            fibers.
                                        Record primary structure as 1 matrix
                                        designated as MSn.   Record 3 fibers  and 1 sub-
                                        matrix as components.   Use code MFn  for
                                        component fibers and code MMn for component
                                        matrix.
                                        Record as 1 matrix designated as M.
     Notes:

     "n" is a three digit number identifying the structure number of the
     primary structure that each component belongs to.

     Dimensions assigned to each component represent the best estimate of the
     maximum length and mean width of that particular component.  Dimensions
     recorded for a primary structure represent the best estimate of the
     overall outside dimensions of that structure.

     Submatrices are only recorded when they incorporate asbestos fibers not
     accounted for after separation of fiber, bundle, and cluster components
     from the primary structure.
Figure C.3:    Examples of Recording of Complex Asbestos Matrices

                                      82

-------
                                               -*-Grid opening
       Figure C.4:  Counting of Structures That Intersect Grid Bars
                                                Grid opening
                                                TEM field of view
Figure C.5:   Counting of Structures That Extend Outside the Field of View
                                83

-------
APPENDIX D - FTBER IDENTIFICATION PROCEDURE

 D.1  INTRODUCTION

    Fibrous structures are classified as asbestos structures by identifying
 one  or more of the constituent fibers  or bundles as asbestos.  Accordingly,
 this section refers  to identification  of fibers rather than structures.

    The criteria which shall be used for identification of asbestos fibers are
 selected,  depending  on the  intended use of the  fiber  counting  results.   In
 some circumstances,  there can be a  requirement  that fibers shall  be unequi-
 vocally  identified as a  specific mineral species.  In other circumstances
 there  can be sufficient  knowledge about the sample that  rigorous  identifica-
 tion of  each fiber need  not be carried out.  The time required to perform
 the  analysis, and therefore the cost of analysis, can vary widely depending
 on the identification criteria which are considered to be  sufficiently
 definitive.  The combination  of criteria considered  definitive for
 Identification  of fibers in a particular analysis shall  be specified before.
 the  analysis is  made, and this combination of criteria shall be referred to
 as  the "Level"  of analysis.  Various factors related  to  instrumental
 limitations and the  character of the sample may prevent  satisfaction of all
 of the specified fiber identification  criteria  for a  particular fiber.
 Therefore, a record  shall be  made  of the identification  criteria  which  were
 satisfied for each suspected  asbestos  fiber included  in  the analysis.   For
 example, if both ED  and  EDXA  were  specified to  be attempted for definitive
  identification  of each fiber, fibers with chrysotile  morphology which,  for
  some reason, do not  give an ED pattern but which do yield an  EDXA spectrum
 corresponding to chrysotile,  are categorized in a way that conveys  the  level
 of confidence to be  placed in the  identification.

  D.2-  ED  AND EDXA TECHNIQUES

    D.2.1 General

          Initially,  classify  fibers into two categories  on the basis of
    morphology: those fibers with tubular morphology,  and those fibers without
    tubular morphology.   Conduct further analysis of each fiber using ED  and
    EDXA  methods.  The following procedures should be  used when fibers are
    examined by ED and EDXA.

          The crystal structures of  some mineral fibers,  such as chrysotile,
    are easily damaged by the  high current densities required for  EDXA
    examination.  Therefore, investigation of these sensitive fibers by  ED
    shall be completed before  attempts  are made  to obtain EDXA  spectra from
    the fibers.   When more stable fibers, such as the  amphiboles,  are
    examined, EDXA and ED may  be used in either  order.
                                       84

-------
 D.2.2   ED Techniques

       The ED  technique  can be  either  qualitative  or  quantitative.
 Qualitative ED consists of visual  examination,  without  detailed
 measurement,  of  the general characteristics  of  the ED pattern obtained on
 the  TEH viewing  screen  from a  randomly orientated fiber.   ED patterns
 obtained  from fibers with cylindrical symmetry, such as chrysotile,  do not
 change when the  fibers  are tilted  about  their axes,  and patterns  from
 randomly  oriented fibers of these  minerals can  be interpreted
 quantitatively.  For fibers that do not  have cylindrical  symmetry, only
 those  ED  patterns obtained when the fiber is oriented with a principal
 crystallographic axis closely  parallel with  the incident  electron beam
 direction can be interpreted quantitatively.  This type of ED pattern
 shall  be  referred to as a zone-axis ED pattern.   In  order to interpret a
 zone-axis ED  pattern quantitatively,  it  shall be  photographed and its
 consistency with known  mineral structures shall be checked.   A computer
 program may be used to  compare measurements  of  the zone-axis ED pattern
 with corresponding data calculated from  known mineral structures.  The
 zone-axis ED  pattern obtained  by examination of a fiber in a particular
 orientation can  be insufficiently  specific to permit unequivocal
 identification of the mineral  fiber,  but it  is  often possible to  tilt  the
 fiber  to  another angle  and to  photograph a different ED pattern
 corresponding to another zone-axis.   The angle  between  the  two  zone-axes
 can  also  be checked for consistency with the structure  of a suspected
 mineral.

       For visual examination of the ED pattern, the  camera  length of the
 TEM  should be set to a  low value of approximately 250 mm  and the ED
 pattern then  should be viewed  through the binoculars.   This  procedure
 minimizes the possible degradation of the fiber by the  electron
 irradiation.   However,   the pattern is distorted by the  tilt  angle of the
 viewing screen.   A camera length of at least 2 m  should be used when the
 ED pattern is photographed,  if accurate measurement  of  the pattern is  to
 be possible.   It is necessary  that, when obtaining an ED pattern to be
 evaluated visually or to be photographed, the sample height  shall be
 properly adjusted to the eucentric point and the  image  shall be focussed
 in the plane of the selected area aperture.   If this is not  done there may
 be some components of the ED pattern which do not originate  from the
 selected area.  In general,  it will be necessary  to use the  smallest
 available ED aperture.

