&EPA
                       Environmental Monitoring     Office of Pesticides   EPA 600/4-85-049
                       Systems Laboratory      and Toxic Substances  November 1985
                       Research Triangle Park, NC 27711   Washington, DC 20460
              Research and Development & Toxic Substances
Measuring Airborne Asbestos
Following An
Abatement Action

-------
   MEASURING AIRBORNE ASBESTOS FOLLOWING
            AN ABATEMENT ACTION
        Quality Assurance Division
Environmental Monitoring Systems Laboratory
    Office of Research and Development
   U.S. Environmental Protection Agency
    Research Triangle Park, N.C.  27711

                    and

       Exposure Evaluation Division
        Office of Toxic Substances
 Office of Pesticides and Toxic Substances
   U.S. Environmental Protection Agency
          Washington, D.C.  20460
              November 1985

-------
                            ACKNOWLEDGMENTS
     Many individuals contributed their time and effort to the
preparation of this guidance document.   The primary author is Dale L.
Keyes of Environmental Sciences, Tucson, Arizona.   Michael Beard (Office
of Research and Development, Quality Assurance Division) and Joseph
Breen (Office of Toxic Substances, Exposure Evaluation Division) served
as Coauthors and Task Managers, and directed the report preparation
process.

     The material presented here is based in part on the results of a
conference held in March 1984 and sponsored jointly by EPA and the
National Bureau of Standards on post-abatement air monitoring for
asbestos.  The conference brought together government officials,
research scientists, asbestos experts,  abatement contractors, and others
to share ideas and experiences.  A draft of the guidance document
incorporating the conference results together with other information was
prepared and reviewed by several persons acting as official peer
reviewers.  The reviewers are:

     Daniel Baxter, Science Applications, Inc.
     Wolfgang Brandner, EPA Region VII
     Lester Breslow, University of California
     Jay Carter, National Institute of  Occupational Saftey and Health
     Eric Chatfield, Ontario Research Foundation
     William Ewing, Georgia Tech. Research Institute
     Peter Frasca, Electron Microscope  Service Lab
     Richard Lee, U.S. Steel
     Marshall Marcus, Marcus Associates
     Anthony Natale, Duall Incorporated
     Bertram Price, National Economic Research Associates
     Tony Restaino, EPA Region V
     John Small, National Bureau of Standards
     Gerald Spencer, The Survey Group,  Inc.
     Eric Steel, National Bureau of Standards
     Ian Stewart, McCrone Associates

     Their effort in reviewing the first draft and in offering valuable
comments is greatly appreciated.

     Others providing valuable comments on various drafts of the
document include Jean Chesson of Battelle Memorial Institute, William
Nicholson of Mt. Sinai School of Medicine, Roger Wilmoth of
EPA/ORD-Cincinnati, David Mayer of EPA/OPTS-Washington, and Ralph Sposato
of EPA/ORD-RTP.

     This report was prepared under subcontract to the Research Triangle
Institute.  Personnel there are also acknowledged for their assistance
in the technical editing process.

     The material contained within the  document is guidance and does not
constitute a Federal regulation.
                                   11

-------
                             TABLE OF CONTENTS
Acknowledgments 	    ii

Nontechnical Summary of the Guidance	     v

Chapter 1 - Introduction and Purpose	1-1

Chapter 2 - Overview of Technical Guidance	2-1

    2.1   The Process for Releasing the Contractor	2-1
    2.2   Analyzing Air Samples for Asbestos	2-1
    2.3   Air Sampling Procedures	2-4
    2.4   Air Testing Criteria for Determining Work Site
          Cleanliness After Abatement 	   2-5
    2.5   Quality Assurance Practices 	   2-6

Chapter 3 - Sample Analysis 	   3-1

    3.1   The Asbestos Measurement Problem	3-1
    3.2   Analysis by Phase Contrast Microscopy (PCM) 	   3-1
    3.3   Analysis by Transmission Electron Microscopy (TEM)	   3-3
    3.4   Analysis by Scanning Electron Microscopy (SEM)	3-5
    3.5   Comparison of the Three Methods for Post-Abatement
          Air Testing	3-7

Chapter 4 - Air Sampling Procedures	4-1

    4.1   Sampling Equipment	4-1
          4.1.1  Filter Media	4-1
          4.1.2  Filter Cassettes 	   4-1
          4.1.4  Flow-Controlled Pumps and Orifices 	   4-2
    4.2   Sampling Procedures 	   4-2
          4.2.1  Checking Filter Assemblies 	   4-2
          4.2.2  Measuring Airflow	4-2
          4.2.3  Determining Sampling Times and Volumes  	   4-5
          4.2.4  Field Operations	-	4-6
    4.3   Sampling Strategy 	   4-6

Chapter 5 - Air Testing Criteria for Determining Work-Site
          Cleanliness After Abatement 	   5-1

    5.1   The Rationale for the Recommended Release Criteria	5-1
    5.2   Statistical Considerations for Using Release
          Criteria	   5-3
                                   111

-------
                            TABLE OF CONTENTS  (con.)

                                                                      Page


    5.3   TEM Release Criterion	5-4
          5.3.1  Sampling Volume and Time	5-4
          5.3.2  The Number and Location of Samplers	5-4
          5.3.3  Comparing Measured Levels of Airborne Asbestos .  .   .  5-8
          5.3.4  Recommended Actions if the Work Site Fails 	  5-8
    5.4   PCM Release Criterion	5-9
          5.4.1  Sampling Volume and Time	5-9
          5.4.2  The Number and Location of Samplers	5-9
          5.4.3  Comparing Measured Levels of Asbestos to the
                 Lowest Quantifiable Level	5-9
          5.4.4  Recommended Actions if the Work Site Fails	5-9
    5.5   Example Applications of the Two Release Criteria	5-9
          5.5.1  PCM Example	5-11
          5.5.2  TEM Example	5-11

Chapter 6 - Quality Assurance Practices 	  6-1

References	R-l

Appendix A - A Random Number Procedure for Selecting a
             Representative Work-site Sample	A-l
                                  IV

-------
                 NONTECHNICAL SUMMARY OF THE GUIDANCE
     One of the most critical points in an asbestos abatement project is
knowing when the work has been completed, the contractor can be
released, and the building can be reoccupied.  This decision should be
based on two factors: (1) satisfactory performance of the abatement
work, and (2) thorough cleaning of the work site.   As outlined below,
these factors should be evaluated by visually inspecting the work site,
and by measuring the level of airborne asbestos there.  The evaluation
should be conducted by the asbestos program manager or the technical
advisor assigned to monitor the abatement work.

Visual Inspection

    •    Once the contractor has completed the abatement work but
         before any containment barriers have been dismantled, the
         project monitor should thoroughly inspect the work site for
         incomplete abatement and for evidence of dust and debris.

    •    Additional abatement and/or work-site cleaning is needed
         if the work site fails the visual inspection.
Air Testing

    •    Air testing should be conducted after the interior
         plastic barriers have been removed but before the final
         barriers separating the work site from the rest of the
         building have been taken down.

    •    Three methods for measuring airborne asbestos are
         available:  phase contrast microscopy (PCM),  scanning
         electron microscopy (SEM),  and  transmission electron
         microscopy (TEM).   TEM is the best method for measuring
         the types of fibers expected to be present at abatement
         work sites, but PCM is more available and practical in
         many localities.  SEM lacks a standard analytical protocol,
         laboratory testing programs, and standard reference
         materials for judging the accuracy of SEM analyses.

    •    Regardless  of which method is used,  air samples
         should be taken "aggressively".  This means.,  _air blowers
         should be used to  dislodge fibers from surfaces, and
         fans should be used to keep them suspended.

    •    At least five samples should be taken inside  the work
         site,  and,  if TEM  is used for sample analysis,  another
         five outside the work site  should be collected.
         Specified sampling equipment, flow rates, and sampling
         volumes should be  used.

-------
    •    One of two alternative criteria should be used to
         determine if the work site has been adequately cleaned:

              For TEM, the average level of airborne asbestos for the
              samples inside the work site should be no higher than the
              average for the samples outside the work site.

              For PCM, the level of airborne fibers for each of the
              samples inside the work site should be less than PCM's limit
              of reliable quantification (0.01 or fewer fibers per cubic
              centimeter if the minimum recommended volume of air is
              collected).

    •    If the work site fails the air test, it should be
         recleaned and retested.
     The following chapters discuss various technical issues regarding
the air test.  Specifications are provided for air sampling and detailed
guidelines are presented for using either TEM or PCM to analyze the
samples.  The information is designed for asbestos program managers and
technical program advisors.
                                   VI

-------
                               CHAPTER 1

                       INTRODUCTION AND PURPOSE
     Asbestos-containing material (ACM) in buildings is a potential
concern for a growing number of building owners.   EPA estimates that
31,000 schools and 733,000 public and commercial buildings contain
friable (easily crumbled) asbestos (USEPA, 1984a and 1984b).  ACM which
is damaged, disturbed, or deteriorated will release asbestos fibers and
possibly create a health hazard for building occupants.

     Many building owners have undertaken or are considering some form
of abatement (removal, enclosure, encapsulation,  or repair of the ACM).
Although EPA's "Friable Asbestos-Containing Materials in Schools;
Identification and Notification Rule" (40 CFR Part 763) does not require
that schools take corrective action when asbestos is detected, the
parent and employee notification requirements of the Rule have
stimulated the majority of school districts to do so (USEPA, 1984a).
Owners of many other types of buildings also have developed asbestos
control programs.

     EPA has published several guidance documents to assist building
owners in understanding the relevant technical issues, determining if
asbestos is present,  planning a control program if necessary, and
choosing a course of action.  The latest update of the EPA guidance is:
"Guidance for Controlling Asbestos-Containing Materials in Buildings,
1985 Edition," June 1985 (USEPA, 1985).

     Once ACM has been detected in a building and the need for abatement
determined, conducting the abatement action in a safe and thorough
manner is crucial.  Releasing the abatement contractor is the final step
in the abatement process (although a continuing operations and
maintenance program may be necessary until the building is demolished*).
This guidance document addresses the question of what criteria can be
used to judge when the contractor can be released.  It supplements and
extends previous EPA guidance by recommending specific procedures for
using air monitoring in making these judgments.  The material is
presented in technical language, and is thus directed to asbestos
program managers, technical program advisors, and others involved
with asbestos abatement work and air testing.

     The guidance offered here is based in part on the results of a
two-day conference sponsored jointly by EPA and the National Bureau of
Standards (NBS)  held in March 1984.   The evidence presented and the
conclusions reached by the conference participants have been examined in
light of other information reported in the open literature and
government studies.   In this sense,  the guidance  document reflects the
interpretation and judgment of EPA in addition to the collective
experience and knowledge of the conference participants.
    See USEPA 1985 for a description of special O&M programs for
    buildings with ACM.

                                   1-1

-------
     EPA will continue to gather data and conduct research on the
subject of air monitoring for asbestos following an abatement project.
To this end, the experience of asbestos program managers, asbestos
consultants, abatement contractors, and others working on asbestos
control projects could prove to be highly informative.  Any information
on measurements of airborne asbestos in buildings with ACM made by phase
contrast microscopy, scanning electron microscopy, or transmission
electron microscopy may be forwarded to:

          U.S. Environmental Protection Agency
          Environmental Monitoring Systems Laboratory
          Quality Assurance Division (MD-77)
          Re: Asbestos Monitoring Data
          Research Triangle Park, N.C.  27711

     EPA is especially interested in measurements of airborne asbestos
made inside the work site during ACM abatement activities,  and,  once the
abatement is completed, after each work-site cleaning by the abatement
contractor prior to his release.   It would be appreciated if data
forwarded to EPA include basic information such as asbestos fiber
concentration, sample volume,  analytical procedure,  number and type of
asbestos fibers counted,  sample preparation technique (direct or
indirect).
                                   1-2

-------
                               CHAPTER 2

                  OVERVIEW OF THE TECHNICAL GUIDANCE
     The guidance offered here addresses the question of how to
determine when an asbestos abatement work site has been sufficiently
cleaned.  It is set within the larger framework of determining when the
abatement work is completed and when to release the contractor, but it
focusses specifically on the use of air sampling and analysis to
determine work-site cleanliness. Procedures for conducting post-
abatement air sampling are specified and methods for measuring airborne
asbestos and interpreting the results are recommended.

