PB82--t96VM8
   Feasibility of  Developing Source  Sampling
   Methods for Asbestos Emissions
   Battelle Columbus  Labs.,  OH
   Prepared for

   Environmental Sciences Research Lab
   Research Triangle  Park,  NC
   Apr  82
U.S. Department of Commerce
National Technical Information Service

-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions an the rerene before completing)
I, REPORT NO.
                                                           3. RECIPIENT'S ACCESSION NO,
                                                                    •S ACCESSION NO,
                                                                    i   19614 8
I. TITLE AND SUBTITLE
                                                           5. REPORT DATE
FEASIBILITY OF DEVELOPING SOURCE SAMPLING
METHODS FOR ASBESTOS EMISSIONS
                                                           6. PERFORMING ORGANIZATION CODE
7. AUTKOR(S)

W. H. Henry, 6.  M.  Sverdrup, E. H. Schmidt and
S. E. Hiller	
                                                           3. PERFORMING ORGANIZATION REPORT NO.
                                                           10. PROGRAM ELEMENT NO.

                                                            C9TAW01-1218(FY-821
). PERFORMING ORGANIZATION NAME AND ADDRESS

Jattelle, Columbus  Laboratories
505 King Avenue
Columbus, Ohio   43201
                                                           li.CONTBACT/G'ffANT ~NO.~

                                                            Contract No. 6B-QE-3169
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP,  NC
3ffice of Research  and Development
J.S. Environmental  Protection Agency
Research Triangle Park, North Carolina  27711	
                                                           13. TYPE OF BEPOHT AND PERIOD COVERED
                                                            Final Report. 12/23/80-6/30/81
 Repo
iORINC
                                                           14. SPONSORING AGENCY CODE

                                                            EPA/600/09
15, SUPPLEMENTARY NOTES
IB. ABSTRACT
The objective  of this program was to determine the  feasibility of developing methods
for sampling asbestos in the emissions of major asbestos sources:  (!) ore produc-
tion and taconite production, (2) asbestos-cement production, (3) asbestos felt and
paper production, and (4) the production of asbestos-containing friction materials.
Potential sampling methods must provide samples compatible with the provisional
analysis methods using electron microscopy (U.S. EPA Report No. 600/2-77-178).

Two general criteria for source sampling methods were identified as:  (1) the samp-
ling method must be capable of collecting a representative sample and (2) the asbestos
emissions must be collected in such a manner that they can be analyzed by the pro-
visional analytical method.  Concurrent investigations of potential emissions in the
industries and of current knowledge of sampling fibers were undertaken to assess the
feasibility of meeting the first criterion.  The industry survey revealed that
asbestos emissions can be divided into two classes:   stack and fugitive.  With res-
pect to the second criterion, it is not feasible to  undertake a methods develop-
ment program for strict compatibility with the recommended procedure of the provisi-
sional analytical method.  However, methods development programs are feasible if
the sampling method is to be compatible with the alternative procedures of the pro-
visional method or general electron microscopy.	^	
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lOENTIFIERS/OPEN ENDED TEAMS
                                                                         c. COSATI Fidd/Gioup
IS. DISTRIBUTION STATEMENT

     RELEASE TO  PUBLIC
                                              IB. SECURITY CLASS (TkliReport)
                                                 UNCLASSIFIED
        21. NO. OF PAGES
              69
                                              2Q. SECURITY CLASS {Thir page)

                                                 UNCLASSIFIED
                                                                         22. PRICE
IP* Fwm 2220-1 (Rev, 4-77)   PREVIOUS EDITION it OBSOLETE

-------
                                         EPA-600/3-82-008
                                         April 1982
                                                 PB82-196Ht8
            FEASIBILITY OF DEVELOPING SOURCE
         SAMPLING METHODS FOR ASBESTOS EMISSIONS
W.M. Henry, G.M. Sverdrup, EiW, Schmidt, and S.E. Miller
             Battelle Colurnbus Laboratories
                     505 King Avenue
                  Columbus, Ohio  43201
                 Contract Nol 68-02-3169
                   Work Assignment 10
                     Project;Officer

                    Kenneth t. Rnapp
       Stationary Source Emissions Research Branch
      • Environmental Science^ Research Laboratory
     Research Triangle Park,!North Carolina  27711
      ! ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
           OFFICE OF SESEARc|t AND DEVELOPMENT
          U.S. ENVIRONMESTAI,! PROTECTION AGENCY
      .RESEARCH TRIANGLE PARK^ NORTH CAROLINA  27711

-------
                                DISCLAIMER
     This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation.  Approval does not signify that the contents necessarily reflect
the views an£ policies of the U.S. pnvironmental Protection Agency, nor
does mentlon!of trade names or commercial products constitute endorsement
or recommendation for use.         !
                                  ii

-------
                               ABSTRACT

          The objective of this program is to determine the feasibility of
developing methods for sampling asbestos in the emissions of major asbestos
sources.  The sources of concern are:  (1) ore production including
asbestos mining and milling aod taconite production, (2) asbestos-cement
production,  (3) asbestos felt and paper production, and (4) the production
of asbestos-containing friction materials.  Potential sampling methods must
provide samples compatible with the provisional analysis methods using
electron microscopy (U.S. EPA Report Ho.  600/2-77-178),
          Two general criteria for source sampling methods were identified
at the onset of the program.  These criteria ares  (1) the sampling method
must be capable of collecting a representative asbestos size distribution
from the local environment, and (2) the asbestos emissions must be collected
in such a manner that they can be analyzed by the provisional analytical
method to provide the required determinations.
          Concurrent Investigations of potential emissions in the industries
and of current knowledge of sampling fibers were undertaken to assess the
feasibility of meeting the first criterion.  The industry survey revealed
that asbestos emissions can be divided into two classes:  stack and fugitive.
Inherent differences between stack and fugitive emission environments may
necessitate the development of two techniques or at least two modifications
of a general technique for sampling.  A development program for sampling
methods is feasible given the nature of the emissions and potential sampling
environments observed in the Industry survey.
          With respect to the second criterion, it is not feasible to under-
take a methods development program for strict compatibility with the recom-
mended procedure of the provisional analytical method.  Strict compatibility
requires the collection of a uniform deposit of proper loading by air fil-
tration onto a 0.4 ym pore size polycarbonate filter.  However, methods
development programs are feasible if the sampling method is to be compatible
with the alternative procedures of the provisional method or general electron
microscopy.  Such procedures require that the collected sample be transferable
to an electron microscope grid for counting.  The method of sample collection
is not precisely specified.
                                     iii