      For routine sample analysis,  calibration films of evaporated gold or
 other materials  shall not be applied to  the  TEM grids.   Such films
 seriously degrade the visibility of fine chrysotile fibers so that they
may not be detected by  the EM operator.   Moreover, the visibility of ED
patterns from chrysotile fibers are also seriously degraded resulting in
 failure to identify them positively.   Both effects lead to significant
reduction of  the  asbestos structure concentrations reported.
                                  85

-------
      To form an ED pattern, move the image of the fiber to the center of
the viewing screen and insert a suitable selected area aperture into the
electron beam so that the fiber, or a portion of it, occupies a large
proportion of the illuminated area.  The size of the aperture and the
portion of the fiber shall be such that particles other than the one to be
examined are excluded from the selected area.  Observe the ED pattern
through the binoculars.  During the observation, the objective lens
current should be adjusted to the point where the most complete ED pattern
is obtained.  If an incomplete ED pattern is still obtained, move the
particle around within the selected area to attempt to optimize the ED
pattern, or to eliminate possible interferences from neighboring
particles.

      If a zone-axis ED analysis is to be attempted on the fiber, the
sample shall be mounted in the appropriate holder.  The most convenient
holder allows complete rotation of the specimen grid and tilting of the
grid about a single axis.  Rotate the sample until the fiber image
indicates that the fiber is oriented with its length coincident with the
tilt axis of the goniometer, and adjust the sample height until the fiber
is at the eucentric position.  Tilt the fiber until an ED pattern appears,
which is a symmetrical, two dimensional array of spots.  The recognition
of zone-axis alignment conditions requires some experience on the part of
the operator.  During tilting of the fiber to obtain zone-axis conditions,
the manner in which the intensities of the spots vary should be observed.
If weak reflections occur at some points on a matrix of strong
reflections, the possibility of multiple diffraction exists, and some
caution should be exercised in the selection of diffraction spots for
measurement and interpretation.  A full discussion of electron diffraction
and multiple diffraction can be found in the references by J. A. Card,
P.B. Hirsch et al, and H. R. Wenk, included in the Bibliography.  Not all
zone-axis patterns that can be obtained are definitive.  Only those which
have closely-spaced reflections corresponding to low indices in at least
one direction should be recorded.  Patterns in which all d-spacings are
less than about 0.3 nm are not definitive.  A useful guideline is that the
lowest angle reflections should be within the radius of the first gold
diffraction ring (111), and that patterns with smaller distances between
reflections are usually the most definitive.

      Five spots, closest to the center spot, along two intersecting lines
of the zone-axis pattern shall be selected for measurement, as illustrated
in Figure Dl.  The distances of these spots from the center spot and the
four angles shown provide the required data for analysis.  Since the
center spot is usually very over-exposed, it does not provide a
well-defined origin for these measurements.  The required distances shall
therefore be obtained by measuring between pairs of spots symmetrically
disposed about the center spot, preferably separated by several repeat
distances.  The distances must be measured with a precision of better than
0.3 mm, and the angles to a precision of better than 2.5°.  The diameter
of the first or second ring of the calibration pattern (111 and 200) must
also be measured with a precision of better than 0.3 mm.

                                   86

-------
         Using  gold as  the  calibration material,  the  camera  constant  is  given


                A x  L =  2.3548 x D  (first  ring)
                A x  L -  2.0393 x D  (second ring)

   D.2.3  EDXA  Measurements

         Interpretation of  the EDXA spectrum may be either qualitative or
   quantitative.  For qualitative  interpretation of a spectrum, the elements
   originating  from the fiber are  recorded.  For quantitative interpretation,
   the net peak areas,  after background subtraction, are obtained for the
   elements originating from the fiber.  This method provides for
   quantitative interpretation for those minerals which contain silicon   To
   obtain an EDXA spectrum, move the image of the fiber to the center of the
   screen and remove the objective aperture.  Select an appropriate electron
   beam diameter and deflect the beam so that it impinges on the fiber.
   Depending on the instrumentation,  it may be necessary to tilt the specimen
   towards the X-ray detector,  and in some instruments,  to use Scanning
   Transmission Electron Microscopy (STEM) mode of operation.

         The time for acquisition of a suitable spectrum varies with the
   fiber, diameter,  and also with instrumental factors.  For quantitative
   interpretation,  spectra should have statistically valid number of counts
   in each peak.  Analyses of small diameter fibers which contain sodium are
   the most critical, since it  is  in the  low energy range that the X-ray
   detector is least sensitive.  Accordingly,  it is necessary to acquire a
   spectrum for a sufficiently  long period that the presence  of sodium can be
   detected in such fibers.   It has been  found that satisfactory quantitative
   analyses  can be  obtained if  acquisition is  continued  until the  background
   subtracted silicon Ka peak integral exceeds  10,000  counts.   The spectrum
   should then be manipulated to subtract  the  background and  to  obtain the
   net areas  of the  elemental peaks.

         After quantitative  EDXA classification of some  fibers by  computer
   analysis of the net peak  areas,  it  may  be possible  to classify  further
   fibers  in  the same sample in the basis  of comparison  of  spectra at  the
   instrument.   Frequently,  visual  comparisons  can be  made  after somewhat
   shorter acquisition times.

D.3  INTERPRETATION OF FIBER ANALYSIS DATA

   D.3.1  Chrysotile

        The morphological structure of chrysotile is characteristic  and
   with experience,  can  be recognized readily.  However, a few other minerals
   have similar  appearance, and morphological observation by itself is
   inadequate for most samples.   The ED pattern obtained from chrysotile  is
   quite specific for this mineral if the specified characteristics of the

                                    87

-------
pattern correspond to those from reference chrysotile.  However,  depending
on the past history of the fiber, and on a number of other factors,  the
crystal structure of a particular fiber may be damaged, and it may not
yield an ED pattern.  In this case, the EDXA spectrum may be the only data
available to supplement the morphological observations.

D.3.2  Amphiboles

      Since the fiber identification procedure for asbestos fibers other
than chrysotile can be involved and time-consuming, computer program such
as developed by Rhoades (see Bibliography) can be used for interpretation
of zone-axis ED patterns.  The published literature contains composition
and crystallographic data for all of the fibrous minerals likely to be
encountered in TEH analysis of air samples, and the compositional and
structural data from the unknown fiber should be compared with the
published data.  Demonstration that the measurements are consistent with
the data for a particular test mineral does not uniquely identify the
unknown, since the possibility exists that data from other minerals may
also be consistent.  It is, however, unlikely that a mineral of another
structural class could yield data consistent with that from an amphibole
fiber identified by quantitative EDXA and two zone axis ED patterns.