     Following are summaries of each major topic in the guidance
document, preceded by a brief discussion of the process for releasing
the abatement contractor.  The major topics include analyzing air
samples for asbestos, air sampling procedures, criteria for determining
work-site cleanliness, and quality assurance practices.
2.1  THE PROCESS FOR RELEASING THE CONTRACTOR

     The most recent EPA guidance on controlling ACM in buildings (USEPA
1985) describes a two-part process for determining when an abatement
project is complete and the contractor can be released. As illustrated
in Figure 2-1, the two steps are: (1) a visual test to determine if the
ACM has been properly abated and if the work site is free of debris and
dust, and (2) an air test to determine if residual asbestos fibers
generated during abatement have been reduced below a predetermined
level, that is, to determine if the air-test release criterion has been
met.  The asbestos program manager or the person appointed to monitor
the abatement work should be responsible for conducting the visual test
and overseeing the air test.

     The visual test is designed, first, to spot any incomplete
abatement work.  If the ACM is surfacing material, abatement could mean
removal, encapsulation, or enclosure (USEPA, 1985).  If the ACM is pipe
or boiler insulation, abatement could mean removal, patching, or
replacement of the protective jacket (USEPA, 1985).  In any case, the
quality and thoroughness of the work is reviewed.  Deficiencies should be
corrected before proceeding with the next phase of the inspection.

     The second role of the visual inspection is to detect obvious signs
of inadequate work-site cleaning.  The abatement contractor should clean
all plastic barriers at the work site using wet cleaning or HEPA
vacuuming techniques (USEPA, 1985).   The inspector should use damp
cloths and a flashlight to check for debris and dust (USEPA, 1985).

     The air test is designed to detect asbestos fibers which were not
removed by the cleaning procedures.   Before the test is conducted, all
plastic barriers are removed except those covering vents, windows,
doors, and all entries to the work site.  This will allow any fibers
trapped between the plastic and floors, walls, and/or ceilings to become

                                   2-1

-------
                            Abatement Action
                                   is
                              Abatement
                              Complete?
                            Wet-Clean and/or
                             HEPA-Vacuum
                               Work Site
                             Is Work Site
                            Visually Clean?
                            Remove All But
                             Final Plastic
                               Barriers
                      Measure Airborne Asbestos
Wet-Clean and/or
 HEPA-Vacuum
   Work Site
No
Yes
           Is Level
        Low Enough?
          Release Contractor
             Figure 2-1. The process for releasing the contractor.

                               2-2

-------
airborne before testing is begun.  If the air test criterion is met, the
contractor is released.  Otherwise, the work site must be thoroughly
recleaned.
2.2 ANALYZING AIR SAMPLES FOR ASBESTOS

     Three microscopic methods are currently being used to analyze
airborne asbestos: phase contrast microscopy (PCM), scanning electron
microscopy (SEM), and transmission electron microscopy (TEM).   Because
asbestos fibers  are small (especially those found in buildings with ACM)
and difficult to distinguish from other types of fibers, the detection
and accurate identification of asbestos requires sophisticated methods
of analysis.

     TEM is the best method for measuring airborne asbestos.  It can
detect the very  thin fibers (typically down to 0.0025 |Jm diameter) found
in buildings with ACM and in the ambient atmosphere, and it has the
capability of identifying asbestos unambiguously.  In addition, a
standard protocol for TEM analysis has been developed by EPA,  and
standard reference materials for instrument calibration and accuracy
checks are available from the National Bureau of Standards (NBS).  PCM
is less sensitive to thin fibers and less specific for asbestos.  When
used according to the National Institute of Occupational Safety and
Health (NIOSH) protocols, PCM cannot detect fibers smaller than 0.25 (Jm in
diameter, and cannot distinguish asbestos from other types of fibers.  As
a result, PCM results can only be considered an index of airborne asbestos
levels.  However, the method has a well-developed protocol, and NIOSH
operates a testing program for PCM laboratories.  SEM is somewhat more
sensitive and specific than PCM but less so than TEM.  Significantly, no
standard protocol nor NBS standard reference materials are available for
SEM.   As a result, SEM analyses are currently of unknown reliability.
Anyone using SEM for measuring airborne asbestos should require the
analytical laboratory to document the relationship between SEM and either
PCM or TEM results.

     With respect to method availability, cost, and "turnaround time"
(i.e., the time between submission of samples and receipt of results),
PCM is superior on all accounts.  It is by far the most available and
(by a factor of 2-10 depending on the level of analysis) the least
expensive.   The turnaround time is usually less than 6 hours compared
with 6-24 hours for SEM and 2-7 days for TEM.   Note, however,  that
availability,  cost,  and time factors may change significantly in the
future.

     RECOMMENDATIONS

         •    Either TEM or PCM should be used to analyze air
              samples for asbestos fibers.   TEM is the method  of
              choice but PCM is more practical.  Without standard
              protocols and reference materials, SEM results are
              difficult to evaluate.
                                   2-3

-------
              Use TEM according to the updated EPA protocol
              (direct sample preparation, if possible; Level I
              analysis may be sufficient), and PCM according to
              the NIOSH P&CAM 239 (or, alternatively, NIOSH 7400)
              protocol.
2.3  AIR SAMPLING PROCEDURES

     Air sampling is conducted by drawing air through a filter at a
known rate.  Typically, flow-controlled pumps and either cellulose ester
or polycarbonate filters (depending on the method of sample analysis)
are used.  Specific sampling procedures should be followed in order to
assure reliablity of the results.

     In addition to the use of appropriate sampling equipment, the
sampling plan must be carefully designed to account for normal
variability in asbestos levels from location-to-location and over
time.   One of the most important factors which influence the
degree of variability in air measurements is the pattern of air movement.
Under conditions of limited movement,  many fibers will settle out of the
air.   Measurements of airborne asbestos under these conditions are likely
to be lower than if all the fibers were suspended.   Artificial agitation
of the air in a building is one way to keep the fibers suspended.

    RECOMMENDATIONS

         •    Use constant-flow sampling pumps and  the following
              filters:

                   For PCM  analysis:   cellulose ester filters
                   with 0.8-1.2 pm pore size.

                   For TEM  analysis:   polycarbonate filters  with
                   0.4 [jm pore size.

         •     Use the  specified procedures  for testing and
              operating sampling equipment.

         •     Sample at a flow rate of between 2  and  12  liters  per
              minute  (L/min).

         •     Sample aggressively:

                  Use  forced-air  equipment  such  as a  leaf blower
                  to  initially dislodge  fibers  from  surfaces.

                  Use  fans as  specified  to  keep  fibers  suspended
                  during sampling.
                                  2-4

-------
2.4  AIR TESTING CRITERIA FOR DETERMINING WORK-SITE CLEANLINESS
     AFTER ABATEMENT

     Regardless of which method is used to analyze air samples for
asbestos, the results of the analyses can only be used for releasing a
contractor if a criterion is available against which the results can be
compared.  In other words, how low do the measured asbestos levels have
to be in order for the work site to be declared sufficiently clean?

     Since an abatement contractor could not be expected to reduce
asbestos levels below those in the air entering the work site, the level
of airborne asbestos in the ambient air (or in the make-up air if
negative pressure ventilation is used at the work site [USEPA, 1985])
appears to be a reasonable reference.  This is the case for TEM.  PCM,
however, is not sufficiently sensitive to thin fibers nor specific for
asbestos to reliably measure asbestos outside the abatement work site.
A criterion based on the limit of reliable quantification of the
analytical method is more appropriate for PCM.

     The recommended number of TEM samples and the minimum sampling
volume needed to compare measured asbestos levels at the work site
against the reference level should take into account the expected
variability in TEM measurements, how low the reference asbestos level is
likely to be, and the detection limit of the TEM method.  Sampling
requirements for PCM should be at least as rigorous as those for TEM,
considering PCM's low sensitivity to thin fibers and lack of specificity
for asbestos.

    RECOMMENDATIONS

         •    If TEM is used:

                   Collect five samples within and five outside
                   the work site, each of at least 3000 liters.
                   (Use blank filters and duplicate samples for
                   reliability checks as specified.)

                   Analyze the samples and express the results as
                   f/cc (or ng/m3 if an indirect sample
                   preparation is necessary).

                   Compare the averages of the inside and outside
                   levels using the statistical t-test.

                   Release the contractor if the inside level is
                   not statistically higher than the outside
                   level; otherwise, have the entire work site
                   recleaned and retested.
                                   2-5

-------
               If  PCM is  used:

                    Collect  at  least  five  samples per work site or
                    one per  room, whichever  is greater, each of at
                    least 3,000  liters.   (Use blank filter and
                    duplicate samples for  reliability checks as
                    specified.)

                    Analyze  the  samples.

                    Release  the  contractor if none of the samples
                    are above the PCM limit  of reliable
                    quantification  (0.01 f/cc if 3,000 liters are
                    sampled); otherwise, have the entire work site
                    recleaned and retested.
2.5  QUALITY ASSURANCE PRACTICES

     Measuring airborne asbestos is a sophisticated and exacting
process.  Errors may be introduced at any one of the many data
collection and analysis steps.  To guard against this possibility
and to assure reliable results, a formal quality assurance
program should be adopted.

     RECOMMENDATIONS

         •    Be sure that all persons and organizations involved
              in sampling and analysis are trained and/or
              experienced.  Check references and documented
              levels of performance.

         •    Use field and laboratory blanks to check for fiber
              contamination,  coded sample labels to avoid analyst
              bias,  duplicate analyses to confirm analytical
              precision,  and  a second laboratory to spot-check
              the accuracy of results.  Be sure  that all
              equipment setup, operation, and calibration
              procedures  are  followed.

         •    Assign responsibility for security of the  samples
              to  specific  persons  at  each stage  of the analysis.
              Document  each step  in the passage  of the sample
              from the  field  to the laboratory to the  building
              owner.

         •    Check  and document  laboratory  results.   The
              building  owner  should retain all test  results and
              records documenting  the  testing process.   Filters
              should  also  be  saved  in  case additional  analyses
              need to be conducted  in  the future.
                                  2-6

-------
                              CHAPTER 3

                           SAMPLE ANALYSIS
     Three options for analyzing air samples for asbestos were
summarized in Chapter 2: PCM, SEM, and TEM.  The recommended option is
TEM based on its superior technical capabilities.  However, PCM is a
more practical alternative in many localities.  SEM needs development
of a standardized protocol and standard reference materials.  The
rationale for these guidelines is presented more fully in this chapter.
3.1  THE ASBESTOS MEASUREMENT PROBLEM

     Analyzing a sample of air for asbestos is a technically challenging
problem.  Asbestos fibers are extremely small and may number several
million for an average size room when friable ACM is present (Chesson, et
al., 1985a).  Thus, only a small fraction of the asbestos fiber popula-
tion can be observed and counted.  Significant errors can be introduced
when the results of the sample analysis are extrapolated to the entire
room. Furthermore, fibers of asbestos may closely resemble those of
hair, cloth, fibrous glass, paper, and other nonasbestos materials.  As
a result, identifying and counting asbestos fibers requires sophisti-
cated instruments, highly trained technicians, and rigorous quality
assurance practices.
3.2  ANALYSIS BY PHASE CONTRAST MICROSCOPY (PCM)

     The Occupational Safety and Health Administration (OSHA) has
adopted a standard protocol for measuring exposure to airborne asbestos
in the industrial workplace.  This protocol, P&CAM 239, (Leidel et al.,
1979) was developed by the National Institute of Occupational Safety and
Health (NIOSH) and specifies PCM as the measurement method-.  The NIOSH
protocol further specifies that only fibers with a 3:1 aspect ratio and
longer than 5 micrometers ([Jm) in length should be counted.