-------
          Viewed on a component-wise basis, the essential areas for research
toward method development concern collection techniques and removal of
nonasbestos material.  Practical options for the collection technique
component are limited to either (1) electrostatic precipitation or (2) collection
by cellulose ester or polycarbonate filters in spite of their known limitations.
These techniques may be supplemented by preeollection with an impinger to
reduce loading.  Past experience o£ analysts indicates that asbestos and
nonasbestos material can be separated from each other in the laboratory
by means of ashing. Bonification, and two-phase liquid separation.  These
sample preparation procedures can alter the asbestos size distribution.  The
usefulness as well as the feasibility of a separation during sampling can be
assessed only after more thorough data characterizing the industry emissions
are obtained and evaluated.  The applicability of inlet and probe technology
appears to be simply an engineering task.
          This report was submitted in fulfillment of Contract 68-02-3169,
Work Assignment 10, by Battelle's Columbus Laboratories under the sponsorship
of the U.S. Environmental Protection Agency.  This report covers the period
December 23, 1980 through June 30, 1981.  Work was completed as of May 29, 1981,

-------
                                 CONTENTS


Abstract	i	    ill

figures  .............. 	 .  •    vii

Tables	    vii

Acknowledgement	 .	   vlii

     1.   Introduction 	 	 ..........      1

               Objective	      1
               Program Design	      1'
             '•  Report Organization  ................      2

     2.   Conclusions  .	      4

     3.   Recommendations  ..... 	      6

     4.   Background	 	      7

               Characteristics of Asbestos  	      7
               Characteristics of Emissions  ...........      9
                    Characteristics of ^articulate Emissions  ...      9
                    Qalssion Environment 	  ...     11
               Source Sampling Methods 	     12
                    Stack Sampling Methods	     14
                    Fugitive Emission Sampling ..........     15

     5.   Criteria and Constraints  ,\	     17

               Criteria	i . • •	     17
               Constraints 	j 	  ......     17
                    Constraints on  the Acquisition of a
                      Representative,' Sample	     17
             :       Constraints Arising from the Analytical
                      Method ....|...  	     25
               Implications of Constraints on Feasibility  ....     27

     6.   Feasibility of Method Development  	     29

               Inlet and Probe . .  .}	,	     31
               Extraneouo Material  Separation  	 .....     31
               Collection Techniques! .	     32

-------
 7.    Feasibility Assessment .........  	  ...     36

           Collection Techniques  ...;..,...	     36
           Extraneous Material Removal  ..............     36
           Applicability of Inlet and iProbe Technology 	     38

 8-    Approach to System Development  <	     39

           Research and Development fdr Extraneous  Material
           Removal	.;	     39
           Research and Development for Collection  fechniques  .  .     39
               i Electrostatic Collector   .....  	     39
               ' Filters  	  J  ........  	     41
&	; Impinger's  .--. -.-.--.  t  .---s  .•-.-.  .-•.  i  .-- .  *  -;•-.--; --  -42-

 References  . '	 ....=  ......	     43

 Appendix                             •

      A.   Description of the Industries	     46

-------
                                   FIGUB1S
Hmnber         '                       ;                                Fage

   1      Factots leading to constraints on a source sampling method
               for asbestos	•	    20

   2      Sampling method development (program flow chart  .....    40
                                   TABLES


   1      General Composition of Industrial Emissions Containing
               Asbestos	:	    10

   2      Conditions Encountered in Industrial Survey of Sampling
               Environmental (Stack Environments) 	 ...    13

   3      Constraints on a Source Sampling Method for Asbestos for the
               Acquisition of a Representative Sample .......    18

   4      Constraints on a Source Sampling Method for Asbestos to be
               Compatible with the Analytical Method for Asbestos
               Determinations ...................    19

   5      Implications of Constraints ;	 .    28

   6      System Components	, . .	    30

   7      Summary of Collection Options . 	    37
                                    vii

-------
                            ACKNOWLEDGMENTS

          The authors would like to acknowledge the helpful advice and
guidance provided by Battelle colleagues Dr. J, A, Gieseke, Dr.  K. W.  Lee,
Mr. C. V. Melton and Ms. S. J. Anderson,  The support and efforts of
Ms. W. N. Cooke and Ms. M. A. Roberts are also greatly appreciated.
          The Project Officer, Dr. Ken Knapp, of the Stationary Source
Emissions Research Branch, Environmental Sciences Research Laboratory,
U. S. EPA, provided direction for the study.
          The industrial survey was conducted in conjunction with Mr.  Michael
Laney and Ms. Laura Conrad of Research Triangle Institute.  Their cooperation
in coordinating the site visits contributed materially to this study.
          The cooperation of many individuals in the industries which  were
visited and the Asbestos Information Association/North America is sincerely
appreciated.
          Likewise we wish to acknowledge the guidance provided by Mr. John
Copeland and Mr, Gilbert Wood of the Industrial Studies Branch,  Office of
Air Quality and Planning Studies, IJ. S. EPA.
                                     viii

-------
                               SECTION 1
                             INTRODUCTION

          Asbestos has been identified as a hazardous air pollutant and is
therefore subject to a National Emission Standard for Hazardous Air Pollutants
(NESHAF).  However, a numerical standard has not yet been promulgated due
in part to the absence of a refsrence source sampling method for asbestos
emissions and a reference method for the analytical determination of
asbestos in collected samples.  A provisional analytical method has been
established based upon electron microscopy (1) .  Research is continuing on
the establishment of a reference analytical method based upon the current
provisional method.  This report describes the first phase of research
leading toward the possible development of a reference source sampling
method for asbestos emissions.

OBJECTIVE
          The objective of this program is to determine the feasibility of
developing methods for sampling asbestos in the emissions of major asbestos
sources.  The sampling methods must provide samples compatible with the
provisional analysis methods described in Sei. 1 (EPA-600/2-77-178),
Information is to be gathered in order that estimates can be made of time
and effort required to develop methods.

PROGRAM DESIGN
          The development of a reference sampling method involves feasi-
bility assessment and subsequently a development effort.  In general, a
feasibility study is designed to gather information on the requirements
that a reference method must meet, constraints placed upon potential methods
by a variety of sources, and needed areas of research.  This information can
then be used to determine the feasibility of conducting a development
program.