      Suspected amphibole fibers should be classified  initially on the
basis of chemical composition.  Either qualitative or  quantitative EDXA
information may be used as the basis for this classification.  From the
published data on mineral compositions, a list of minerals which are
consistent in composition with that measured for the unknown fiber should
be compiled.  To proceed further,  it is necessary to obtain the first zone
axis ED pattern, according to the  instructions in D.2.1.

      It is possible to specify a particular zone-axis pattern for
identification of amphibole, since a few patterns are  often considered to
be characteristic.  Unfortunately, for a fiber with random orientation on
a TEM grid, no specimen holder and goniometer currently available will
permit convenient and rapid location of two pre-selected zone-axes.  The
most practical approach has been adopted, which  is to  accept those low
index patterns which are easily obtained, and then to  test their
consistency with the structures  of the minerals  already pre-selected on
the basis of the EDXA data.  Even  the structures of non-amphibole minerals
in this pre-selected list must be  tested against the  zone-axis data
obtained for the unknown fiber,  since non-amphibole minerals in some
orientations may yield similar patterns consistent with amphibole struc-
tures .
                                   88

-------
                                  SPOT I.
          SPOT 3
                      SPOT 2.
           SPOT 4
             SPOT 5
Figure D.I:    Measurement of Zone Axis SAED Patterns
                        89

-------
         The zone-axis ED interpretation must include all minerals previously
   selected from the mineral data file as being chemically compatible with
   the EDXA data.   This procedure will usually shorten the list of minerals
   for which solutions have been found.  A second set of zone-axis data from
   another pattern obtained on the same fiber can then be processed,  either
   as further confirmation of the identification, or to attempt elimination
   of an ambiguity.  In addition, the angle measured between the orientations
   of the two zone-axes can be checked for consistency with the structures of
   the minerals.  Caution should be exercised in rationalizing the inter-zone
   axis angle,  since if the fiber contains c-axis twinning the two zone-axis
   ED patterns may originate from the separate twin crystals.   In practice,
   the full identification procedure will normally be applied to very few
   fibers,  unless  for a particular reason precise identification of all
   fibers is required.

D.4  FIBER  CLASSIFICATION CATEGORIES

   It is not always possible to proceed to a definitive identification of a
fiber;  this may be due  to instrumental limitations or  to  the actual nature
of the  fiber.   In many  analyses  a definitive  identification of each fiber
may not actually be necessary  if there is other  knowledge available about
the sample, or  if the concentration is below  a level of interest.  The
analytical procedure must therefore take account of both  instrumental
limitations and varied  analytical requirements.  Accordingly, a system for
fiber classification  is used to  permit accurate .recording of data.  The
classifications are shown in Tables Dl and D2, and are directed towards
identification  of chrysotile and amphibole respectively.  Fibers shall be
reported in these categories.

   The general  principle to  be  followed in this  analytical procedure  is  first
to define the most specific fiber classification which is to be attempted,
or the  "level"  of analysis to be conducted.  Then, for each fiber examined,
record  the classification which  is actually achieved.  Depending on the
intended use of the results, criteria for acceptance of fibers as "iden-
tified" can then be established  at any time after completion of the
analysis.

   In  an unknown sample,  chrysotile will  be  regarded  as confirmed  only if a
recorded, calibrated ED pattern  from one fiber in the CD categories is
obtained, or if measurements of  the ED pattern are recorded at the
instrument.  Amphibole will be regarded as confirmed only by obtaining
recorded data which yields exclusively amphibole solutions for fibers
classified in the AQZ, AZZ or AQZZ categories.

  D.4.1 Procedure  for Classification of  Fibers with  Tubular Morphology,
         Suspected  to  be Chrysotile

         Occasionally, fibers are encountered  which have tubular morphology
  similar to that of  chrysotile, but which cannot be characterized further
  either by ED  or EDXA.  They  may be non-crystalline,  in  which case ED

                                     90

-------
techniques are not useful, or  they may be  in a position on  the  grid which
does not permit an EDXA  spectrum  to be obtained.  Alternatively,  the fiber
may be of organic origin, but  not sufficiently definitive that  it can be
disregarded.  Accordingly, there  is a requirement to record the fiber, and
to specify how confidently each fiber can  be identified.  Classification
of fibers will meet with various  degrees of success.  Figure D2 shows the
classification procedure to be used for fibers which display any tubular
morphology.  The chart is self explanatory, and every fiber is  either
rejected as a non-asbestos mineral (NAM),  or classified in  some way which
by some later criterion  could  still contribute to the chrysotile fiber
count.

      Morphology is the  first  consideration, and if this is not similar to
that usually seen in chrysotile standard samples, designate the initial
classification as TM.  Regardless of the doubtful morphology, examine the
fiber by ED and EDXA methods according to  Figure D2.  Where the morphology
is more definitive, it may be  possible to  classify the fiber as having
chrysotile morphology (CM).
are:
      For classification as CM, the morphological characteristics required
      (a)   the individual fibrils should have high aspect ratios
            exceeding 10:1, and be about 20 to 40 nm in diameter;

      (b)   the electron scattering power of the fiber at 60 to 100 kV
            accelerating potential should be sufficiently low for internal
            structure to be visible; and,

      (c)   there should be some evidence of internal structure suggesting
            a tubular appearance similar to that shown by reference UICC
            chrysotile, which may degrade in the electron beam.