     The NIOSH protocol involves collecting airborne fibers on a
standard 37-millimeter (mm), 0.8-[Jm pore-size cellulose ester filter.   A
pie-shaped section of filter is then analyzed by dissolving the filter
and counting the fibers with PCM at 400X magnification."  Phase contrast
increases the light contrast between the object and the background, thus
enhancing the microscopist's ability to see fibers.  Normally, 100
    NIOSH has published a revised protocol—NIOSH 7400 (NIOSH, 1984).
    OSHA is now reviewing this revised protocol, but has not adopted it
    yet.
                                   3-1

-------
 microscopic fields or 100 fibers are counted, whichever occurs first.*

      PCM, as employed in the NIOSH protocols, has two serious limitations
 for measuring airborne asbestos.  First,  PCM can not distinguish asbestos
 from nonasbestos fibers; all elongated particles with the required length
 and aspect ratio are counted.   PCM-measured fibers thus can only serve as
 an index of asbestos fibers.  Second,  only particles larger than about
 0.25 l-im in diameter can be detected owing to inherent limits of resolution
 of PCM, and only particles longer than 5  [Jm are counted due to the
 counting protocol.

      These are not serious limitations for the use of PCM in asbestos
 workplace settings where asbestos fibers  are a significant fraction of
 all airborne fibers.   Moreover,  variation over time in levels of
 PCM-measured fibers and asbestos fibers appears to be correlated in the
 asbestos workplace; that is, the higher the level of PCM fibers, the
 higher the level of asbestos.   These relationships are borne out by
 studies of the health of workers in asbestos industries in which levels
 of PCM-measured fibers serve as  the index of exposure to asbestos (NEC,
 1984).

      With one  exception,  conditions in buildings with friable ACM are
 believed to  be quite  different.   Although evidence is limited,  asbestos
 fibers  appear  to be smaller  in size (fewer fiber bundles)  and a smaller
 fraction of  all airborne fibers  than those in asbestos industry settings
 (Chatfield,1983).   A  recent  study of schools with friable  ACM found very
 low correlations of fiber levels measured by PCM compared  with asbestos
 levels  measured by  SEM and TEM (Chesson et al.,  1985a).

      The exception  to  this general  rule may be fiber levels  generated
 during  asbestos  abatement activities.   Levels  of both PCM  and asbestos
 fibers  are likely to be  elevated during abatement,  especially during
 removal  of friable  ACM.   In  this  sense, the abatement work site  may
 approximate  conditions  in the  asbestos  industry workplace.   Thus,  OSHA
 requires  measurements  of  airborne  fibers  by PCM during abatement
 projects  as  an  indication of asbestos  exposure.

     The  justification  for using  PCM-measured  fibers  as  the  basis  for
 determining  when the abatement worksite has  been  sufficiently cleaned,
 that is,  as  the  release  criterion,  follows  from  the  above  argument.  If
 levels of both PCM  and asbestos  fibers  are  elevated  during abatement
 activities,  then removing PCM  fibers should  remove asbestos  fibers  as
well.  In other  words, work-site  cleaning practices which  reduce  levels
 of airborne  cellulose, hair, and  other  large fibers  detected  by  PCM
 should likewise  reduce levels of  residual asbestos fibers.   However,
 this rationale rests on logical deduction; no  simultaneous measurements
of PCM and SEM or TEM levels during an abatement action and  following
work site cleaning operations have been made to test  the rationale.
   A minimum of 10 fibers needs to be counted for reliable quantifi-
   cation (Leidel et al., 1979).  Counting more than 100 fibers
   or 100 fields would be unnecessarily time-consuming and would add
   little to the reliability of the results.

                                   3-2

-------
     Of the three methods for measuring airborne asbestos, PCM is the
least expensive  (about $25-$50 per sample) and the most readily
available.  In addition, results of PCM analysis can usually be
communicated to  the building owner in less than 6 hours.  Finally, the
NIOSH protocol has been extensively tested and an active laboratory
evaluation program, the Proficiency Analytical Testing (PAT) Program, is
maintained by NIOSH.*
3.3  ANALYSIS BY TRANSMISSION ELECTRON MICROSCOPY (TEM)

     The limit of a microscope's ability to detect objects is related to
the wavelength of the source of "illumination".  Since electrons have a
much shorter wavelength than does light, the electron microscope is
inherently superior to the optical microscope for detecting small fibers
typical of asbestos fiber populations found in buildings with ACM.

     Of the two types of electron microscopy used for measuring airborne
asbestos, TEM is considered the method of choice (Chatfield, 1983; Steel
and Small, 1985).  Following the EPA provisional methodology for TEM
analysis (Samudra et al., 1978), fibers are collected on a 0.4 [Jm pore
size polycarbonate filter (or on a 0.45 (Jm pore size cellulose ester
filter if significant levels of contaminating organic materials are
present in the air).  Sample preparation involves either (1) direct
transfer of collected fibers from the polycarbonate filter to an
electron microscope (EM) grid after the filter is first carbon-coated,
or (2) an indirect transfer whereby a section of the cellulose ester
filter is ashed, the asbestos fibers are sonicated in water and
refiltered on a polycarbonate filter, and then carbon-coated and
transferred to the EM grid.**  Direct transfer is preferred since it
does not cause fiber breakage.  The mounted fibers are then examined at
20,OOOX magnification, identified as asbestos, measured, and counted.
The mass of each fiber may also be estimated if estimates of mass
concentration are desired.   No more than 100 fibers or 10 grid openings
need to be observed.

     TEM is the method of choice for analyzing asbestos based on its
sensitivity to the smallest fibers and on its specificity for asbestos.
Since the sample of fibers  is mounted on an extremely thin substrate on
the EM grid,  electrons can  pass through the substrate, be diffracted by
the fibers and other materials, and be refocussed into an image on a
fluorescent screen, all without substantial back-scatter of electrons.
This allows high electron beam voltage (approximately 100 kilovolts) and
high magnification of the specimen (up to 100,OOOX).   Extremely thin
asbestos fibers (typically  0.0025 [Jm in diameter) can be detected.
    The results of the PAT program should be used in selecting a
    laboratory for PCM analysis. Call (513) 841-4357 for a copy of the
    latest evaluation results.

    A direct transfer technique for cellulose ester filters has also been
    reported (Burdett and Rood, 1983).
                                   3-3

-------
      TEM can be used to indicate the likely presence of asbestos in a
 population of fibers based on fiber shape and configuration alone.
 However, in order to confirm the identity of the fibers, chemical and
 crystal analysis of individual fibers is needed.  The relevant
 analytical techniques are known as energy dispersive X-ray spectrometry
 (EDXA or EDS) and selective area electron diffraction (SAED).   In EDXA,
 X-rays emitted from interactions between the electron beam and the
 fibers are analyzed, and in SAED, the electron diffraction patterns
 created by the same interactions are analyzed.

      A TEM instrument outfitted with EDXA and SAED capabilities is
 sometimes called an analytical electron microscope.   The high  electron
 beam voltage characteristic of TEM combined with the thinness  of the
 fiber substrate on the EM grid allows EDXA and SAED to be performed on
 single fibers.  Although each fiber observed is not always subjected to
 EDXA and SAED analysis,  the preliminary identification of asbestos-like
 fibers combined with chemical and crystal analysis of a representative
 subset of fibers allows  the fiber population to be characterized with a
 high level of confidence.   In addition,  SAED can be  performed  visually
 (by quickly observing the diffraction pattern on the fluorescent screen)
 or quantitatively (by photographing the  diffraction  pattern at an angle
 and measuring the photograph).   The latter is a definitive means of
 identifying asbestos.

      The extreme sensitivity of TEM does  make the task of detecting and
 assessing thick fibers  (larger than about 1.0 |Jm) and fiber clusters  and
 bundles  more  difficult.   Because thin fibers greatly outnumber thick
 ones in  air samples  from buildings  with  friable ACM  (Chatfield,  1983;
 Chesson  et  al.,  1985), counting fibers on the EM grid may stop before
 any large fibers  are observed.   Although  the failure to  observe  thick
 fibers will not  significantly affect  fiber counts, it will  bias  downward
 the estimation  of  fiber  mass,  since a  single large fiber may equal  the
 mass  of  several  thousand small  fibers.  Likewise,  fiber  clusters  and
 bundles  may be  infrequently found.  However,  the clusters  and  bundles
 are so difficult  to  accurately  measure and are  so large  in  mass  compared
 to  individual fibers  that the revised EPA protocol specifies that the
 presence  of clusters  and bundles  be noted but not included  in  fiber
 counts or estimates  of mass  concentrations  (Yamate,  1984).

      Scanning transmission  electron microscopy  (STEM)  has been employed
 by  some  laboratories  to  aid  in  identifying  large  fibers.  STEM is
performed by scanning the field  of view at  a  lower magnification
 (typically  1000X).  Not  all  transmission  electron microscopes  have
 scanning capabilities.

     Given  the complexity and sophistication  of  the TEM  analysis  for
asbestos, the need for highly skilled microscopists and  detailed
protocols is apparent.   Early attempts to  compare analyses  of  the same
air sample by several laboratories revealed  that  the analytical  results
varied by several orders of magnitude (USEPA, 1977a).  Since then,
efforts to standardize sample preparation and analysis protocols and to
develop strict quality assurance practices have greatly  improved the
reliability of asbestos  analyses by TEM.  A  recent study by the NBS
revealed that TEM microscopists in the study had a greater  than 90

                                   3-4

-------
percent chance of identifying chrysotile fibers longer than 1 pm (Steel
and Small, 1985).  However, the authors also noted that instrument
characteristics  (especially, the mechanical stage, image quality, and
electron diffraction capabilities) can be a significant source of error.
On the other hand, the availability of NBS standard reference materials
for asbestos (Small et al., 1985) and the use of special counting
procedures like  the one described by the NBS authors (Steel and Small,
1985) can be used to evaluate the reliability and accuracy of TEM
results.

     The major disadvantages to using TEM for post-abatement clearance
monitoring are the cost and time for analysis.  Partly because few
laboratories currently offer TEM services, costs for analysis may be $500
or higher per sample, and the time until results are received may be
several days.  To reduce the cost and turnaround time, EPA has proposed
three levels of TEM analysis (Yamate, 1984):

         Level I - Identification of asbestos fibers is based on fiber
                   morphology and the observed SAED pattern on the
                   fluorescent screen.

        Level II - Analysis of the chemical composition of each fiber by
                   EDXA is added to the Level I procedures.

       Level III - Quantitative analysis of SAED patterns from a few
                   representative fibers is added to Level II procedures.