-------
          The development program, which is outside the scope of this study,
involves conducting research on one  or more potential sampling methods.
After research on unanswered questions, potential methods are tested and
either dropped from consideration or refined.
          This feasibility  study was designed to provide data which could
be used to determine whether or not it is technically feasible to initiate
a development program  for a reference source sampling method and, if so,
to estimate  the required time  and effort.
          The hazardous pollutant of concern is asbestos.  For purposes of
sampling, asbestos is  primarily chrysotile, amosite, and crocidolite.
These exist  primarily  as fibers or groups of fibers of various diameters
and lengths.  By convention, fibers are those particles having parallel
sides and length-to-diameter ratios of at least 3:1.  The diameter range
extends from 0.03 pn o.d. for hollow chrysotile fibrils to on the order of
10 pm for clumps of fibers.   The diameter range of individual amphibole
fibers is 0.1 to 0.2 vm.  Commercial asbestos has diameter ranges of 0.75 to
1.5 ym and 1.5 to 4 pm for  chrysotile and amphiboles, respectively.
          The four industries  of concern are:
          9  ore production
             —asbestos production
             —taconite production
          •  asbestos-cement
          •  asbestos  felt  awd paper
          *  asbestos  friction materials.
          Taconite production  differs from the other industries in that
the fibers which are present are sn extraneous impurity, not a desired
component of Che product.   Fiber emissions from r.ll the industries oecm
as stac'c emissions and fugitive emissions.

REPORT ORGANIZATION
          This report  is organized into 8 sections.  Section 2 contains the
conclusions  derived from the feasibility study.  Section 3 contains
recommendations on the initiation of a development program for source
sampling methods.

-------
          A background discussion of the characteristics of asbestass
industries, and asbestos source sampling methods is given in Section 4,
This is followed in Section 5 by identification of two criteria for choosing
an acceptable source sampling method and associated constraints upon
potential methods.  Potential components of a sampling method are presented
in Section 6.  Section 7 contains a discussion of the feasibility of source
sampling methods.  A feasible development program for a source sampling
method is presented in Section 8.

-------
                                  SECTION 2
                                 CONCLUSIONS

          After review of each of the key system components of a sampling
systems(s) it has been concluded that the development of a standard method
for sampling asbestos emissions is feasible.  This study has not uncovered
any limiting industry anomalies or insurmountable technical problems.
          It is not feasible to undertake a methods development program
for strict compatibility with the recomanded procedure of the provisional
analytical method.  Strict compatibility requires the collection of a
uniform deposit of proper loading by air filtration onto a 0.4 urn pore size
polycarbonate filter.  However, methods development programs are feasible
if the sampling method is to be compatible with the alternative procedures
of the provisional method or general electron microscopy.  Such procedures
require that the collected sample be transferable to an electron microscope
grid for counting.  The method of sample collection is not precisely specified.
          Inherent differences between stack and fugitive emission environments
may necessitate the development of two sampling techniques or at least two
modifications of the same technique.
          Viewed on a component-wise basis, the essential areas for research
toward method development concern collection techniques and removal of non-
asbestos material.  Practical options for the collection technique component
are limited to either (a) electrostatic precipitation or (b) collection by
cellulose ester or polycarbonate filters; although each of these options
possesses negative features for the overall sampling and analysis procedure.
          The negative features of cellulose ester filters include high pres-
sure drop and sample losses in the transfer of collected asbestos to an EH
grid.  The negative features of polycarbonate filters include less than
100 percent collection efficiency and the tendency for collected asbestos to
become detached from or move around on the filter during handling operations.

-------
These collection techniques may be supplemented by preeollection with an
impitiger to reduce loading.  Past experience of analysts indicates that
asbestos and nonasbestos material can be separated from each other in the
laboratory; however, ashing, Bonification, and two-phase liquid separation
techniques can alter the asbestos sise distribution,  fhe usefulness as well
as the feasibility of a separation during sampling can be assessed only after
more thorough data characterizing the Industry emissions is obtained and
evaluated.  The applicability uf inlet and probe technology appears to be
a straightforward engineering task.

-------
                               SEC1IOK 3
                            BECOMMENDATIONS

          A development program for a source sampling system should proceed on
a component-wise basis.  This would entail mutual pursuance of research efforts
on collection techniques and extraneous material separation during sampling.
Subsequently the most promising of the techniques should be incorporated with
each other and current state-of-the-art inlet and probe designs to form a
sampling system(s).  Finally the complete system must be laboratory checked
and field demonstrated.
          Investigation of  collection techniques should center on electro-
static collectors and  on collection by cellulose ester and polycarbonate
filters despite  the limitations of each of these options.  More industrial
data further characterizing the extraneous material needs to be obtained to
assess whether a development program for removal of extraneous material
should focus on  separation  during sampling, in the laboratory, or both.
The advantages of precollection with impingers or other means to reduce
loading should be evaluated experimentally.

-------
                               SECTION 4
                              BACKGROUND

          Three basic considerations which must be addressed when considering
the feasibility of source sampling methods are:  (a) the characteristics
of asbestos, (b) characteristics of the total emissions, and (c) current sampling
methods.  Pertinent information on these areas provides a background
upon which to base the feasibility assessment.

CHARACTERISTICS OF ASBESTOS
          A variety of terms have been applied to asbestos with different
connotations to mineralogists, the general scientific and technical
community, and the public,  for purposes of sampling it is sufficient to
restrict attention to six classes of asbestos arising from two minerals:
serpentines and amphiboles.  About 95 percent of the asbestos used in the
industries of concern is chrysotile, a serpentine mineral.  The retaining
five types of asbestos are amphiboles.  They are amosite, crocidolite,
anthophyllite, tremolite, and actinolite.
          The several types of airborne asbestos may be iti any or
all of three forms; fibrils, fibers and fiber bundles with bundles
being less prevalent than fibers or fibrils.  It is important that
sampling, sample preparation and analysis are conducted such that
tte integrity of the airborne form of asbestos is maintained.
Disruption, either by breaking bundles or fibers apart or by clustering
fibrils or fibers, during any step of the method will lead to false
representation and conclusions especially regarding number concentration
and size.
          Characteristics of fibers which are important in the design of a
sampling method include:  (a) aerodynamic behavior in force fields,
(b) light scattering if used as a direct detection technique, and (c) inter-
action with the collection medium.  A fundamental characteristic of fibers
is the length-to-dianieter or ^aspect ratio which by common working definition
must have a value greater than three for classification as a fiber.