      Examine every fiber having these morphological characteristics by
the ED technique,  and classify as chrysotile by ED (CD) only those which
give diffraction patterns with the precise characteristics shown in Figure
D3.  The relevant features in this pattern for identification of
chrysotile are as follows:

      (a)   the (002) reflections should be examined to determine that
            they correspond approximately to a spacing of 0.73 nm;

      (b)   the layer line repeat distance should correspond to 0.53 nm;
            and,

      (c)   there should be "streaking" of the (110)  and (130)
            reflections.
                                  91

-------
                         FIBER  WITH  TUBULAR-MORPHOLOGY
                           Is  fiber morphology characteristic
                           of  that displayed by reference chrysotile?
                             NO
           Examine by SAED
   Pattern not
   chrysotile
Chrysotile
pattern
         Pattern not present
           or indistinct
     Examine by quantitative EDXA
Composition not
that of chrysotile
   Chrysotile
   composition
              No Spectrum
                 TM
                             YES
                                      Examine by SAED
Chrysotile
pattern
Pattern  not
chrysotile
                                     Pattern not present
                                        or  indistinct
                               Examine by  quantitative EDXA


Chrysotile
composition
No S
CMQ


Composition not
that of chrysoti
Dectrum
CM
le

NAM
                              Examine by quantitative EDXA
                        Composition not
                        that of chrysotile
                            Chrysotile
                            composition
                                      No Spectrum
                                                             {CDQ j
   Figure D.2:     Classification Chart for Fiber With Tubular Morphology
                                           92

-------
            /ass
                nm
                          X
               0002,
Figure D.3:     Chrysotile SAED Pattern
                 93

-------
      Using the millimeter calibrations on the TEM viewing screen,  these
observations can readily be made at the instrument.  If documentary proof
of fiber identification is required, record a TEM micrograph of at least
one representative fiber, and record its ED pattern on a separate film or
plate.  This film or plate shall also carry calibration rings from a known
polycrystalline substance such as gold.  This calibrated pattern is the
only documentary proof that the particular fiber is chrysotile, and not
some other tubular or scrolled species such as halloysite, palygorskite,
talc or vermiculite.  The proportion of fibers which can be successfully
identified as chrysotile by ED is variable, and to some extent dependent
on both the instrument and the procedures of the operator.  The fibers
that fail to yield an identifiable ED pattern will remain in the TM or CM
categories unless they are examined by EDXA.

      In the EDXA analysis of chrysotile there are only two elements which
are relevant.  For fiber classification, the EDXA analysis must be
quantitative.  If the spectrum displays prominent peaks from magnesium and
silicon, with their area in the appropriate ratio, and with only minor
peaks from other elements, classify the fiber as chrysotile by
quantitative EDXA, in the categories CQ, CMQ, or CDQ, as appropriate.

D.4.2  Procedure for classification of Fibers Without Tubular
        Morphology, Suspected to be Amphibole

      Every particle without tubular morphology and which is not obviously
of biological origin, with an aspect ratio of 5:1 or greater, and having
parallel or stepped sides, shall be considered as a suspected amphibole
fiber.  Further examination of the fiber by ED and EDXA techniques will
meet with a variable degree of success, depending on the nature of the
fiber and on a number of instrumental limitations.  It will not be
possible to identify every fiber completely, even  if time and cost were of
no concern.  Moreover, confirmation of the presence of amphibole can be
achieved only by quantitative interpretation of zone-axis ED patterns, a
very time-consuming procedure.  Accordingly, for routine samples from
unknown sources, this analytical procedure limits  the requirement for
zone-axis ED work to a minimum of one fiber representative of each
compositional class reported.  In some samples, it may be necessary to
identify more fibers by  the zone-axis technique.  When analyzing samples
from well-characterized  sources, the cost of identification by zone-axis
methods may not be justified.

      The 0.53 nm layer  spacing of  the random orientation ED pattern is
not by itself diagnostic for amphibole.  However,  the presence of c-axis
twinning in many fibers  leads to contributions to  the layers in the
patterns by several individual parallel crystals of different axial
orientations.  This apparently random positioning  of the spots along the
layer lines, if also associated with a high fiber  aspect ratio, is a
characteristic of araphibole asbestos, and  thus has some limited diagnostic
value.  If a pattern of  this type is not obtained, the identity of the
                                   94

-------
fiber is still ambiguous, since the absence of a recognizable pattern may
be a consequence of an unsuitable orientation relative to the electron
beam, or the fiber may be some other mineral species.

      Figure D4 shows the fiber classification chart to be used for
suspected amphibole fibers.  This chart shows all the classification paths
possible in analysis of a suspected'amphibole fiber, when examined
systematically by ED and EDXA.  Two routes are possible, depending on
whether an attempt to obtain an EDXA spectrum or a random orientation ED
pattern is made first.  The normal procedure for analysis of a sample of
unknown origin will be to examine the fiber by random orientation ED,
qualitative EDXA, quantitative EDXA, and zone-axis ED, in this sequence.
The final fiber classification assigned will be defined either by
successful analysis at the maximum required level, or by the instrumental
limitations.  Record the maximum classification achieved for each fiber on
the counting sheet in the appropriate column.  The various classification
categories can then be combined later in any desired way for calculation
of the fiber concentration, and a complete record of the results obtained
when attempting to identify each fiber is maintained for reassessment of
the data if necessary.

      In the unknown sample, zone-axis analysis will be required if the
presence of amphibole is to be unequivocally confirmed.  For this level of
analysis,  attempt to raise the classification of every suspected amphibole
fiber to the ADQ category by inspection of the random orientation ED
pattern and the EDXA spectrum.  In addition,  examine at least one fiber
from each type of suspected amphibole found by zone-axis methods to
confirm their identification.   In most cases, some ambiguity of
identification can be accepted,  because information exists about possible
sources of asbestos in close proximity to the air sampling location.
Lower levels of analysis can therefore be accepted for these situations.
                                  95

-------
                          [        FIBER WITHOUT TUBULAR MORPHOLOGY
                     Ootk 
-------
Table Dl
Classification of fibers with tubular morphology
        TM



        CM


        CD


        CQ


        CMQ



        CDQ



        NAM
Tubular Morphology, not sufficiently characteristic for
classification as chrysotile

Characteristic Chrysotile Morphology

Chrysotile SAED pattern

Chrysotile composition by Quantitative EDXA

Chrysotile Morphology and composition by Quantitative
EDXA

Chrysotile SAED pattern and composition by Quantitative
EDXA

Non-asbestos Mineral
                                    97

-------
Table D2
Classification of fibers without tubular morphology
        UF

        AD


        AX


        ADX


        AQ

        AZ

        ADQ


        AQZ


        AZZ


        AQZZ


        NAM
Unidentified Fiber

Amphibole by random orientation SAED (shows layer
pattern of 0.53 nm spacing)