     Where asbestos is known to be present, fibers which appear to be
asbestos-like by shape and by qualitative analysis of crystal structure
(i.e., visual SAED) can reasonably be assumed to be asbestos.  This is
the case for abatement work sites in buildings with ACM.  As a result,
Level I analysis should be sufficient for post-abatement testing
purposes. However, where definitive confirmation of airborne asbestos is
needed, for legal or other purposes, a Level II or III analysis will be
necessary. Thus, if only Level I analysis is employed, EM grids for all
samples should be archived for future Level II or III analysis as may be
needed.
3.4  ANALYSIS BY SCANNING ELECTRON MICROSCOPY (SEM)

     As an electron microscopic method, SEM holds promise for greater
sensitivity to thin fibers and better specificity for asbestos as
compared with light microscopy.  Technically, however, it currently
falls short of TEM's capabilities.  SEM differs from TEM in that the
fiber substrate mounted on the EM grid is considerably thicker.  As a
result, electrons bombarding the specimen are scattered and reflected
rather than being transmitted.  The thick substrate also reflects and
scatters electrons which are detected as "noise" by the microscope.  As
a result, the object being viewed must be larger than a TEM-observed
object in order to be seen. In terms of fiber dimensions, the limit of
resolution obtained under typical conditions is a fiber diameter of
0.20 pm.
                                   3-5

-------
      SEM is also less powerful than TEM in its  ability to  distinguish
 asbestos from other types of fibers.   SAED is not  feasible with  SEM  due
 to the thick substrate and the signal noise problem  noted  above.
 Chemical analysis with EDXA is possible (for fibers  with a diameter  of
 at least 0.20 urn),  but EDXA alone does not provide definitive  evidence
 for asbestos.  (Some nonasbestos  fibrous  materials have similar  chemical
 compositions.)  However,  in a setting such as an asbestos  abatement  work
 site where airborne asbestos is likely to be present,  morphological
 identification of asbestos-like fibers by SEM combined with detection of
 asbestos-like chemical compositions for a few of these fibers  would  be
 strong support for the presence of asbestos.

      Without doubt,  SEM can be superior to PCM  for indicating  the
 presence of airborne asbestos.   In addition, the scanning  feature of SEM
 used at a magnification of 1,000-2,OOOX provides a useful  means  of
 rapidly observing fields  of view  and  locating large  fibers, clusters,
 and bundles.

      Unlike PCM and  TEM,  no standardized  protocol  for  sample preparation
 and analysis  using  SEM is  currently available.  Although samples are
 usually collected on 0.4-0.8  pm pore  size polycarbonate filters,
 cellulose  ester filters have  also  been employed.*  Likewise,  most
 laboratories  use  a  relatively  simple  protocol for  sample preparation
 (generally,  direct carbon  coating  of  the  filter),  but  the  specific
 features  of the protocol  differ significantly among  laboratories.  The
 same  is  true  for  instrument specifications  (e.g.,  raster scan  rate,
 magnification,  electron beam  strength)  and  fiber identification and
 counting procedures.**  Without standardized protocols, it is not
 possible to  characterize  analytical accuracy and reliability of SEM
 results.   It  is difficult  to know  how much  confidence  can be  placed  in
 the results of  an SEM analysis  until  (1)  a  standardized protocol is
 developed,  evaluated, and  adopted,  (2) NBS  reference materials are made
 available  for calibrating  instruments and procedures, and  (3) a
 laboratory  evaluation program is initiated.  EPA and NBS are  both
 initiating programs which address  these deficiencies.

     Based on evidence presented at the NBS/EPA conference  (NBS/EPA,
 1984), SEM service for asbestos analysis appears to be more available
than TEM but less available than PCM.  In addition, both the  cost and
time of analysis appear to be intermediate between  PCM and  TEM.
    Polycarbonate filters are preferred since problems with signal noise
    are fewer than with cellulose ester filters.

    An example of an SEM protocol developed by Verein Deutscher Ingenieur
    appears in Spurny (1985).
                                   3-6

-------
3.5  COMPARISON OF THE THREE METHODS FOR POST-ABATEMENT TESTING

     As summarized in Table 3-1, PCM, TEM, and SEM offer clear if not
easy choices for measuring airborne asbestos following an abatement
project.  PCM is currently the most widely used technique, due to its
history of use for meeting OSHA workplace exposure standards and to its
availability.  Since OSHA monitoring is required for many asbestos
removal projects, it seems only natural to many building owners to
specify a PCM-based criterion for determining project completion.  Not
only is PCM the most popular method, the PCM analytical protocol and the
laboratories offering PCM service are the best characterized.  Finally,
PCM's low cost and short turnaround time make it an attractive choice
when contractors are waiting to complete an abatement project.  But PCM
measurements are, at best, only rough indicators of asbestos
contamination following abatement.

     TEM is distinguished from PCM on all characteristics.  It is more
sensitive to thin fibers and more specific for asbestos, on the one
hand, and less available, more costly, and more time consumptive on the
other.  With respect to method characterization and development, TEM has
shown substantial improvement during the last few years.  The
availability of qualified laboratories offering TEM service also should
improve as the demand for service increases.  This should put downward
pressure on future costs and turnaround times.  In the short term,
however, users of TEM for determining the completion of abatement
projects will be faced with relatively high costs and long delays in
obtaining results of analyses.  Specifying EPA Level I analysis may
mitigate these problems to some extent, as suggested by the low end of
the cost and time estimates for TEM in Table 3-1.  (A portion of the
range of cost and time estimates for TEM reflects direct compared with
indirect methods of sample preparation.)

     SEM appears to lie between PCM and TEM on most characteristics: it
is potentially more sensitive to thin fibers and more specific for
asbestos than PCM but less so than TEM; it is more readily available
(and popular) than TEM but less so than PCM; and estimates for both cost
and time of analysis are higher than for PCM but lower than for TEM.
SEM's greatest handicap is inadequate method characterization, including
the lack of a standardized protocol for sample preparation and analysis.
Efforts by EPA and NBS to evaluate the utility of SEM, to provide
standard reference materials, and to develop a laboratojy testing
program should improve SEM characterization.

     It is important to note that the estimates in Table 3-1 assume
up-to-date instruments, skilled analysts,  good operating conditions, and
strict quality assurance practices.   Where these assumptions do not
hold, the estimates of method sensitivity and specificity may not apply.
In addition,  both TEM and SEM can be conducted with various degrees of
sophistication.   The three EPA levels of analysis reflect this for TEM,
as does the range of cost and time estimates for SEM.   For example, SEM
could be conducted with reduced sensitivity in order to detect
                                   3-7

-------
    TABLE 3-1.  COMPARISON OF METHODS FOR MEASURING AIRBORNE ASBESTOS
                        PCM
                            SEM
                          TEM
Standard
Methods
Quality
Assurance

NIOSH P&CAM 239
Method"
Proficiency
Analytical Test-
ing Program; no
NBS reference
materials.
No standard
method.
No lab testing,
or NBS reference
materials .

EPA provisional
method & update**
Limited lab test-
ing, NBS refer-
ence materials
available.

 Cost

 Availability

 Time
 Requirements
 Sensitivity
 (Thinnest
 Fiber Visible)

 Specificity
$25-50

Most available.

1 hr preparation
& analysis, <6
hrs. turnaround
0.15 pm at best;
0.25 pm typical.
Not specific for
asbestos.
Collection
Filters
0.8-1.2 pm
cellulose ester.
$50-300

Less available.

4 hr preparation
& analysis, 6-24
hrs. turnaround
0.05 pm at best;
0.20 pm typical.
More specific
than PCM but not
definitive for
asbestos (SEM
with EDXA)

0.4-0.8 pm poly-
carbonate best,
cellulose ester
also used.
$200-600

Least available.

4-24 hr prep-
aration & anal-
ysis, 2-7 days
turnaround

0.0002 pm at best;
0.0025 pm typical.
Definitive for
asbestos (Level
III — TEM with
EDXA & SAED)
0.4 pm polycarbon-
ate, or 0.45 pm
cellulose ester
if organic con-
taminants present.
* Leidel et al., 1979.  NIOSH 7400  (NIOSH, 1984) is an alternative.

** Samudra et al.,  1978; Yamate, 1984.

Source: Based on information from the EPA/NBS conference on post-
abatement air monitoring (NBS/EPA,  1985), the open literature, and
government reports, and on peer review comments.
                                   3-8

-------
PCM-equivalent fibers (i.e., fibers with diameters greater than 0.25 jJra
and at least 5 (Jm in length) and without EDXA analysis to distinguish
nonasbestos fibers from those that are asbestos-like.  Such an analysis
may cost as little as $50 and take only a few hours,  but obviously would
provide no more information than a PCM analysis.   A more sophisticated
SEM analysis, which counts all fibers 0.020 |Jm or more in diameter and
which uses EDXA on some fibers, would likely approach the high end of
the cost and time range ($300) and requires more  than one day for
analysis.

     Whichever method of measuring airborne asbestos  is chosen, the
exact specifications of the analysis (sensitivity, specificity, cost,
and turnaround time) should be clearly communicated to the laboratory.
Limits of fiber resolution should then be verified by the laboratory.

     To summarize:

     (1)  If PCM is selected as the method of analysis, the results
          should be reliable, the cost modest, and the turnaround time
          rapid.  Strictly speaking, however, the results can only
          indicate success in removing large airborne fibers of both
          asbestos and other composition—the relationship between PCM
          fibers and asbestos fibers in this situation rests solely on
          deduction.

     (2)  If TEM is selected as the method of analysis, the results
          should be reliable and should indicate  the level of all
          asbestos fibers; but the cost will be high and the turnaround
          slow.  (Remember, the specificity of a  Level II or III
          analysis is higher than that of Level I, but cost and
          turnaround time are also higher.)
     (3)  If SEM is selected as the method of analysis, the results
          should indicate the level of most airborne asbestos fibers
          (although some nonasbestos fibers may also be counted and the
          smallest fibers will not be counted), and the cost and
          turnaround time will be between those for PCM and TEM; but
          the results will not necessarily be reliable.
                                   3-9

-------
                               CHAPTER 4

                        AIR SAMPLING PROCEDURES
     This chapter describes recommended air sampling equipment and
procedures for use in post-abatement clearance monitoring.   As
summarized in Chapter 2, the recommended approach is "aggressive
sampling."  Specifications for aggressive sampling are provided,
including the characteristics of forced-air equipment.  The volume of
sampled air needed to evaluate alternative release criteria is also
discussed.  The number of samplers and their location within the work
site are noted here and discussed more fully in Chapter 5.
4.1  SAMPLING EQUIPMENT

     In the sampling process, air is drawn through a filter at a known
rate by a flow-controlled pump.  The sampler components are described
below.

4.1.1  Filter Media

     The selection of a filter for sample collection will depend on
which method is used to analyze the sample for asbestos.  When PCM is
employed, the filter should be cellulose ester with a pore size of
0.8-1.2 [jm, as specified in the P&CAM 239 protocol (Leidel et al.,
1979)."  When either TEM or SEM is used, the first choice in filter
media is polycarbonate with a pore size of 0.4 |Jm.  When substantial
quantities of airborne organic materials are present, a 0.45 pm
cellulose ester filter is recommended together with the indirect method
of sample preparation for TEM (Yamate, 1984).  Cellulose ester filters
have also been used for SEM analysis, although they tend to cause
additional background "noise" problems (see Chapter 3).  Select the
proper filter type and check each filter lot for low background asbestos
counts prior to sampling.  (See Chapter 6 for additional information on
quality assurance.)