-------
          Both filler diameter, d_, and aspect ratio (L/d|) influence
the aerodynamic behavior of asbestos.  The fundamental physical unit  of
chrysotile is the fibril, a hollow crystal with mean internal diameter of
0.018 iim (1 urn = 10   m) and mean outside diameter of 0,034  pm within the
range 0.03 to 0.04 urn.  A large number of fibrils constitute a ctrysotile
fiber which commonly has a diameter between 0.75 and 1.5 urn.  Chrysotile
fibers are not perfectly straight and are often frayed and contain fibrils
projecting frcm the fiber.
          The amplibole asbestos  fibers  are rod-like with straight sides.
The mean diameter range of  elementary amphibole asbestos fibers is 0,1 to
0.2 vim-  The  diameter range of  commercial amphibole asbestos fibers is about
1.5 to 4 pm.
          There  is  not  much reported in  the open  literature on determination
of  equivalent aerodynamic diameters,  d  , for asbestos.  Equivalent aero-
                                       36
dynamic  diameter is defined as  the diameter of a  sphere of unity density
whose settling velocity is  equal  to  that of the actual particle or fiber
under consideration.  The following  relationships have been reported for
amosite  and crocidolite(2),  the two  amphiboles most often encountered within
the four industries of  concern.
                                      ft)
                                          0.116
d   = 2.18 d, f~~|         amosite
 ae         f
                                          0.171
                              2.19dff-7-l         crocidolite
          Additional properties which affect aerodynamic characteristics
in force fields are specific gravity and electric charge.   The specific
gravity of chrysotile has been given as 2.4 to 2.6(3).  Amosite and
crocidolite have specific gravities in the ranges 3.1 to 3.25 and 3,2
to 3.3, respectively.  Normal electric charge of chrysotile is positive
while that of the amphiboles is negative.

-------
          The chemical nature of asbestos includes stability with respect
to solvents and temperature.  Chrysotile is the most soluble form of
asbestos.  Acids readily decompose the MgOH surface.  Amphiboles are more
resistant to acid attack.  Chrysotile begins to lose its water of crystal-
lization at about 300 C.  At 850 C Chrysotile is transformed to nonfibrous
magnesium olivene,  Amphiboles are more refractory than Chrysotile.  Loss
of water and fiber deterioration of amphiboles occurs at higher temperatures
(ca, 1000 C).

CHABACTE1ISTICS OF EMISSIONS
          The four industries of concern were listed in the Introduction.
Site visits were made to at least one plant in each of the four industrial
categories.  Descriptive summaries of these visits are provided in Appendix A.
Information gathered on particulate emission source characteristics and  the
emissions environments are discussed below.
Characteristics of Particulate Emissions
          The composition of emissions is determined by both the industrial
process and the existing control technology.  The likely general composition
of particulate emissions is shown in Table 1.  Chrysotile is the major
asbestos component in all industries but taconite production.  Amphiboles
are present in the AC pipe, friction products, and felt and paper products
industries.
          Composition of the particulate emissions from the production process
varies at different stages along the process.  In the asbestos mining and
milling industry, the percentage of asbestos in the material being handled
increases from its initial concentration in the ground of 2-60 percent up
to nearly 100 percent in the bagging operation as the ore is processed.   In
the taconite industry the concentration of fibers in the tailings collected
along the process changes as Increasing amounts of iron  are separated from
the ore.  In the manufacturing industries, the concentration of asbestos
in the emissions from the production line decreases from near 100 percent
at the point of Introduction of asbestos to an amount comparable to the
asbestos concentration in the manufactured product.

-------
                      TABLE 1.  GENERAL COMPOSITION OF INDUSTRIAL
                                EMISSIONS CONTAINING ASBESTOS
          Industry                     Asbestos                      Bonaebestos

Asbestos mining and milling   Chrysotile                      Rock containing asbestos
                   (a)
Taconite production           Mineral clevage fragments       Rock containing fibers

Asbestos-ceaent pipe          Chrysotile, crocidolite         Portland cement   40-55%
                              Asbestos   15-35%               Silica            24-33%

Friction products             Asbestos   30-80%               Friction compounds,  ZnO,
                                                              sulfur, rubber, resin,
                                                              "brass wire

Felt and paper products       Chrysotile  80-90%              Resins, latex, cement,
                                                              gypsum, starch, glue,
                                                              fiberglass
(a)  Controversy exists as to whether the amphibole fibers produced by the
     crushing of ore in the taconite Industry are truly asbestos or not,
     Zoltai (4) reported that the appropriate mineralogical term for the fibers
     derived from the Peter Mitchell ore of Reserve Mining Company is:  fibrous
     clevage fragments of cunningtonite-grueerite; however, he also indicated
     that there is no conclusive means by which these fibers and fibers of
     commercial amosite can be distinguished in micrometer size samples.  For
     the purpose of sampling, there appears to be no reason to distinguish
     between cleavage fragments and amosite.
                                              10

-------
          The composition of emissions leaving stacks is dependent upon
control technology.  As will be shown in Table 2, baghouses are used for
specific parts of the production process and also for collecting emissions
simultaneously from several operations thereby effectively mixing various
compositions.
          Both production processes and control technology influence the
size of emitted particles.  The size distribution of particles (including
fibers) passing through a baghouse will have a smaller mean size than most
fugitive emissions.  Asbestos containing particles emitted from finishing
operations will be present in the emissions of manufacturing plants.
These particles consist of asbestos embedded in small chunks of tha
finished product and possess a size larger than that of the asbestos itself.
Emission Environment
          A source sampling method must be able to extract a sample from
the local sampling environment.  The variables which characterize these
environments can be used to categorize the environments with respect to
feasibility of sampling methods.
          The primary categorization is based upon control of air flows
which potentially contain asbestos emissions.  These two categories
are (a) stack environments in which the air flow is constrained by a duct,
and (b) fugitive emissions in which asbestos is entrained by uncontrolled
air flow.  Fugitive emissions of concern can occur as (i) ventilation air
leaves a plant, (ii) indoor plant air escapes through open doors, windows,
or panels, or (iii) outdoor emissions from mining, transport, and disposal
operations.
          Additional variables of the sampling environment include tempera-
ture, relative humidity, air flow velocity, temporal variations of the
characteristics of the sampling environment, and physical accessibility.
Physical accessibility is a practical constraint.  The physical character-
istics of process machinery and building structures limit the sampling
volume itself, access to the sampling volume (e.g., suitable sampling ports)
and the amount of working space around the sampling volume.
                                    11

-------
          The characteristics of stack environments in the four industries
are shown in fable 2.  The values shown are approxinate In nature tat
sufficient for establishing feasibility.  Parameters shown include stack
diameter, volumetric flow rate, the computed average gas velocity, temperature
and moisture content.
          Sampling environments likely to be encountered when sampling
fugitive emissions of ashestos can be divided into two classes:  (a)outdoor
 and  (b)  emissions from industrial plants;  although,  these environments are quite
similar.  Temperature and moisture content are at or near ambient for all
environments.  Air velocity in the outdoor environment is the ambient
wind velocity for emissions from disposal sites.  Air velocities encountered
around mining operations are also close to the ambient wind velocity out-
side of areas in  close  proximity to blasting operations.  Air velocities
for fugitive emissions  from plants are the ambient air currents through
openings such as  windows, doors, or natural draft ventilators.  Some ventila-
tors use large fans  to  exhaust air from drying operations.
          The time dependence of the characteristics of the emissions and
sampling environment place an additional constraint on a sampling method.
The ability to collect  a time-integrated sample over a period of time long
in duration compared to the period of parameter fluctuation is necessary
in order to collect a sample representative of the emissions.
          This constraint has  further consequences for a sampling method.
Samples could be  collected continuously or intermittently over a specified
time period.  As  the length of sampling Is increased to achieve time inte-
gration, the sampling rate must be correspondingly decreased if the same
amount of asbestos is  to be collected.  The ability to determine accurate
sample volumes of air must be maintained as the sampling rate and/or
sampling durations are  reduced.