Amphibole by qualitative EDXA.  Spectrum has elemental
components consistent with amphibole

Amphibole by random orientation SAED and qualitative
EDXA

Amphibole by quantitative EDXA

Amphibole by one Zone-Axis SAED pattern

Amphibole by random orientation SAED and Quantitative
EDXA

Amphibole by Quantitative EDXA and one Zone-Axis SAED
pattern

Amphibole by two Zone-Axis SAED patterns with consistent
inter-axial angle

Amphibole by Quantitative EDXA, two Zone-Axis SAFD
patterns, and consistent inter-axial angle

Non-asbestos Mineral
                                    98

-------
APPENDIX E - CALCULATION OF RESULTS

 E.1  INTRODUCTION

    Calculate the results using the procedures described below.   The results
 can be conveniently calculated using a computer program.

 E.2  TEST FOR UNIFORMITY OF FIBER DEPOSITS ON TEM GRIDS

   A check shall be made, using the chi-square test,  to determine whether the
 asbestos structures found on  individual grid openings are randomly and
 uniformly distributed/among the grid openings.  If• the total number found  in
 k grid openings  is n, and the areas of the k individual grid openings are
 designated At to Ak,  then the  total  area  of TEM specimen examined is:
                       A  =
 The fraction of the .total area examined' which is represented by the
 individual grid opening area, pi(  is given by A±/A.   If the  structures  are
 randomly and uniformly dispersed over the k grid openings examined, the
 expected number of structures falling in one grid opening with area ^  is
 np^   If the observed number of structures; found on that grid opening is n,,
 then:
                 x2
Z
                                    (nt -
 This value shall be compared with significance points of the x2 distribu-
 tion, having (k - 1) degrees of freedom.  Significance levels lower than
 0.1% may be  cause for the sample analysis to be rejected,  since this
 corresponds to a very inhomogeneous deposit.   If the structure count fails
 this test, the precision of the result will be uncertain,  and new analytical
 filters should be prepared from the original collection filter.
                                     99

-------
E.3  CALCULATION OF THE ANALYTICAL SENSITIVITY

   Calculate  the  analytical  sensitivity S,  in  structures/liter, using the
formula:
where:
             Af
             N
             AS
             V
             F
S  -  Af/(N x Ag x V x F)


 -  Active area of  sample  collection filter in mm2
 —  Number of grid openings examined
 -  Mean area of grid openings
 -  Volume of air sampled in liters
 —  Concentration factor
E.4  CALCULATION OF THE MEAN AND CONFIDENCE LIMITS FOR A REPORTED
          ASBESTOS  STRUCTURE CONCENTRATION

   In the asbestos  structure count made according to  this  method,  a number  of
grid openings have been sampled from a population of grid openings and it is
required  to determine the mean grid opening asbestos structure count for the
population on the basis of  this small sample.  The upper and lower 95%
confidence limits are also  to be reported with every mean.  When such limits
are constructed as indicated in Section E.4.2, the interval between the two
limits should contain the true mean 90% of the time  (95% of the time if the
lower confidence limit is zero).

   E.4.1  Calculation of the Mean Asbestos Structure  Concentration

         Calculate  the mean concentration from a particular measurement,  C,
   in structures/liter:
               C  -  S x N x n
   where:
         S  -  Analytical sensitivity in structures/liter
         N  -  Number of grid openings examined
         n  -  mean number of asbestos structures found per grid opening

   E.4.2 Calculation of Upper and Lower 95% Confidence Limits

         If the total structures counted from a filter is 30 or less, the
   Poisson distribution is recommended to be used for constructing confidence
   limits.  Let P(x;m) represent the cumulative probability function from a
   Poisson distribution with mean m.  That is, P(x;m) is the probability of
   observing x or fewer counts given by
P(x;m) - e'm(l
          m
                             m2/2!
+ mx/x!).
                                     100

-------
The upper 95% confidence limit, xa, on the mean count, based on an
observed count of x, is the value  that satisfies:

      P(X)XU) = 0.05,

and the lower 95% confidence limit, XL,  on the mean count, based on an
observed count of x, is the value  that satisfies

      1 - P(x-l,xL)  - 0.05

(e.g., Miller and Freund 1965)5.  Tables of P(x;m) for values of m less
than 26 may be found in Miller and Freund.  A computer program can be
easily written for calculating P(x;m) for larger values of m.  If such a
program is not available, confidence limits may alternately be based on
the normal distribution approximation to the Poisson distribution (see
Miller and Freund 1965).  Corresponding approximations to the upper and
lower 95% confidence limits are
         — X
               1.65*x1/2
and
         =• x -  1.65*x1/2.
      The Poisson distribution is correct whenever structures are randomly
distributed in the sampled air and on the filter.  However, as a hedge
   Confidence limits derived as described in this section are one-
   tailed limits meaning that 95% of the distribution lies below the
   upper 95% confidence limit and, similarly, 95% of the distribution
   lies above the lower 95% confidence limit.  If the two limits are
   combined to create a confidence interval, the interval between the
   two limits represents 90% of the distribution (which is different
   than a 95% confidence interval).

   Construction of confidence limits in this manner, rather than the
   more traditional approach of constructing confidence intervals
   (where, for example, the 95% confidence interval corresponds to
   upper and lower 97.5% confidence limits), is recommended to avoid
   confusion when referring to an asymmetric distribution or when the
   lower confidence limit falls to zero.  When the lower confidence
   limit falls to zero, for example, the upper 95% confidence limit now
   corresponds to the upper 95% confidence interval (rather than the
   90% confidence interval, as indicated in the last paragraph) because
   the lower confidence limit (zero) removes nothing from the
   distribution.   Thus, we have chosen to refer to one-tailed
   confidence limits throughout and avoid emphasis on confidence
   intervals.
                                  101

-------
against possible clumping of structures, it is recommended that, whenever
the total structure count on a filter is greater than 30, confidence
limits also be calculated using the following procedure and the more
extreme limits (i.e., larger Xu and smaller XL) be  reported.