4.1.2  Filter Cassettes

     Commercial filters are sold as filter and cassette combinations.
If cassettes are loaded with filters outside the manufa-eturers'
facilities, loading should take place only under clean laboratory
conditions (i.e., either in a clean room or on a class 100 clean bench
with a laminar-flow hood).   In order to obtain a uniform distribution of
collected particulates across the surface of the collecting filter, a
5.0 pro pore-size cellulose ester backing filter should be placed behind
    The new NIOSH Method 7400 prescribes the use of 25-mm diameter
    rather than the more common 37-mm diameter filters (NIOSH, 1984).
    Since the area of a 25-mm filter is 45 percent less than that of a
    37-mm filter, 45 percent less air needs to be sampled to achieve the
    same fiber density (f/mm2) on the filter.

                                   4-1

-------
 the collecting filter.  This is followed by the cellulose support pad
 and the cassette base (see Figure 4-1).  The filters should be sealed
 evenly without wrinkles.

      The movement of air through the filter may cause a significant
 buildup of static charge on the cassettes.  The static charge, in turn,
 is likely to affect the distribution of fibers on the filter and may
 cause fibers to collect on the cassette walls rather than on the filter.
 To guard against static buildup, the European Reference Method for
 Asbestos Measurement published by the Occupational Medicine and Health
 Laboratory recommends that a metal cowl or electrically conductive
 cassette be used in conjunction with the sampling train (OMHL, 1984).

 4.1.3  Flow-Controlled Pumps and Orifices

      Air samples should be collected using constant flow sampling pumps.
 A typical pump and sampling train is shown in Figure 4.2.   Critical
 orifices are used because they are convenient and accurate in
 controlling flow.   However, slight changes in size and shape of the
 orifice due to wear or accumulation of particles  will change the orifice
 characteristics.   Therefore,  orifices should be monitored  before,
 during, and after use in sampling.   Pump and filter combinations must  be
 matched to flow rate requirements since some filters produce high back
 pressure which limits pump capacity.   Double orifice pumps can be  used
 for collecting samples on two types  of filters  simultaneously.
 4.2  SAMPLING PROCEDURES

 4.2.1  Checking  Filter Assemblies

     The  cassette  assembly and  sampling  train  should be  checked for
 leaks before use.  This  can be  accomplished by connecting the train to a
 metered vacuum reservoir.  The  apparatus must  pass a leak check of less
 than 4 percent of  the average sampling rate at a vacuum  equal to or
 greater than the maximum value  reached during  the sample run (USEPA,
 1977c).   Alternatively,  a rotameter  can  be used to check for leaks (see
 Section 4.2.2).

 4.2.2  Measuring Airflow

     In most applications, a high quality rotameter with arbitrary unit
 graduations  is sufficient to monitor the sample flow rate through the
 sampling  apparatus.  When rotameters are not used, flow measurement
 devices such as  mass flow meters and dry gas meters may be employed.
 The  flow measuring device should be  inserted behind (downstream of) the
 filter and the pump assembly.  All measurement  equipment should be
 capable of ranges at least 1.5 times and readable to at least 0.01 of
 the desired  flow rate.   All flow measurement equipment should be
 calibrated against standards of higher accuracy before and after
 sampling.   Specific calibration procedures for dry gas meters,  mass flow
meters,  and  rotameters are found in EPA,  1977b.
                                   4-2

-------

                                       — Collection Filter






                                       	  Backing Filter






                                       ~-^Pad
Figure 4-1.   Filter and  cassette assembly.
                     4-3

-------
               Orifice
               Detail
Brass Disk 0.209" Dia.
1/16" Thick Center
Drilled No. 68 & Soft
Soldered in Place    Drill No. 4
                   (0.2090")
                                                   8 Turns of
                                                   1/4" Copper
                                                   Tubing Wound
                                                   4" Diameter
Gelman Filter Holders
Model 4202 47 mm Open
Faced Magnetic
                                                                            Clamp, Medium
                                                                            Utility 3-Finger
                                                                            Jaw Vinylized
                                                                           36" Long Rod
 3 foot 1/4" x 3/16"
 Rubber Vacuum Tubing
  Swagelok B-2-MHC4T
  Hose Connector to Male
  Pipe 1/8" Male Pipe to
  1/4" I.D. Tubing
Thomas Industries Inc.
Pump Model 107CA18
                                           — Tube Fitting, Male Elbow 90°
                                             1/8" Male Pipe Threaded to
                                             1/4" Tube
                                             Swagelok B-200-2-4

Elapsed Time
Indicator
WW Grainger
6x136

	 "^•"•n

7 Day
Programmable Timer
Grainger 2E214
Power Cord
                               Figure  4-2.   Typical  pump.
                                               4-4

-------
4.2.3  Determining Sampling Times and Volumes

     Regardless of which method is used for sample analysis, a minimum
of about 3,000 liters (L) of air should be filtered at a rate of 2-12
L/min.  The total sampling volume needed will depend on the criterion
used to determine abatement project completion (see Chapter 5).  For
example, if the criterion is 0.01 PCM fibers per cubic centimeter
(f/cc), then the total volume required to detect fiber levels this low
is 2,850 liters:

          - (10 f/100 fields)    (855 mm2/filter)    (1 L)    _  „ . T
          ~    (0.01 f/cc)    X (0.003 mnrVfield) X (lOOOcc)    Z"  L
where:   (a) 10 f/100 fields is the minimum fiber loading on the filter
         required for reliable quantification by the P&CAM 239 Method
         (Leidel et al., 1979).*

         (b) 0.01 f/cc is the release criterion.

         (c) 855 mm2/filter and 0.003 mm2/field are, respectively, the
         area of a 37-mm diameter filter and the area of each viewing
         field.  (Some optical microscopes have viewing fields as large
         as 0.006 mm2.  Larger fields of view will improve [decrease]
         the limit of reliable quantification for a given sampling
         volume. )

     At 2-12 L/min, collecting 2,850 liters would require sampling for
about 24 hours (2 L/min) or about 4 hours (12 L/min).

     Likewise for TEM, a volume of 3,054 liters would be needed to
detect asbestos levels down to, for example, 0.005 asbestos f/cc:**

         _ (1 f/10 gd.sq.)     (855 mm2/filter)        (1 L)         ,
       V ~  (0.005 f/cc)    X (0.0056 mm* /gd.sq.)  X (lOOOcc)    JU°4 L

where:   (a) 1 f/10 gd.sq. is the minimum fiber loading (per 10 grid
         squares) for fiber detection.

         (b) 0.01 f/cc is the release criterion.

         (c) 855 mm2/filter and 0.0056 mm2/gd.sq. are, respectively, the
         area of a 37-mm filter and the area of one grid square, 75  |Jm
         on a side in a 200 mesh EM grid.  (Grid squares may vary in
         size from 0.0056 to 0.0081 mm2.  Larger squares will improve
         [lower] the detection limit for the same sampling volume.)

     These examples point out the need to estimate sampling volumes  on
the basis of how many fibers need to be collected for reliable measurement
    The NIOSH 7400 Method lowers the minimum fiber loading to 5 f/100
    fields (NIOSH, 1984).

    If an indirect sample transfer technique is used, additional dilution
    terms must be added to the equation.  This will increase the minimum
    sampling volume needed to detect one fiber.
                                   4-5

-------
 at the air level selected as the release criterion.  The number of fibers
 is based on the limit of reliable quantification for PCM and on the
 theoretical detection limit for TEM.  The former is the preferred
 measure of the minimum required fiber loading on the filters.   Unfortu-
 nately, data on which to base an estimate of the limit of reliable
 quantification for TEM are not available.

 4.2.4  Field Operations

      Samplers should be located in a room or area so that they are not
 unduely influenced by the configuration of the space or by each other.
 For example,  samplers should not be placed in room corners,  under
 shelves,  or in other locations where airflow is restricted.

      Once the sampling equipment is in place, the location,  time,  filter
 number, pump  number, and other pertinent information are recorded.  (See
 Chapter 6 for a detailed discussion of quality assurance requirements.)
 The end cap is removed from the front of the cassette and the  pump is
 started.   Normally,  the cassette face is oriented in a downward position
 to prevent contamination of the filter by large particles falling from
 the ceiling.   In a  clean work site,  however, the ceilings should be free
 of any large  particles.   Placing the filter cassette in an upward
 position  is thus feasible.   This has the added advantage of  preventing
 the collected fibers from becoming  dislodged from the filter when the
 vacuum is  released.   After the pump  is started,  the flow rate  is
 recorded  and  verified after 15-30 minutes of operation to guard against
 leaks  or  constrictions in the  sampling train.   Timers are frequently
 used when  the  sampling time exceeds  a few hours.   When the pump needs to
 be shut off for any  reason,  the cassette should  first be oriented  in an
 upright position (if sampling  has been conducted with the filter facing
 downward)  to preclude the chance of  collected fibers falling from  the
 filter when the vacuum is released.

     When  the  requisite  sampling volume  has  been reached,  the  time,
 intermediate  flow rate  checks,  and the final flow rate  are recorded.
 Samples on  cellulose  ester  filters are usually mailed  to  the laboratory
 for analysis without  further treatment.   The polycarbonate filters
 should  be  treated with  special  care.   They  should be  hand carried  to  the
 laboratory  if possible.   To guard against fiber  loss  from polycarbonate
 filters, keep the filters in a  horizontal position  with the  collection
 surface up.
4.3 SAMPLING STRATEGY

     Guidelines for the number of sampling locations needed to evaluate
the various release criteria are described in Chapter 5.  Regardless of
which criteria is selected, air sampling to evaluate compliance should
be conducted "aggressively", that is, after any settled fibers have been
resuspended and while fans are operated to keep them airborne.

     Aggressive sampling should begin after the work site has been
wet-cleaned and HEPA-vacuumed (see Chapter 2) and all plastic except the
                                   4-6

-------
final containment barrier removed.   (That is,  plastic should remain on
windows, doors, and air vents.)  Any negative  filtration units used
during abatement should remain on.   The samplers are located as
indicated by the sampling design.  Before any  sampling begins, floor,
ceilings, and walls are swept with the exhaust from a high-speed air
circulating device such as a 1-horsepower,  electrically operated leaf
blower.   This activity should continue until the exhaust has been swept
across all surfaces, or for at least 5 minutes per 1,000 square feet of
floor area.  Stationary fans (20-inch minimum  in size) on 2-meter high
stands are then placed at central locations so as to induce area-wide
circulation.  In addition, they are directed at the ceiling and operated
at low speeds so as to avoid high rates of air flow in the vicinity of
the sampling equipment. One fan should be used for each 10,000 cubic
feet of space.  The fan(s) should be left on for the duration of
sampling. Aggressive sampling greatly increases the probability that
fibers,  if present, will be dislodged and distributed in a relatively
homogeneous manner throughout the air space.
                                   4-7

-------
                               CHAPTER 5

            AIR TESTING CRITERIA FOR DETERMINING WORK-SITE
                      CLEANLINESS AFTER ABATEMENT
     The overall process for determining when an abatement contractor
can be released was outlined in Chapter 2 and is discussed in more
detail in the updated EPA guidance document on controlling asbestos in
buildings (USEPA, 1985).  A two-phased approach is recommended: visual
inspection followed by air testing.  Air testing is designed to
determine whether the work-site has been cleaned adequately.  The measure
of work-site cleanliness will depend on which method is chosen for
measuring asbestos fibers.  The recommended criteria are:

    •    If TEM is used, the average of measured work-site levels should
         be statistically no larger than the average of measured levels
         outside the work site.

    •    If PCM is used, all measured work-site levels should be no
         higher than 0.01 f/cc (or less, if a lower level of reliable
         quantification is used).

     The basis for these recommendations is presented in this chapter.
The discussion includes the rationale for each of the two release
criteria, and the statistical basis for applying them.