SOURCE SAMPLING METHODS
          As indicated by the industry review both stack sampling and fugitive
emissions sampling must be considered.  This section reviews each.
                                   12

-------
TABLE 2.  CONDITIONS ENCOUNTERED IN INDUSTRIAL SURVEY OF
          SAMPLING ENVIRONMENT (STACK ENVIROIWENTS)

Icdystry
Taconiie Frod«e tiara/
iron Ore Ifineflciation







Asbestos «ining Ik Millltig
Sice 1








Sice 2





AsbesEDt-CtiMifiE Products
Sice 1



Sit* 2


Aibci.it.ai F*lt



Afttttoi Friction
Product*

*""
Process
dr Pusiper Plant
Fine Crusher Plant
Floe Crusher Cooveyor-To*
Concentrator
Dry Cabbing Plant
CoDC«str*tor FJ«iitE
Filter Plant
Felieeigittg, Plant
Sack Pellet Stora^ Silo

Ore PEepiraiion


UryiBg

IT UBS? art-Dry Koct
Storage EC mil
Hilling & Bagging

Kaii. Dryer
Ha in Bigiing

Jhryer
gagger
i*g Cleaner

UCL End
tfet End
Finishing
Fidi lilting
•leading, Practising,
flnlihjng
Air Over Caring Oven
HeadinK
Drying
Tricing
Acphilt Saturator
Kiting
Pr* forming 1
rinlaMnE
Sti
Ointrol Tecbuology
Bagiujyse
Eaghause
Baghoase

Baghouse
Cyclones
Cyclones
Vet ESP
UnconETolIe'3

3 Baghoyses


2 Bajjhmises

Sagliause

Bagtvoiise
|4-E*tiatiisI points j
1 (roa fiber dust!
I "sie° j
Baghogsc
2 6« Rhesuses

Bmghousft
Saghagse
Bajghoiise

Baghouse
Baghouse
Sag.Kjuse
.ijsliua^u
IlgbOUSB

Hone
llei Impingfrr
Hone
Cyclone * Saghouse
FiberglBBB Hat
liihoaie

Bciboufe
*** ' *
Cincftes)
61
32
to





18

3t
16
40
43
50
S6

52x53

SO
24
30130
ISxlB
ISj.26
lt>16

311(0
41«51
27xJ2
S?xl22
30

60
22
—
16
—
30

30


67, 000
13,000
30,000





i.OCO

21,000
12.600
33,000
29,000
38,000
32,000

»,000

38,000*
«,»«
13,000*
7
1,500
2.500"

12,000
22,000
22,000
28, SOU
26,300

50.000
7.500
—
«,toe*
—
23,000

•6,000

,„_,
3,300
2,100
3,400





3,*00

s,loa
2.2W1
t.600
2,300
2, (00
1,»M>

1.000

1..SOB
2,000
2,100
J
320
1.400

l.'OO
1,500
2,000
1,200
5,400

i.joo
1,200
—
3,200
—
4,200

t.100

If) rh
Aab -tlO Aab
Azb +15 Alb
/tab -fi teb





Anb ^50 Sicurated

Affib teb
Aab Aib
Acb Acb
100 S.turatcd
120 Saturated
6€t &nb

Aab A&b

2bO Elevated
ABQ A&l)
Aab Amb
250 Elevatel
A»b Acb
A3b Affib

A>& Aeb
Afflb Aab
Amb Acb
Aab Ml,
Asb ABO

Aab «sb
A^b Saturated
„
Aab Aatt,
—
Aab tab

Aab Amb
                          13

-------
 Stack Sampling Methods
          Historically,  stack sampling methods for asbestos emissions have
been based upon total particulate sampling methods.  The Canadian standard
reference method (5) specifies an in-stack filter In a sampling train essen-
tially equivalent to U.S. EPA Method 17 (6).  The filter holder and filter must
be capable of withstanding temperatures up to 200 F.  A cellulose ester membrane
filter with 0.8 iim pore size is required.  The probe must have a heating system
capable of maintaining the temperature of the gas at the exit end of the probe
high enough to prevent condensation.
          The U.S. EPA has recommended (see Appendix G in lef. 7) a method for
sampling asbestos emissions which is also based on Method 17.  Inasmuch as
asbestos emissions are not affected by temperature below 300 F, the collection
temperature of 250 F for total particulate sampling need not be maintained.
Particulate matter may contain condensible material; asbestos does not.  Relax-
ation of this constraint eliminates the necessity of employing a heated probe
and filter system.  Sampling in the stack at stack temperature reduces the
distance travelled by the fibers going from the stack environment to the filter.
Elimination of heating has the consequence that this method is no longer suitable
for environments containing saturated water vapor or liquid drops.
          Sampling conducted at iron ore beneficiation plants for fiber
emissions has used both in situ and extractive sampling ;(7,8).  Extractive
sampling was used at a dock pellet storage silo ventilator stack because of
saturated conditions in the stack.  The sampling train was heated from the inlet
through the 47-mm polycarbonate filter (7).  Sampling of the baghouse exhausts
from  the ore car dump, fine crusher, and fine crusher conveyor—to-concentrator
storage silos was accomplished by in situ filtration (7).  With the exception
of one test using a cellulose ester filter, all tests were conducted using a
47-mm polycarbonate filter.  Sampling duration ranged from 15 seconds to 7
minutes depending upon the expected loading.
          Fiber emission measurements have also been made for pelletizing oper-
ations (8).  Temperatures at the four locations encountered in Ref. 8 ranged
from  157 F to  270 F.  Deviations from Method 1 to 5 (9,10,11) included:  (a)
the use of a 115-mm cellulose acetate filter instead of a glass fiber filter,
 (b) maintenance of 180 F for the sampling probe and heated filter, and (c)
installation of a glass cyclone in the heated filter box ahead of the filter to
remove some particulate material.  A temperature of 180 F was chosen after deter-
ioration of the cellulose acetate material was detected at 200 F.
                                      14