      Suppose k grid openings are examined and let xt be the structure
count for the ith grid opening.  Note that a minimum of four grid openings
(k>4) should be examined and they should be of equal size.  Then 95% upper
and lower confidence limits are calculated according to the formulas

      xu - 1.65s

      XT  - 1.65s
Where
s2 - tk/(k-l)
k

E
- x/k)2,
and 1.65 is the cumulative 95th percentage point of the standard normal
distribution.

      Upper or lower confidence limits, for airborne concentrations derived
from a particular measurement are computed by multiplying the
corresponding upper or lower confidence limits derived as described above
for the structure count by the analytical sensitivity of the measurement.


E.4.3 Reporting the Mean and Confidence Limits for a Measured
      Concentration

      Where a result is to be interpreted as a single isolated
measurement, the following procedures shall be used for reporting:

No structures detected - the concentration shall be reported as "NF"
meaning "not found" and it should be accompanied by the upper and lower
95% confidence limits derived as described in Section E.4.2.  Note that
the lower 95% confidence limit is zero in this case.

At least 1 structure detected - report the mean concentration derived as
described in Section E.4.1 along with the upper and lower 95% confidence
limits derived as described in Section E.4.2.

      Rules for combining the results of individual measurements as part
of a site evaluation are presented in the Technical Background Document,
Part 2 of this report.
                                  102

-------
  E.5   DISTINGUISHING DETECTED ASBESTOS FROM ANALYTICAL BACKGROUND
        non-parametric Wilcoxon Test (Hollander and Wolfe 1973) is employed in
  this method  to test  for significant differences between asbestos
  concentrations (in units of s/cm2) observed on sample filters and
  concentrations observed on blanks.  It must be emphasized, however,  that  the

  than indf^/1?   atSi0n °f ^ teSt iS tO ^°^S °f samPle resu^  "ther
  than individual results.  This should not be a severe limitation because
  conclusions based on the totality of data from an entire study tend  to be
  more important than the results from individual analyses.

    The  Wilcoxon test  performs  better on multiple samples.   A test comparing
  the concentration from an individual filter to those from a group of blanks

         0^   '  ^ SUCh a teSt 1S Ukely t0 be of low P°w*r ansensiSve
          di!fe"nces).  In fact,  data from at least 20 blanks are likely
             Hr~    teS? beC°meS  CSpable °f exhibiting a statistically
             dlfference (p < -0.05) between blank concentrations and  the
         n   ? °bSTed °n S Slngle Sample filter-   ^ P°wer of the  test
 a    SS    £    ^   d ^^ comParing multiple  sample filter analyses to
 a limited number  of blanks.   For  example,  at least seven sample filters
 b~lankl           ^  ^ C°mbined  t0 find S siSnific-nt difference from two
                             WU1 *enerally be available from two sources:
 are       i         hlf°rical-   Within a particular study,  two filter blanks
 be fSS M lyvanalred fr°m eaCh lot °f filt^s.   Typically,  there will also
 be field blanks    Depending on  the  length of time that a particular
 ac^af0?    ^6n USlng  thlS  meth°d'  the laboratory should also be
 accumulating a historical  database  of blank analyses!  Either set of blank
 data may be  combined with  care  to provide the needed data to  perform the
 Wilcoxin Test and  distinguish sampled asbestos from analytical

 levels  of"* bTk H°Ter> lf different lots of alter, exhibit different
 levels  of background contamination  or if results  from a laboratory  chanee
 over time (due possibly to  changes  in methodology or analys^s'then i* Xy
      er0lae           ? !jist°rical  blank d^a in analysis .   On the
                             l data  are representative,  use of this  data may
                 the power  of the analysis  for detecting airborne
 this  methi    hist°rical blank  da*a are used  in «%tatl.tlc.l  test in
 tnis  method,  this is to be  noted and  results of corresponding tests  based
 only  on concurrent blanks shall also be reported.      ponainS Cests  based
                Pr°cedur,e can be used  f°r  applying the  Wilcoxon test to
                    results'  More detailed discussions of the procedure are
                     113 S°UrCeS ^^luding Hollander and Wolfe 1973).
        M blank and N non-blank samples are available.  Let X,       X« be
the measured filter concentrations from the blank samples and Y" "  %! the

 "      COcnoCebtrat:irSMfMr0m ^ n°n-blank -"»Pl- 
-------
 exhibiting the same concentration) among the M+N observations, assign the
 average rank to each group of tied observations.  (E.g., if the observations
 are 0, 0, 0, 0, 0.001, 0.002, then the corresponding ranks would be 2.5,
 2.5, 2.5, 2.5, 5, 6.)  Let W be the sum of the ranks assigned to the
 concentrations from the non-blank samples.  Large values of W are evidence
 that concentrations from non-blank samples are larger than those from blank
 samples, and the hypothesis that the blank and non-blank samples have equal
 concentrations of fibers is rejected if W is large enough.

    Tables are  available for determining when W is  large  enough to be
 statistically significant in cases in which there are no ties and M and N
 are small (see e.g., Table A.5 in Hollander and Wolfe 1973).  As an example
 of the application of  these tables, if N - 4 and M - 5, then W, the sum of
 the ranks of the concentrations from the non-blank samples, would have to be
 28 or larger before it could be concluded at the 0.05 level of significance
 that the concentrations from the non-blanks are higher  than those from the
 blanks (see Table A.8  in Hollander and Wolfe 1973, page 276).

    If ties are present in the data or if the number of samples is  so  large
 that tables are not available, a normal approximation to the distribution of
 W may be used.  Define

    W* - W -  rN(M+N+l)/21
              [Var(W)]1/2
where
Var(W) - M
         12
                  M+N+1 -
S tj(tj2-l)
j-1	
(M+N) (M+N-1) J
 where  g  is  the  number  of groups  of tied concentrations  (included among the
 combined set  of blank  and nonblank samples)  and tj is the size of tied group
 j.  The  hypothesis  test is based on an assumed normal distribution for W*.
 Thus one would  reject  the hypothesis of equal  concentrations on blanks and
 non-blanks  at the 0.05 level of  significance if W*>1.65.  (Note:  This
 normal approximation may not work well if the  number of samples is small.
 Consequently, if  there are ties  in the data  and M or N  is less than about 4,
 consideration should be given to consulting  with a statistician regarding a
 more accurate approach to computing the statistical significance of W.)