     Recommendations for the required sampling volumes and the number
and location of air samplers follow from the choice of release criteria.
As summarized in Chapter 2, sampling design recommendations are:

    •    If TEM is used, at least five samples inside and five samples
         outside each homogeneous work site should be collected.  Sampling
         volume should be at least 3,000 liters.

    •    If PCM is used, at least five samples inside each homogeneous
         work site should be collected.  Sampling volume should be at
         least 3,000 liters.
5.1  THE RATIONALE FOR THE RECOMMENDED RELEASE CRITERIA_

     The objective of measuring airborne asbestos following an abatement
project is to assure that asbestos fibers released during the abatement
action have been reduced to an acceptable level.  Unfortunately, no
safe level of exposure to asbestos exists.  Any exposure to the fibers
carries some risk.  The point is to reduce levels to the lowest level
technically possible.  Hence, the lowest airborne asbestos level that can
be attained within practical limitations depends upon the technological
feasibility in the analytical methodologies.
                                   5-1

-------
      Since outdoor levels are typically very low compared with levels  in
 buildings with ACM,  outdoor levels would appear to  be the next best
 basis of comparison.   However,  where negative air ventilation systems
 are used during abatement,  the  make-up or "background" air comes  from
 other parts of the building rather than directly from outdoors.   In  this
 situation, the more  appropriate reference point is  the level  of asbestos
 in air outside the work site but inside the building.   Thus,  the
 recommended release  criterion,  if TEM is used,  involves comparisons
 between measurements  of asbestos inside the work site with those
 outside,  either outdoors or immediately outside the work site.

      The  use of PCM  requires additional considerations.   Since  PCM used
 according to the NIOSH protocol detects many types  of fibers  other than
 asbestos  and is not  sensitive enough to detect  the  very small  fibers
 typical of asbestos  in the  ambient environment,  outdoor fiber measurements
 using PCM provide little if any information on  ambient asbestos.  The  same
 is  true for airborne  asbestos inside buildings  other  than at  the
 abatement work site.

      As an alternative to inside-outside comparisons,  the recommended
 release criterion for use with  PCM involves  comparing  work-site asbestos
 levels  with the PCM limit of reliable  quantification.   Since the  lowest
 level  of  airborne fibers quantifiable  with  PCM  depends  on the volume of
 air  collected,  the criterion could,  in concept,  specify  any level of
 fibers.*   A level of  0.01 f/cc  is  recommended as  the  least stringent
 level  that should be  considered.

     Regardless  of which method  for  measuring asbestos  is used, the
 release criterion should be  specified  in terms of fiber  rather than mass
 concentrations.   The  number  of  fibers  rather than their mass is believed
 to be a better  indicator of  health effects  (NEC,  1984).   In addition,
 mass concentrations are unduly  influenced by a few large  fibers.  Thus,  in
 principle,  a superficial cleaning  of the work site could  significantly
 reduce  levels of  asbestos mass by  removing primarily a few large fibers,
 while leaving the concentration  of total asbestos fibers almost unchanged.
 However, if an  indirect sample preparation is used for TEM analysis in
 which fibers may be broken thus  increasing the number of fibers in the
 sample, the concentration of  fibers measured is likely to be higher than
 that in the air.  Under these conditions, mass concentration is the
preferred measure of asbestos levels.
    The asbestos measurement protocol for PCM specified by NIOSH (P&CAM
    239) requires that at least 10 fibers be collected for 100 fields of
    view on the filter.  This corresponds to a sample volume of about
    3,000 liters (see Section 4.3.3).  Thus, if a very stringent release
    criterion (very low concentration of asbestos) were desired, a very
    large sampling volume would be specified.
                                   5-2

-------
5.2  STATISTICAL CONSIDERATIONS FOR USING RELEASE CRITERIA

     If samples of air are taken in the same general area but at
slightly different locations or at different times at the same location,
the measurements of sampled material will differ.  Likewise, side-by-side
samples taken at the same location and time will vary.  Thus,' measure-
ments of airborne asbestos at an abatement worksite will be variable
irrespective of abatement activities or post-abatement cleaning efforts.
The task at hand is to understand this variability, and, using standard
statistical procedures, to determine whether two measurements are truly
different or differ only due to normal (expected) variability.

     The variability of measurements of airborne asbestos has two
components — sampling and analytic variability.  Sampling variability is
due to random fluctuations in the constituents of an air mass, and to
systematic factors such as air circulation patterns in a room.  Analytic
variability is associated with the instruments and procedures used to
sample air and analyze the samples.

     Recent EPA research studies provide information on the magnitude of
sampling and analytic variability for measurements of airborne asbestos
using TEM (USEPA, 1983; Chesson et al., 1985a; 1985b).  The results of
the analyses of variability in these studies are expressed as the
coefficient of variation (CV).   The CV is simply the standard deviation
of a series of measurements divided by the mean value.*

     The first study (USEPA, 1983) produced estimates of sampling
variability.  A CV of 0.88 was found in simultaneous measurements of
airborne asbestos among rooms with ACM in 25 school buildings within a
single school district.  Since the measurements of asbestos were
adjusted for between-school variation in mean asbestos levels, the
measurements can be considered reflective of spatial variability in a
single area.  Variability over time was estimated as the CV for average
weekly levels of asbestos in three different schools (CV = 0.42).
Spatial and time variability combined would thus be a CV of about 1.0.**

     The other two studies (Chesson et al., 1985a; 1985b) estimated the
analytic component of variability.  Using the variation between
laboratories in 49 pairs of TEM measurements of asbestos as the
indicator, analytic variability was estimated as a CV of-about 1.0.

     Based on these limited studies, expected variability in asbestos
levels at a single location (e.g., an abatement work site) as measured
by TEM may be characterized by a CV of between^1.0 and 1.5.
    Since the variability of measurements of airborne asbestos tends to
    be larger if the average value is high, a high standard deviation
    may reflect a high variability and/or a high mean for the
    measurements.  Dividing the standard deviation by the mean thus
    allows the variability of measurements with large means to be
    compared with the variability of those with small means.

    CV's are combined by taking the square root of the sum of each CV(
    squared.
                                   5-3

-------
      This information on the normal or expected variability of asbestos
 fibers is used in the following sections to calculate the required
 number of samples for determining compliance with the TEM release
 criterion.  (A different approach is used for the PCM criterion.)  Of
 course, the actual variation in asbestos levels at any site is
 calculated directly from the measurements at that site.   The degree of
 expected variability is assumed solely for the purpose of determining
 sample design specifications before the measurements  are made.
 5.3  TEM RELEASE CRITERION

      As  noted previously,  the recommended criteria  for releasing  the
 abatement contractor if TEM is used involves  comparing asbestos levels
 at the work site with those measured outside.   Only if the  asbestos
 levels inside are not statistically larger than those  outside  the work
 site, would the  contractor be released.

 5.3.1  Sampling  Volume and Time

     The  required sampling  volume  is determined  by the  lowest level of
 asbestos to which the work-site environment must be reduced.   As  shown
 in Table 5-1,  typical ambient asbestos  levels are on the order of 0.001
 in rural areas and somewhat higher  in urban areas.   Based on these data,
 enough air  must  be sampled to detect a  concentration of approximately
 0.005 f/cc.   As  described  in Chapter 4, a  volume of at  least 3,000
 liters per  sample is  required if  the sample preparation involves  direct
 transfer to the  EM grid, more if  the indirect sample preparation
 technique is  used.  At a rate of  2-12 L/min, sampling would require from
 3.5  to 21 hours.

 5.3.2  The  Number and Location of Samplers

     The number  of samples  needed to reliably determine compliance with
 the  release  criterion depends  on  several  factors.   Table 5-2 lists these
 factors  and  illustrates their  influence on  sample size.

     The first two  factors  are  the  expected errors  regarding decisions
 on satisfactory  cleaning of  the abatement work  site.  These are the
 probabilities  that  no  difference  between  levels  inside  and outside the
 work site will be  detected when the  work-site asbestos  levels  are
 actually too high  (false negatives),  or that a  difference will be
 detected when  the work-site  levels  are actually  low enough (false
positives).

     The  third factor  ("inside-to-outside multiple") is related to the
 false positive and negative error rates.  Since  small differences
between inside and outside asbestos  levels are more difficult  to detect
than large differences, more samples are needed to maintain the same
 rates of  making errors in decisions.  For example,  as shown in Table
5-2, if the CV for TEM is 1.5, seven samples are required to detect a
5-fold difference between inside and outside levels  with a 10-percent
chance of making a wrong decision.  However, only four samples  are
required  to detect a  10-fold difference.  In other words,  if seven

                                   5-4

-------
                          TABLE 5-1

 REPORTED LEVELS OF AMBIENT ASBESTOS FIBERS MEASURED BY TEM-
   Setting
Level (f/cc)
         Reference
    Urban
    Urban
  0-0.024
    Rural        0-0.004

  Industrial   0.0002-0.011
    Rural
  0-0.045
(most below
   0.01)
  0-0.003
Murchio, 1973.  "Asbestos Fibers
in the Ambient Air of California,"
California Air Resources Board.

            Same

John et al.,  1976. "Experimental
Determination of the Number and
Size of Asbestos Fibers in Ambient
Air," NTIS Report # PB-254086

Chatfield, 1983. "Measurement of
Asbestos Fibre Concentrations in
the Ambient Atmosphere," Royal
Commission on Asbestos.

            Same
All TEM analyses reported to have been made following direct sample
preparation procedures.
                               5-5

-------
                                TABLE 5-2

              NUMBER OF SAMPLES REQUIRED TO TEST COMPLIANCE
                     WITH THE  TEM RELEASE CRITERION
False False Inside-to-
Positive Negative Outside
Rate Rate Multiple
0.10 0.10 5
0.10 0.10 7
0.10 0.10 10
Number of Samples
CV=1.0 CV=1.5 CV=2.0
4 79
4 57
3 45
       0.10     0.05           5             5         9      12

       0.10     0.05           7             5         7      9

       0.10     0.05          10             4         5      7




Source:  Based on the method used in Breen et al., 1985  (Table 5)
                                   5-6

-------
samples are used, there is at most a 10-percent chance that the work
site would pass the air test when levels inside were actually five times
higher than outside.  If the true multiple is larger than 5 and seven
samples were collected, the likelihood of passing is less than 10
percent.  The inside-to-outside multiple is selected when planning the
air test.  A multiple in the range of 5 to 10 is typically used.

     The final factor is the coefficient of variation.  As discussed in
Section 5.3, the CV expresses the relative variation in asbestos
measurements for samples at the same site.  As shown in Table 5-2, the
higher the CV, the larger the sample size must be.

     Table 5-2 indicates that the number of required samples varies from
4 to 12.  To make the task of air testing following an abatement  project
practical, a minimum sample size of 5 is recommended.   This corresponds
to false positive and negative error rates of 0.10  each, an
inside-to-outside multiple of 5, and a CV of between 1.0 and 1.5.

     The recommended sample size of 5 applies to samples inside as well
as outside the work site.  Thus, a total of at least 10 samples would be
required if TEM is employed.  In selecting the location of the 10 air
samplers, the following guidelines should be considered:

    •    For indoor locations, first determine if the work site is
         homogeneous.   "Homogeneous" refers to contiguous areas with the
         same type of ACM and in which one type of  abatement process was
         performed.  For example, an auditorium with accoustical  ceiling
         plaster containing chrysotile that was removed would qualify as
         homogeneous.   Similarly, a corridor and connecting rooms on a
         single floor would be homogeneous if all areas contained the same
         type of ACM and the same abatement method  were used.   Separate
         floors within a building and separate buildings are usually
         considered different work sites.   Collect  five samples in each
         homogeneous work site.