-------
          Measurements of asbestos emissions from baghouse-controlled
sources have been reported (12),  The industries included two asbestos
cement plants, an asbestos textiles plant, and two asbestos mills.  An
extractive sampling system was used upstream of the baghouses.  Samples
were drawn through a cyclone prior to filtration by a 10 em, 0.8 vim
pore size membrane filter.  On the downstream side the cyclone was not
used.  In some instances sampling locations for extractive isokinetic
sampling were Inaccessible.  High volume samples with mem'.-rane filters
were used within the baghouse itself for the downstream measurements.  A
recent study (13) suggests sampling simultaneously using 3 filters at
different flow rates in an attempt to insure proper loading.
fugitive Emission Sampling
          Commonly used sampling strategies for measuring fugitive
emissions have been categorized (14,15) as:
          *  The quasi-stack method which involves capturing the entire
             emissions stream with an enclosure or hood and sampling
             these confined emissions with standard stack sampling
             techniques.
          •  The roof-monitor method which involves measurement of the
             emissions by traverses across well defined openings such
             as ventilators, windows, and access doors (16).
          •  The upwind-downwind method which involves measurement of
             upwind and downwind concentrations using ground based samplers.
             The source strength is calculated using a diffusion model and
             measured meteorological parameters.
          •  The exposure-profiling method which involves the direct
             measurement of particle flux downwind of a source by
             simultaneous multi-point sampling over an effective cross-
             section of the fugitive emission plume.  The sampling conditions
             must be isokinetic.
          Several devices have been used for monitoring airborne asbestos (17).
The most common method is high-volume filtration using cellulose ester membrane
filters (18,19).  An array of hi-vol samplers is commonly used to measure fugi-
tive emissions outdoors.
                                       15

-------
          Passive samplers "have also been used to collect particulate matter
for measurement of particle flux,  AB example is the isokinetic sampler
reported in Refs. 20 and 21 which collects particles electrostatically on
a metal foil as the air stream passes through the sampler under the air
stream's own inertia.  While such samplers meet environmental constraints,
an additional constraint on the amount of sample collected is imposed upon.
the sampler by virtue of its design.  That is, the sampling volume is
limited by the product of the effective cross section of the sampler and the
prevailing air velocity.  To obtain a measurement of airborne concentration,
as opposed to particle flux, a separate continuous record of local air
velocity must be maintained.
                                  16

-------
                               SECTION 5
                       CRITERIA AND CONSTRAINTS

CRITERIA
          A source sampling method for asbestos emissions must meet
certain requirements if it is to be accepted as an approved sampling
method.  Two standards upon which to base a judgment of acceptability
were determined at the onset of the program.  The first criterion is that
the sampling method must be capable of collecting a representative asbestos
size distribution from the local environment.  The second criterion is
that the asbestos must be collected in such a manner that it can be analyzed
by the provisional analytical method to provide the required determinations.

CONSTRAINTS
          A number of constraints, arising from different sources, restrict
potential sampling methods if they are to meet the two basic criteria.  The
establishment of these constraints provides the framework for the conduct
of the feasibility study on the development of a source sampling method.
General constraints identified at the onset of the program are presented
in Tables 3 and 4,  These constraints arise from several factors as shown
in Figure 1.
Constraints on the Acquisition of a Representative Sample
          Constraints on a method for the collection of a representative
sample arise from two basic areas:  (a) the required determinations and
(b) the characteristics of the particulate emissions and sampling environment.
Required Determinations—
          Health concerns have led to the establishment of national
emission standards for hazardous air pollutants.  The standard for asbestos
is contained in 40 CFR Subpart 61b,  The emission standard for the four
                                  17

-------
      TABLE 3.  CONSTRAINTS ON A SOURCE SAMPLING METHOD FOR ASBESTOS
                FOR THE ACQUISITION OF A REPRESENTATIVE SAMPLE
RSI  Ability to collect asbestos fiBrils and fibers over the diameter
     range 0,03 <_ <$£  _<10 pm for determination of number and mass con-
     centration by counting techniques.
RS2  Ability to collect asbestos fiber bundles over the diameter range
     0.2 iim to several tens of urn for the determination of number
     concentration by counting.
RS3  Ability to extract a sample from the local environment characteri-
     zed by air velocity, temperature, and moisture content.
RS4  Ability to collect a time-integrated sample.
                                    18

-------
                    TABLE  4.   CONSTRAINTS  ON A SOURCE SAMPLING METHOD
                                 FOR  ASBESTOS TO  BE  COMPATIBLE WITH  THE
                                 ANALYTICAL METHOD FOR ASBESTOS  DETERMINATIONS
             Compatible Kith the Provisional Method
          Strictly Compatible
          With Recommendations
                                    Compatible With
                                     Alternatives
                                 Compatible With
                               Electron Microscopy
AMI
AM3


AHA






AH5



AM6



AM?


AM8
The sample must be collected
  uniformly over a 0,4 urn
  pore size polycarbonate
  filter

The collection filter must
  have an asbestos loading
  in the proper range for
  counting

The collection method is
  air filtration

the collection medium is 0.4
  urn pore size polycarbonate
  filter material
The collection of non-
  asbestos patter must be
  minimized

Special care in tna handllni;
  of polycarbonate filters
  must be exercised
                                       The capability to take the collected sample, alter it
                                         (e.g., by ashing), and obtain a uniform dispersion
                                         on a  polycarbonate filter is required
                                       The capability of obtaining an asbestos  loading an a
                                         polycarbonate filter in the propar range for
                                         counting ie required
The collection method
  is air filtration
the collection method is not
  United to air  filtration
The collection medium      The collection medium is
  is 0.4 um polycarbonate    not specified;  however,  it
                            must he compatible  with  a
                            procedure to transfer the
                            collected asbestos  to an
                            em grid
                                   or cellulose ester
                                   filter material
The capability  to reduce the amount of collected non-
  asbestos material  (>3.g., by ashing) must be  available
Polycarbonate  filters ore not necessarily required  for use
  in the field
Fiber bundles must  be collected for counting


Count and equivalent volume determinations must be made
                          The counting at  fiber bundles is
                            not necessarily  required

                          The specific determination
                            is not specified

-------
     Industry
  Health/Environmental Concerns
        i
                1
   Particulate
  Emissions and
Sampling Environment
Standards/Required Determinations
                       1
                Constraints for
              Representative Sample
                 Sampling Method
                     Analytical Method
         Figure 1.   Factors leading to  constraints on a
                    source sampling method for  asbestos.
                                    20