 E.6    CALCULATION OF ASBESTOS STRUCTURE LENGTH,  WIDTH,  AND ASPECT RATIO
       DISTRIBUTIONS

    The distributions all approximate to logarithmic-normal,  and so the size
 range  intervals for calculation  of the distribution shall be spaced
 logarithmically.  The  other characteristics  required for the choice of size
 intervals are that  they should allow for a sufficient number of size
 classes, while  still retaining a statistically-valid number of asbestos
                                      104

-------
structures in each class.  Interpretation is also facilitated if each size
class repeats at decade intervals.  A ratio from one class to the next of
1.47 satisfies all of these requirements and this value shall be used   The
distributions, being approximately logarithmic-normal, when presented
graphically, shall be plotted using a logarithmic ordinate scale and a
Gaussian abscissa.
  E.6.1
 Calculation of Asbestos Structure Length Cumulative Number
 Distribution
        This distribution  allows  the  fraction of  the total number of
  asbestos structures either shorter  or  longer than a given length to be
  determined.  It is calculated using the relationship:

                         i=k
                          E
                          i-N
                                            x 100
  where:
          C(N)k  =
           nl
           N
           Cumulative fraction asbestos structures (expressed in
           percent) which have lengths less  than  the upper bound of
           the k'th class
           Number of asbestos structures in the i'th length class
           Total number of length classes.
 E.6.2
Calculation of Asbestos Structure Width Cumulative Number
     Distribution
       This distribution  allows  the  fraction of  the  total number of
 asbestos structures either narrower or wider  than a given width to be
 determined.  It is calculated in a  similar way  to that used in E.6.1, but
 using the asbestos structure widths.
                                   105

-------
E.6.3  Calculation of Asbestos Structure Aspect Ratio Cumulative
            Number Distribution

      This distribution allows the fraction of the total number of
asbestos structures which have aspect ratios either smaller or larger than
a given aspect ratio to be determined.  It is calculated in a similar way
to that used in E.6.1, but using the asbestos structure aspect ratios.
                                   106

-------
 APPENDIX F - BIBLIOGRAPHY

 Asbestos International Association (1979):  Reference method for the deter-
 mination of asbestos fibre concentrations at workplaces by light microscopy
 (membrane filter method).  AIA Health and Safety Publication, Recommended
 Technical Method No., 1 (RTM1);   Asbestos International Association  68
 Gloucester Place, London, W1H 3HL, England.

 Bradley, D.E.  (1965):  Replica and shadowing techniques.  In: Techniques for
 Electron Microscopy, (D.H. Kay,  Ed.).  Blackwell Scientific Publications
 Alden Press,  Oxford, England. 96-152.

 Burdett, G.J.  and Rood,  A.P.  (1982):   A Membrane-filter, direct transfer
 technique for the analysis of asbestos fibers or other inorganic particles by
 transmission electron microscopy.   Environmental Science and Technology  17
 643-648.                                                                   '

 Campbell,  W.J.,  Blake,  R.L.,  Brown,  L.L.,  Gather,  E.E.  and Sjoberg,  J.J.
 (1977):   Selected silicate minerals  and their asbestiform varieties.   Mineral-
 ogical definitions and identification-characterization.   Information Circular
 8751.   United States Department  of the Interior,  Bureau of Mines,  Washington,
 D • C . *        •

 Chatfield,  E.J.,  Dillon,  M.J. and  Stott, W.R.  (1983):   Development of improved
 analytical techniques for determination of asbestos  in  water  samples    EPA
 Report 600/4-83-042.  Available  through National  Technical Information
 Service,  5285  Port Royal  Road, Springfield,  VA: Order Number  PB83-261-471.

 Chatfield,  E.J.  (1986):   Asbestos  measurements  in workplaces  and ambient
 atmospheres.   In:  Electron Microscopy in Forensic, Occupational, and  Environ-
 mental Health  Sciences  (S. Basu  and J.R. Millette, Eds.).  Plenum Publishing
 Corporation, 233  Spring Street,  New York,  NY 10013,  149-186

 Chatfield,  E.J.  (Editor)  (1987):   Asbestos  fibre measurements in building
 atmospheres.  Ontario Research Foundation,  Sheridan  Park Research Community
 Mississauga, Ontario, Canada L5K 1B3.

 Chatfield,  E.J. and Lewis, G.M.  (1980):  Development and application of an
 analytical  technique for measurement  of asbestos fibers  in vermiculite   In-
 Scanning Electron Microscopy/1980/I,  (0. Johari, Ed.).   SEM Inc   AMF O'Hare
 Chicago, Illinois 60666, U.S.A.

 Cliff, G. and Lorimer, G.W. (1975):  The quantitative analysis of thin
 specimens.  Journal of Microscopy  103, 203-207.

Deer, W.A., Howie, R.A.  and Zussman,  J. (1963):  Rock-forming minerals.
Longman Group Limited, London, England, or Halsted Press, U.S.A.
                                     107

-------
        Reeister (1987):  Asbestos-containing materials in schools.   U.S.  En-
             rotection Agency.   Federal Register,  Vol.  42, No.  210,  October
30, 1987, 41826-41905.

Card, J.A. (Editor) (1971):  The Electron Optical Investigation of Clays.
Mineralogical Society, 41 Queen's Gate, London S.W.7.

Hawthorne, F.C. (1983):  The crystal chemistry of the amphiboles.  Canadian
Mineralogist, Vol. 21, Part 2, 173-480.

Hirsch, P.B., Howie, A., Nicholson, R.B., Pashley, D.W. and Whelan  MJ
(1965): Electron microscopy of thin crystals.  Butterworths, London, 18-23.

Hollahan, J.R. and Bell, A.T. (Editors)  (1974):  Techniques and  applications
of plasma chemistry.  Wiley, New York.

Hollander M, Wolfe D.   1973.  Nonparametric  Statistical Methods.  John Wiley
and  Sons, Inc., New York.