    •    Place the samplers within the homogeneous  work site so as to
         collect representative samples.   If the work site is a single
         room, disperse the samplers throughout the area.  If the work site
         contains up to five rooms, place at least  one sampler in each
         room.  If the work site contains more than five rooms, select a
         representative sample of rooms.   The random number procedure in
         Appendix A is one way to select a representative sample.   Place
         each sampler so that it is subject to normal air circulation;
         avoid room corners, obstructed locations,  and sites near windows,
         doors, or vents.  Samplers placed outside  the work site  but within
         the building should be located to avoid any air that might escape
         through the containment barriers.  Minimum recommendations are at
         least 50 feet from the entrance to the work site, and 25 feet
         from the plastic barriers.

    •    Outdoor samplers should be placed at ground level (about 6 feet
         high), if possible, and away from obstructions that may
         influence wind patterns.  If access to electricity and concerns
         about security dictate a roof-top site, avoid locations  near
         vents or other structures on the roof.
                                   5-7

-------
      The above guidelines are designed to assure that representative
 samples of airborne asbestos are collected.

 5.3.3  Comparing Measured Levels of Airborne Asbestos

      The appropriate statistical test for comparing levels of asbestos
 measured at the work site with those measured outside is known as the
 "difference between means" using Student's "t" test:

          •    Compute the natural logarithm of fiber concentration for each
               sample.

          •    Compute means of the log-transformed data for inside samples
               and for outside samples.

          •    Form the ratio:
          Where:

               Y.  = the  average  of log concentrations  inside  the  work  site

               Y   = the  average  of log concentrations  outside  the work site

               S   = {[I(Y..  - Y.)2 +  Z(Y  -  -  Y  )2]  /  (n. + n   -
                         ij    i       °J    o        i    o
               n.  = number of samples inside  the work  site

               n   = number of samples outside the work site
               o
          •     Then compare  t to 1.86 if  10 samples were collected
               (the 95 percentile  point of  a  "t" distribution  with
               8 [n. + n  -  2] degrees of freedom).  If t exceeds  1.86,
               the  work site fails the test (consult a  statistics  text
               for  the appropriate t  value  if the degrees" of freedom are
               other than 8).

     The  fiber level data is log-transformed because  frequency distri-
butions of asbestos levels usually are highly  skewed.  The transformed
data are  adequately approximated  by  a normal distribution on which
standard  statistical methods can  be  used.

5.3.4  Recommended Actions If the Work Site Fails

     For each homogeneous work  site which fails the test (i.e., average
asbestos levels inside the work site are statistically greater than
those outside), the entire work site should be thoroughly recleaned.
Wet cleaning methods should be  used  (see Section 2.1 and the companion
EPA guidance document [EPA,  1985]).  New samples (at least 5) should be

                                  5-8

-------
 collected  in  the work  site  and  analyzed for asbestos as described above.
 This process  should be repeated until the work site passes the test.
 Note that  for an abatement  project with more than one homogeneous work
 site (as defined in Section 5.3.2), the release criterion should be
 applied to each work site independently.
5.4  PCM RELEASE CRITERION

     The contractor release criterion recommended for use with PCM
employs PCM's limit of  reliable quantification as opposed to asbestos
levels outside the work site for  comparison purposes.  As a result, some
of the specifications for using the PCM criterion differ significantly
from those  for TEM.

5.4.1  Sampling Volume  and Time

     A minimum of 3,000 liters of air should be collected by each
sampler.  Since PCM can only be used as an indirect measure of asbestos,
the sampling requirements should be at least as stringent as those for
TEM.  Based on the illustration in Chapter 4, a sampling volume of about
3000 liters will allow  the PCM method to reliably quantify fibers levels
as low as about 0.01 f/cc.

5.4.2  The Number and Location of Samplers

     The recommended minimum number of samplers is five per homogeneous
work site, or one per room, whichever is greater.  Again, the rationale
is that the minimum requirements for the PCM test should be at least as
stringent as the TEM requirements since PCM is only an indirect
indicator of asbestos.   (In fact, the PCM specifications are slightly
more stringent for work sites with more than five rooms.)  All of the
guidelines for locating samplers discussed in Section 5.3.2 (except the
need for outside samplers) apply to PCM as well.

5.4.3  Comparing Measured Levels of Asbestos to the Lowest
       Quantifiable Level

     The recommended test for the PCM release criterion is that each of
the five or more samples must be less than the PCM limit of reliable
quantification.   If 3,000 liters is the sampling volume,-this limit is
approximately 0.01 f/cc.  Using each sample in the test is more
stringent than averaging the sample values and using the mean, as
illustrated in Table 5-3.  As shown, the probability that the work site
would pass the test is  only about 0.12 if the true asbestos levels is
actually equal to 0.01  f/cc, due to variation in PCM measurements.
Thus,  the work site needs to be cleaned so that the actual air level is
lower than 0.01  f/cc to be assured that it will pass the test.

5.4.4  Recommended Actions If the Work Site Fails

     As with the TEM criterion, each homogeneous work site should be
completely recleaned if it fails the test.   Recleaning is followed by
resampling and reanalyzing the samples.

                                   5-9

-------
                           TABLE 5-3

      THE  PROBABILITY OF  CONTRACTOR RELEASE FOR DIFFERENT
              PCM FIBER LEVELS IN THE WORK SITE*
     Actual Fiber Level                Probability of Release
           (f/cc)
0.001
0.002
0.003
0.004
0.005
0.01
0.02
0.05
0.998
0.94
0.81
0.64
0.49
0.12
0.01
0.0003
A negative binomial distribution and a CV of 1.0 to 1.25 are
assumed for PCM fiber levels.
                               5-10

-------
 5.5  EXAMPLE APPLICATIONS OF THE TWO RELEASE  CRITERIA

      Examples  of how the recommended PCM and TEM  criteria  for  contractor
 release  are applied will illustrate the  guidelines  described in this
 chapter  and problems that may arise in certain  situations.  The following
 examples assume  that 3,000 liters  of air are filtered by each  sampler.

 5.5.1 PCM Example

               Work Site:   Five school classrooms, one auditorium, and one
               connecting corridor.   Samplers are  located in each area
               (seven samplers altogether).

               Release Criterion:   All measurements  must be less than 0.01
               f/cc.

               PCM Results:  <0.01,  <0.01,  <0.01, 0.045, <0.01,  <0.01,
               <0.01  f/cc.

               Interpretation of Results:  Work  site fails.

               First  PCM  Retesting  Results (after  recleaning the entire work
               site):   <0.01,  <0.01,  0.012, <0.01, <0.01, <0.01, <0.01

               Interpretation of Results:  Work  site fails.

               Second Retesting Results:  All seven  samples, <0.01 f/cc.

               Interpretation of Results:  Work  site passes.
     In this example, all PCM samples in the first test were below the
release criterion except the fourth sample, which is significantly
higher (0.045 f/cc).  According to the guidelines, the entire work site
should be recleaned.  Some may argue that only the room with the high
measurement should be recleaned.  However, the seven PCM measurements
only represent a sample of the entire work site.  Air samplers placed in
other locations may also show high levels.  In addition, the low
sensitivity and specificity of the PCM test argues for a comprehensive
response if the test fails.  (Remember that a work site could pass the
PCM test while failing the more sensitive and specific.TEM test.)  The
example also shows that one sample in the first retest (the third sample
this time) was still above the 0.01 f/cc criterion.  The fact that the
work site again failed to pass the test could reflect either inadequate
cleaning and/or normal variability in PCM measurements of airborne
fibers.  Nevertheless, a second recleaning and retesting is recommended.
Note that the five or more samplers need not be placed at the same
locations for retesting.

5.5.2  TEM Example

              Work Site:  The entire first floor of an office building
              (30 offices,  two rest rooms, a reception area, and a
              connecting corridor).  The five work site air samplers are

                                   5-11

-------
 placed in three offices,  one rest room,  and the  corridor.
 The five outside samplers are placed in  two stairwells
 between the first and second floors.

 Release Criterion:   The average of inside  samples  must  be
 statistically no greater  than the average  of the outside
 samples.
 TEM Results:
             Fiber  Level   Natural  Log  of     Mean  of
              (f/cc)	   Fiber  Level         Logs
                                            -3.48  (Y.)
Inside
Samples



Outside
Samples



0.073
0.032
0.008
0.057
0.026
<0.005*
0.010
0.024
0.009
0.015
-2.62
-3.44
-4.83
-2.86
-3.65
-5.30
-4.61
-3.73
-4.71
-4.20
                                           -4.5i  (YO;
Difference of means test:
                           71  . - Y
            ^j     	°J    o
                  n  - 2
                   o
=  {^

   [(2.98^ 2.26)]1/2  _
                             (-3.48) - (-4.51)
  t =
                             0.809
Interpretation of Results:   Work site fails because
                            t > 1.86.
                    5-12

-------
                                 ERRATUM
      This sheet replaces the TEM release calculations in example 5.5.2 on

 pages 5-12 and 5-13 of the document "Measuring Airborne Asbestos Following

 An Abatement Action."  (EPA 600/4-85-049)
Difference of means test:
                                           (p. 5-12)
S =
      [Z(Yii - Y^2
                     • - Y  )2
                   OJ    O'
              n  + n  - 2
 [2.98 +  1.381 1

 L5 + 5- 2j
                    /2   =
       J




  0.738
                  Note:
                                                           =  1.38
t =
                 =   (-3.48) - (-4.51)    =  2.21
        Vnj + Vn0      0.738 /(V5 + V5)
Interpretation of results:  Work site fails because t > 1.86.
Difference of means test:
                              (Retest, p. 5-13)
S =
f(4.80 + 1.38)1 l/2


L     ~    J
=   0.879
     (-4.17) - (-4.51)
t =
     0.879/(V5 + V5)
                         0.612
Interpretation of Results:  Work site passes because t < 1.86.

-------
              TEM Retesting Results  (after recleaning the entire work
              site):
                        Fiber Level    Natural Log of
                            (f/cc)      Fiber Level
                Inside      0.081
                Samples    <0.005~'
                            0.011
                            0.025
                            0.008
                          -2.51
                          -5.30
                          -4.51
                          -3.69
                          -4.82
                                          Mean of
                                            Logs
-4.17
                Outside    Same as before (no new sampling is needed)
                Samples
                * Assumed to be 0.005 for purposes of calculation.
              Difference of means test:

                              2.26)!1/2
  _ F(4.80 +
  "L      8
                                        = 0.940
t =  (-4.17) - (-4.51) =
                                          0.57
                       0.940
              Interpretation of Results:  Work site passes because
                                          t < 1.86.
     This example illustrates how the use of average fiber concentra-
tions in the TEM criterion influences the results.  After the initial
cleaning, levels inside the work site were significantly higher than
those immediately outside.  Since the t-test is greater than 1.86
(2.01), the entire work site is recleaned.  The second set of work-site
samples reveals lower air levels with one exception—the first sample
(0.081 f/cc) is higher than any of the samples in the first or second
rounds of testing.  However, the mean of all second round samples is
lower than the mean of the first, and the contractor is released since
the t-test is less than 1.86 (0.57).  Some may be concerned about the
single high level found during the second round of sampling, and would
argue for another recleaning of the work site. Recall, however, that TEM
measurements are expected to be highly variable, and that a single high
value is not necessarily a cause for concern.  On the other hand, a
simple rule such as "if any single value is more than 'x' times the mean
for all values, the work site must be recleaned" could be used as a
supplement to the recommended TEM criterion.  Such a rule would help
guard against the possibility of a single contaminated room in an
otherwise clean work site.
                                  5-13

-------
                               CHAPTER 6

                      QUALITY ASSURANCE PRACTICES
     Regardless of which method for measuring airborne asbestos  is  used,
reliable results can be obtained only if the collection,  transfer,
handling, analysis, interpretation, and documentation of  the data  follow
specified procedures.  Procedures for data collection, analysis, and
interpretation were described in Chapters 3, 4,  and 5; procedures  for
transferring and handling the data are included  in this chapter.   In
order to insure that all of these procedures are carefully followed,  a
quality assurance (QA) program is essential.