-------
industries of concern (i.e., ore production, asbestos-cement,  asbestos
felt and paper, and asbestos friction products) specifies that there be
no visible emission to the outside air from any operations or  that the
emission containing asbestos be cleaned before such emissions  escape to
the environment,
          OSHA regulations for the work place environment specify an
exposure regulation for workers in 29 CPR Section 1910,1001.   The 8-hour
time-weighted average airborne concentration of asbestos fibers with
                                                           3
length greater than 5 pm is not to exceed two fibers per cm of air.
          As shown in Figure 1» these standards in principle specify
certain required determinations of asbestos.  For example the  OSHA regula-
tion requires that a count determination be made on the collected sample.
          Required determinations arise from sources other than codified
standards.  Based upon discussions with the EPA Project Officer and other
EPA scientists, determinations of asbestos have been defined which are
more stringent than those identified above.  These determinations are
compatible with the provisional method (1) for asbestos determinations.
                                              3
          The number of asbestos fibers per cm  of air must be determined
over a fiber size range including fibrils, fibers, and fiber bundles.  The
concentration of bundles is reported separately.  Classification of
asbestos into one of the following categories is made after electron
diffraction:
          •  chrysotile
          •  amphibole group
          •  ambiguous (incomplete spot patterns)
          •  nonasbestog
          *  unknown (no spot pattern).
          Determination of the length and diameter of fibrils  and fibers
is required for subsequent calculation of the mass concentration.  This
determination is not made for fiber bundles.
          Two constraints placed upon a sampling method by the required
determinations are listed in Table 3 as the first two constraints.  The
ability to collect asbestos fibrils and fibers for determination of
number and volume concentrations by counting places the following require-
ments on a sampling method.
                                  21

-------
          (a)  The overall collection efficiency for asbestos must be
known over the diameter range 0.03 pm to 10 urn.  Furthermore, the collection
efficiency must be such that combined with the variables of sampling rate
and ttoe and asbestos concentration, sufficient asbestos (based on number)
can be collected to provide good counting statistics for the analytical
determination.
          lor example, asbestos contributing to the majority of the number
concentration may not be concentrated in the same diameter size range as
the asbestos contributing to the preponderance of the volume concentration.
Per unit length, one fiber of 0.4 pm diameter will contribute 100 times as
much volume to the total volume as will a fibril with d, = 0.04 urn.  From
previous experience It can be expected that the fibrils Bill contribute
the most to the total asbestos number concentration; fibers will contribute
the most to the volume concentration.  Ideally, an asbestos sampling method
must be able to collect sufficient asbestos for both number and volume
determinations without Impairing one of the two determinations by collecting
too much material (e.g., too many fibrils for the number determination with
an appropriate loading for the volume determination) or too little material
(e.g., appropriate loading for the number determination with too few fibers
for the volume determination).  This ideal may, in fact, be very difficult
to obtain.  The number of fields which can be counted under the electron
microscope is small because of economic constraints to be discussed later.
This implies that if the same number of fields are to be counted for both
number and volume determinations, approximately the same number of fibrils
and fibers should be present for equal counting statistics for the number
and volume determinations.  Considerations of the effects of competing
constraints on selection of a sampling method will be discussed in the
summary of the requirements for a sampling method.
          (b)  No fiber size separation during sampling will be required
if sufficient numbers of fibrils and fibers can be collected simultaneously
by the same mechanism.  An alternative approach could consist of collecting
more fibers by selective concentration or sampling larger air volumes
while simultaneously collecting a second sample of fibrils.
                                    22

-------
          the second constraint in Table 3 provides for determination of the
number concentration of  fiber bundles.  If an accurate determination of the
number concentration of  fiber bundles is required, the sampling method must
be capable of collecting sufficient  quantities of these bundles for such a
determination.  The large bundle  sizes coupled with the wide size range of
such bundles suggests  that  a separate sampling method could be required to
collect  these bundles  efficiently over their size range.  Information on
the effective size of  such  bundles with respect to the collection mechanism
under consideration (e.g.,  aerodynamic diameter for capture on a surface
from an  air stream) is required in order to determine if an additional
sampling method would  be required for bundles.
          An additional  requirement  of a sampling train for the collection
of fiber bundles  is that the bundles be collected without fragmentation.
Fragmentation will increase the number of fibrils and fibers and, if
fragmentation is  complete,  reduce the number of fiber bundles.
Emissions and Sampling Environment—•
          As shown in  Figure 1, characteristics of the particulate emissions
and their local environment place additional constraints on a sampling
method for the acquisition of a representative sample.   The type of asbestos
emissions and their environments are dependent upon specific industries.
Within the four industries of concern material and process  variations
and control technology combine to establish emission and sampling environ-
ment characteristics.
           The following variables constitute the characteristics of the
 emissions and their environment which directly affect the choice of a
 sampling method:
           •  relative amount and composition of nonasbestos
              particulate matter
           *  type of asbestos
           m  asbestos concentration and size distribution
           •  presence of corrosive gases
           •  air flow dynamics—duct or stack flows vs.
              fugitive emissions
           *  temperature
           »  time dependence of the variables above.
                                    23

-------
          The sampling environment provides constraints on the ability of
the sampling method to extract a representative sample from the environment.
the third constraint in Table 1 requires that this be accomplished under
two basic conditions:  (a) in duct or stack flows and (b)  in open air for
fugitive emissions.  The relative extent to which asbestos emissions fall
into these two categories is determined by processes within the four
industries of concern.  This third constraint necessarily requires that
the sampling apparatus operate under local temperature conditions.
          The time dependence of the five variables listed above places
an additional constraint on a sampling method.  The fourth constraint in
fable 3 for a sampling method, the ability to collect a time-integrated
sample, is necessary in order to collect a sample representative of the
emissions over a period of time long in duration compared to the period of
parameter fluctuation.
          This constraint has further consequences for a sampling method.
Samples could be collected continuously or intermittently over a specified
time period.  As the length of sampling is increased to achieve time
integration, the sampling rate must be correspondingly decreased if the
same amount of asbestos is to be collected.  The ability to determine
accurate sample volumes of air must be maintained as the sampling rate
and/or sampling durations are reduced.
Combined Constraints for Representative Sampling—
          The combined constraints on a sampling system resulting from
characteristics of the source emissions, their local environment, and
the required determinations on the collected sample can be categorized
according to potential components of a sampling system.  These components
are;
          •  inlet system
          *  transport system
          •  collection system.
          The inlet system must capture a representative sample from the
air.  The sample is then transported to the collection system by the
transport system.  The inlet system may contain a precutter which
segregates particles according to a particular property (e.g., aerodynamic
                                  24