International  Centre  for Diffraction Data (1987):  Powder diffraction file.
International  Centre  for Diffraction Data, 1606 Park Lane, Swarthmore,
Pennsylvania 19081, U.S.A.

International  Mineralogical Association (1978):   Nomenclature  of amphiboles
 (compiled by B.E.  Leake).   Canadian Mineralogist,  16:501-520.

 International  Organization for  Standardization (1982):  Draft  International
 Standard DIS8672   Determination of airborne inorganic fibre concentrations  in
workplaces  by light microscopy (membrane filter method).

 Jaffe, M.S. (1948):   Handling and washing fragile replicas.  J.  Applied
 Physics, 19:1187.

 Joy, D.C.,  Romig Jr., A.D. and Goldstein, J.I. (Editors) (1986):  Principles
 of analytical electron microscopy.  Plenum Press, New York and London.

 Ledoux, R.L. (Editor)  (1979):  Short course in mineralogical techniques of
 asbestos determination.  Mineralogical Association of Canada,  Department of
 Mineralogy, Royal Ontario Museum, 100 Queen's Park,  Toronto,  Ontario,  Canada
 M5S 206.

 Lehmann E.  1975.  Nonparametries.  Statistical Methods  Based on Ranks.
 Holden-Day, Inc., San  Francisco.

 Levadie  B  (Editor)  (1984):  Definitions for asbestos and other health-
 related'silicates.   ASTM  Special  Technical  Publication 834.  American Society
 for Testing and Materials, 1916 Race  Street,  Philadelphia, Pennsylvania  19103,
 U.S.A.
                                       108

-------
 Michael,  J.R.  and Williams,  D.B.  (1987):   A consistent definition of probe
 size and spatial resolution in the analytical electron microscope   J  Mic
 147:289-303.

 Michaels,  L.  and Chissick,  S.S.  (Editors)  (1979):   Asbestos (Volume 1) ,
 properties,  applications and hazards.   Wiley,  New York.

 Miller I,  Freund J.   1965.   Probability and Statistics for Engineers.
 Prentice-Hall,  Inc.,  Englewood Cliffs,  New Jersey.

 National  Bureau of Standards Special Publication 506  (1978):   Workshop  on
 asbestos:  definitions and measurement methods.   U.S.  Government Printing
 Office Washington,  D.C.  20402.

 National  Bureau of Standards Special Publication 619  (1982):   Asbestos
 standards: materials  and analytical  methods.   U.S.  Government  Printing  Office
 Washington, D.C.  20402.

 National  Institute  for Occupational  Safety and Health (1984):   NIOSH Method
 7400.   U.S. Department of Health  and Human Services,  Public Health  Service,
 Centers for Disease Control,  National Institute  for Occupational  Safety  and
 Health, 4676 Columbia Parkway, Cincinnati,  Ohio  45226, U.S.A.

 National  Institute  for Occupational  Safety and Health (1986):   Asbestos
 Fibers, NIOSH Method  7402.   NIOSH Manual of Analytical Methods  5/17/86
 (Draft).

 Natrella, M.G.  (196.6):   Experimental statistics: National  Bureau  of  Standards
 Handbook 91.  U.S. Government Printing Office, Washington  D.C.  20402.

 Ortiz,  L.W. and Isom,  B.L. (1974):  Transfer technique for electron microscopy
 of membrane filter samples.   American Industrial Hygiene Association Journal
 35, 7,  423-425.

 Pearson, E.S. and Hartley, H.O. (1958):   Biometrika tables  for  statisticians,
Volume  1.  Cambridge University Press, 32  East 57th Street, New York  New
York, U.S.A.

Rhoades, B.L. (1976):  XIDENT- A computer  technique for the direct indexing of
electron diffraction spot patterns.  Research Report 70/76.  Dept. of Mechani-
cal Engineering, Univ. of Canterbury, Christchurch, New Zealand.

Russell, P.A.  and Hutchings,  A.E.  (1978):  Electron microscopy and X-ray
applications to environmental and occupational health analysis.  Ann Arbor
Science Publishers Inc.,  P.O: Box 1425,  Ann Arbor, Michigan 48106, U.S.A.
                                     109

-------
     Small,  J.A. ,  Heinrich, K.F.J., Newbury, D.E. and Myklebust, R.L. (1979):
     Progress in the development of the peak-to-background method for the quantita-
     tive analysis of single particles with the electron probe.  In: Scanning
     Electron Microscopy/1979/II, (0. Johari, Ed.)-  SEM Inc.,  AMF O'Hare, Chicago,
     Illinois 60666, U.S.A.

     Small,  J.A. ,  Newbury, D.E. and Myklebust, R.L. (1983):  The visibility of
     asbestos fibers in the scanning electron microscope.  Proceedings of the 18th
     Annual Conference of the Microbeam Analysis Society,  (R. Gooley, Ed.), San
     Francisco Press, 148.

     Smith,  J.E.  and Jordan, M.L. (1964):  Mathematical and graphical interpreta-
     tion of the log-normal law for particle size distribution analysis. J. Colloid
     Science 19,  549-559.

     Spurny, K.R., Stober, W., Opelia, H. and Weiss, G.  (1979):  On  the evaluation
     of fibrous particles in remote ambient air.  Science  of the Total Environ-
     ment II, 1-40.

     Spurny, K.R.   (Editor)  (1986):  Physical and chemical  characterization of
     individual airborne particles.  Wiley, New York.                 ,
                                                                      ;
     Wenk, H.R. (Editor)  (1976):  Electron microscopy  in mineralogy.  Springer-
     Verlag, New York.

     Yada, K. (1967):  Study of chrysotile asbestos by a high  resolution  electron
     microscope.  Acta Crystallographica 23, 704-707.

     Yamate, G., Agarwal,  S.C. and Gibbons, R.D. (1984):   Methodology for the
     measurement of  airborne asbestos by electron microscopy.   Draft Report,
     Environmental Monitoring  Systems Laboratory, Office of Research and
     Development, U.S. Environmental Protection Agency,  Research Triangle Park,
     North Carolina  27711.
*U5,COVEmMENrPRJNTINGOmCE:1990 -7H8-1 39/0 0»79
110

-------