     Following are the key elements of a comprehensive QA program.

     Training and Experience

         •    Everyone involved with measuring airborne asbestos
              should be properly trained and should understand his  or
              her role.

         •    Only qualified air sampling firms  and analytical
              laboratories should be hired.   As  noted previously,
              NIOSH's Proficiency Analytical Testing Program is  a  good
              source of information on qualifications of  laboratories
              offering PCM services.  Information should  be requested
              from each PCM or TEM laboratory on:

                   the laboratory's quality  control program.

                   the lowest fiber counts (f/cc)  that are routinely
                   reported.

                   the thinnest fibers that  are  routinely detected.

     Quality Control Checks

         •    All sampling equipment should  be calibrated and checked
              as described in Section 4.2.

         •    All analytical  instruments  should  be calibrated with  NBS
              reference materials,  and checked as  described in Leidel,
              et al.,  1979,  or NIOSH,  1984 (PCM),  and Yamate, 1984
              (TEM).

         •    One field blank and one  laboratory blank per work  site
              should be analyzed to check for asbestos contamination
              of blank filters.

         •    All labels on filters should be coded to avoid  possible
              bias  by laboratory analysts.
                                   6-1

-------
    •    One filter per work site should be split for duplicate
         analysis by a second laboratory.  Where the duplicate
         analysis is significantly different, procedures used by
         either or both laboratories should be investigated until
         the source of the discrepancy is identified and
         corrected.

Data Handling Chain-of-Custody

    •    Responsibility for samples should be assigned to an
         individual at each stage of the testing process.

    •    Each step in the transfer of the data from field to
         laboratory to building owner should be recorded.

Documentation

    •    All testing procedures and test results should be
         documented.

    •    Unused filters and portions of filters should be saved
         for possible reanalysis at a later date.
                             6-2

-------
                              REFERENCES
Burdett,  G. J.,  and Rood, A. P.   1983.  "Membrane Filter, Direct
     Transfer Technique  for the Analysis of Asbestos Fibers or Other
     Inorganic Particles by Transmission Electron Microscopy," Environ.
     Sci. Technol., 17,  643-648.

Chatfield, E. J.   1983.  Measurement of Asbestos Fibre Concentrations in
     Ambient Atmospheres.  Ontario, Canada:  Ontario Research
     Foundation.

Chesson,  J., Margeson, D. P., Ogden, J., Reichenbach, N. G.,  Bauer, K.,
     Constant, P.  C., Bergman, F. J., Rose, D. P., Atkinson,  G. R., and
     Lentzen, D. E.   1985a.  Evaluation of Asbestos Abatement
     Techniques, Phase 1:  Removal.  Final Report. Washington, DC:
     Office of Toxic  Substances and Environmental Monitoring Systems
     Laboratory, U.S. Environmental Protection Agency.  Contracts
     68-01-3938  and 68-02-3767.

Chesson,  J., Margeson, D. P., Ogden, J., Bauer, K., Constant, P. C.,
     Bergman, F. J.,  and Rose, D. P.  1985b.  Evaluation of Asbestos
     Abatement Techniques; Phase  2:  Encapsulation.  Draft Report.
     Washington, B.C.:  Office of Toxic Substances, U.S. Environmental
     Protection  Agency.  Contracts 68-01-6721 and 68-02-3938.

Chesson,  J., Price, B. P., Stroup, C. R., and Breen, J. J.,   1985.
     Statistical Issues in Measuring Airborne Asbestos Levels Following
     an Abatement Program. American Chemical Society Symposium on
     Environmental Applications of Chemometrics in Philadelphia in
     August 1984.  (To appear in  forthcoming proceedings volume.)

National  Research Council.  1984.  Asbestiform Fibers, Non-occupational
     Health Risks.  Washington, B.C.:  National Academy Press.

Leidel, N. A., Boyer, S. G., Zumwalde, R. D., and Busch, K. A.  1979.
     USPHS/NIOSH Membrane Filter Method for Evaluating Airborne Asbestos
     Fibers.   Washington, D.C.:  National Institute of Occupational
     Safety and Health.

NBS/EPA.  1985.  Proceedings of the Workshop on the Monitoring and
     Evaluation of Airborne Asbestos Levels Following an Abatement
     Program.  Cosponsored by U.S. Environmental Protection Agency and
     National Bureau of Standards.  Held at NBS, Gaithersburg, Maryland,
     March 12-13, 1984.

NIOSH.   1984.  Fibers, Method 7400.  Cincinnati, Ohio:  National
     Institute of Occupational Safety and Health.

OMHL.  1984.   MDHS 39, Methods for the Determination of Hazardous
     Substances,  Asbestos Fibers in Air,  Determination of personal
     exposure by the European Reference version of the membrane filter
     method,  Health and Safety Executive.   London:  Occupational
     Medicine and Hygiene Laboratory.

                                   R-l

-------
 Samudra, A., Harwood, C. F.,  and Stockham,  J.  D.   1978.   Electron
      Microscope Measurement of Airborne Asbestos  Concentration:  A
      Provisional Methodology Manual.   Washington,  D.C.:   Office  of
      Research and Development.  U.S.  Environmental Protection Agency.
      EPA 600/2-77-178.

 Small,  J.  A., Steel,  E.  B., and Sheridan, P. J.   1985.   "Analytical
      Standards for the  Analysis of Chrysotile  Asbestos  in Ambient
      Environments,"  Anal.  Chem.,  57, 204-208.

 Spurny, K.  R.  1985.   "Measurement of Asbestos  and Other Mineral Fibers
      in Ambient and Indoor  Air,"  Proceedings of the Workshop on  the
      Monitoring and Evaluation of Airborne  Asbestos Levels Following an
      Abatement Program.   Cosponsored  by U.S. Environmental Protection
      Agency and National Bureau of Standards.  Held at NBS,
      Gaithersburg,  Maryland,  March 12-13, 1984.

 Steel,  E. B., and Small, J. A.   1985.   "Accuracy of Transmission
      Electron Micrscopy  for the Analysis of Asbestos  in  Ambient
      Environments,"  Anal.  Chem.,  57,  209-213.

 USEPA.   1977a.   Montgomery  County  Asbestos  Study.   RTP,  N.C.:  Office of
      Research and Development,  USEPA.

 USEPA.   1977b.   Quality  Assurance  Handbook  for Air  Pollution Measurement
      Systems,  Vol.  II: Ambient  Air Specific Methods. Washington, D.C.:
      Office  of Research  and Development, USEPA.  EPA-600/4-77-027a.

 USEPA.   1977c.   "Method  5: Determination of Particulate  Emissions From
      Stationary Sources," Federal  Register, Vol. 42, No.  160, pp.
      41776-41781.   (40 CFR  60,  Appendix A,  Method 5.)

 USEPA.   1983.   Airborne  Asbestos Levels in  Schools.  Washington, D.C.:
      Office  of  Toxic  Substances, USEPA.  EPA 560/5-83-003

 USEPA.   1984a   Evaluation of  the EPA Asbestos-in-Schools  Identification
      and Notification Rule.    Washington, D.C.:   Office of Toxic
      Substances, USEPA.  EPA  560/5-84-005.

 USEPA.   1984b.  Asbestos in Buildings:  National Survey of
     Asbestos-Containing Friable Materials.   Washington, D.C.:   Office
      of Toxic Substances, USEPA.  EPA 560/5-84-006.

USEPA.  1985.  Guidance for  Controlling Asbestos-Containing Materials in
     Buildings, 1985 Edition.   Washington,  D.C.:   Office of Toxic
     Substances, USEPA.   EPA 560/5-85-024.

Yamate,  G.,  Agarwal, S.  C.,  and Gibbons, R.  D.   1984.   Methodology for
     the Measurement of Airborne Asbestos by Electron Microscopy.  Draft
     Report.  Washington, D.C.:  Office of Research and Development,
     U.S. Environmental Protection Agency.   Contract 68-02-3266.
                                   R-2

-------
                              APPENDIX A

              A RANDOM NUMBER PROCEDURE FOR SELECTING A
                   REPRESENTATIVE WORK-SITE SAMPLE
     A table of random numbers can be used to pick a sample of rooms for
air testing as follows.   (A table of random numbers is simply many
sequences of single-digit numbers presented in a random order.)

    (1)  If the number of rooms is less than 10, each room is
         assigned a unique single-digit number.  Rooms with
         numbers corresponding to the first five numbers in a
         selected random number sequence in the table constitute
         the air test sample.

    (2)  If the number of rooms is greater or equal to ten (but
         less than 100), each room is assigned a unique two-digit
         number.  The random number table is then considered
         sequences of two-digit numbers.  The first five room
         numbers to appear in a selected sequence constitutes the
         air test sample.

    (3)  Similarly, if the number of rooms is greater than or
         equal to 100, each room is assigned a unique three-digit
         number and the random number table is considered
         sequences of three-digit numbers.

     Tables of random numbers can be found in any statistics textbook,
such as:   Snedecor, G. W., and Cochran, W. G., 1976.  Statistical
Methods (6th Ed.). Iowa State U. Press, Ames Iowa. 593 pp.
                                   A-l

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing/
1. REPORT NO.
  EPA 600/4-85-049
                              2.
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
  Measuring Airborne Asbestos Following  an Abatement
  Action
             5. REPORT DATE
               November 1985
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

  D.L. Keyes,  J.J.  Breen,  and M.E. Beard
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Environmental  Sciences,  6845 North Cocopas Road
  Tucson, Arizona,  85718,  under contract  to  the
  Research  Triangle Institute, P.O. Box 12194,
  Research  Triangle Park,  N.C.  27709
                                                            10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.

               68-02-3767
12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Protection Agency
  Office of Research and Development
  Environmental Monitoring Systems Laboratory
  Research Triangle  Park, N.C.  27711
                                                            13. TYPE OF REPORT AND PERIOD COVERED
             14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
    One of the most  critical points in an asbestos abatement project is knowing
  when the work has  been completed, the contractor can be released, and the  building
  can be reoccupied.   This decision should be based on two factors:  (1)  satisfactory
  performance of  the  abatement work, and  (2) thorough cleaning of the work site.   As
  outlined herein, these factors should be evaluated by virtually inspecting the  work
  site and by measuring  the level of airborne asbestos there.  The evaluation should
  be conducted by the asbestos program manager  or  the technical advisor assigned  to
  monitor the abatement  work.

       This report discusses various technical  issues regarding the air test.  Speci-
  fications are provided for air sampling and detailed guidelines are presented for
  using either TEM or PCM to analyze the samples.   The information is designed for
  asbestos program managers and technical program  advisors.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b. IDENTIFIERS/OPEN ENDED TERMS
                           c. COSATi I leld/Group
 Airborne Asbestos
 Asbestos Abatement
 Asbestos-Containing Materials
 Asbestos Control Program
 Asbestos Exposure
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (Tins Report I
                                                                          21. NO. OF PAGES
                                              20. SECURITY CLASS (Tim pj.ro
                                                                          22. PRICE
EPA Form 2220-1 (R«v. 4-77)   PREVIOUS EDITION is OBSOLETE

-------