-------
diameter or electrical mobility for charged particles in a field) and
delivers particles or fibers with  specified values of that property to
the downstream portion of  the  inlet.  In the downstream portion of the
inlet system or  in the transport system provision can be made for
incorporation of dilution  air  into the sample flow stream.
          The collection system must he capable of handling a range of
incoming particulate concentrations.  Constraints placed upon the collection
system by the sampling environment and required determinations do not
limit the collection system to collection by a membrane filter.  Potential
collection strategies include  filter collection from the airstream, the
use of cyclones, electrical and thermal deposition, and collection by
impingers or impactors.  However,  the analytical method further constrains
the collection methods which can be utilized.
Constraints Arising from theAnalytical Method
          As shown in Figure 1, the analytical method for making the required
determinations places additional constraints on a source sampling method.
These constraints can be classified as technical and economic.
Technical Constraints—
          EPA desires that a source sampling method for asbestos be compatible
with the provisional analytical method reported in Ref. 1.  At the same
time it is recognized that the analytical method is only a tool,  subject
to change, used  to make certain required determinations.   The constraints
placed upon a sampling method by the requirement of compatibility are
presented in Table 4.   The compatibility requirement is broken down into
three alternative requirements.  These are (a)  strictly compatible with
the recommended procedures of the provisional method,  (b)  compatible with
alternative procedures (not considered optimal)  of the provisional method,
and (c)  compatible with electron microscopy but not necessarily with the
procedures of the provisional method.
          As shown in fable 4,  relaxation of the requirement of strict
adherence to the recommended procedures of the provisional method provides
for many potential alternatives for a source sampling method.  If the sampling
method is to remain compatible with alternative recommendations of the
provisional method, polycarbonate filters need  not be used for sample collection.
                                    25

-------
Sample preparation to reduce nonasbestos material Is permitted.   Dilution
of the sample to achie¥e a different loading is also possible.  Finally,
if the sampling method is only to be compatible with general electron
microscopy, the collection principle in the sampler need not be air filtra-
tion.
          Three practical concerns associated with EM analysis which influence
the selection of a sampling method are:  (a) uniformity of fiber dispersion
across the IM grid, (b) loading, and (c) nonasbestos interference.  If the
collected participate matter is directly transferred to the EM grid as out-
lined by the provisional method, these three concerns apply directly to the
filter collection.  Otherwise these concerns apply to the transfer process
resulting in a deposit on the EM grid.
          (a)  Only a small portion of the sample is viewed under the electron
microscope.  Consequently, it is vital that the portion of the sample selec-
ted for viewing be representative of the loading of the entire sample col-
lection.  For example, if the sample is collected on a polycarbonate filter
and transferred to an EM grid, the collected fibers should be uniformly
distributed across the filter surface.
          (b)  loading is an additional constraint.  The optimum loading on
an EM grid is in the range 10 to 20 fibers per 200-mesh grid opening (90 x
     2
90 UB ).  Ideally, the sampling method will provide such a loading regardless
of the conditions encountered.
          (c)  Nonasbestos material on the EM grid interferes with the micros-
copist's ability to distinguish the asbestos from other material.  Extraneous
material can be removed during sample preparation by ashing (22,23), Bonifi-
cation (22,23), or two-phase liquid separation techniques (24,25); however,
such a procedure may also alter the collected asbestos.  Fibers and bundles
can be broken apart into fibrils and asbestos losses can occur.  Therefore,
the sampling method should separate asbestos from the coexisting nonasbestos
particulate matter to the extent possible.
                                   26

-------
Economic Constraints—
          The provisional method is a relatively expensive analytical method.
Both the sophisticated equipment and the labor intensive nature of the
counting procedure contribute to this characteristic of the provisional
method.  As a result, a source sampling method should collect an asbestos
sample in such a manner as to minimize the labor required for the counting
and classification procedure.  This can be accomplished by (a) providing a
sample for counting as free from nonasbestos material as possible, and
(b) providing a sample with an asbestos loading in the optimal range for
counting.

IMPLICATIONS OF CONSTBAINTS ON FEASIBILITY
          The implications of the constraints identified in Tables 3 and 4 are
presented in Table 5.  The terms RS and AM refer to constraints arising from
representative sampling and the analytical method respectively.  The numerals
refer to the order of the constraints listed in Tables 3 and 4.
          The implications listed in Table 5 indicate that compromise is
necessary between practicality of sampling and the ideal requirements for a
collected sample imposed by the desired determinations of count and volume.
Three levels of compatibility between a sampling method and analytical
methods are presented in Table 4.  It is not feasible to undertake a methods
development program for strict compatibility with the recommended procedures
of the provisional analytical method.  However, methods development programs
are feasible if the sampling method is to be compatible with the alternative
procedures of the provisional method or general electron microscopy.
                                      27

-------
               TABLE  5.   IMPLICATIONS  OF  CONSTRAINTS


                        Implication for the Development of a Source
Constraint*3*                   Sampling Method for Asbestos

     RS 1           A single collector cannot be  used  to simultaneously
     AH 8           collect asbestos over the diameter range 0.03 to
                    10 pm and provide optimum loading  for both lUEsbeir
                    and volume concentrations by  counting.

     SS 2           The potential breakup of  fiber bundles must be jain—
     AM 7           inslzed by providing a short straight transport path
                    between the sampling inlet and the collector.

     RS 3           The difference in air velocity between stack and
                    fugitive emission environments necessitates the de-
                    velopment of at least two sampling techniques
                    designed for air velocities in the two types ef
                    environments.

     RS 3           Saturated conditions will be  encountered.  The
                    sampling system must be able  to collect samples in
                    these environments.

     RS 3           Elevated temperatures are not a constraint.

     RS 4           A continuous monitor to assess the level of asbestos
     AM 2           loading in the collector  is not practical,  A series
                    of sample volumes could be collected separately to
                    provide one with an acceptable loading.

     AM 1           Strict compatibility with the recoroEEendeil practices
                    of the provisional method is  not possible if collec-
                    tion methods other than air filtration by polycarbo-
                    nate filters are to be considered,

     AM 1           If the sampling method Is strictly compatible with
                    the provisional analytical method, the sampling rate
                    through the filters must  be within the range for opti-
                    mal filtration by a polycarbonate  filter.

     AN 1, 3, 4,    Direct air filtration or  filtration of a liquid con-
     6              fainlng collected asbestos is feasible.  Uniform
                    electrical deposition of  asbsstos  on a surface meeds
                    further research.

     AM 5           the size and chemical characteristics of the asbestos
                    and non-asbestos particulate  emissions preclude the
                    use of iuertial or magnetic forces in a sampling
                    system for material separation*  It is highly probable
                    that material separation  techniques vill need to be
                    used during aapple preparation.


(a) RS "• Constraint for representative campling,   SS 1 is the first
         constraint listed in Table 3,
    AH « Constraint: imposed by the analytical method.  AH 1 is the
         first constraint listed in Table 4.
                                   28

-------
                                 SECTION 6

                     FEASIBILITY OF METHOD DEVELOPMENT

     In determining the feasibility of developing a sampling method it is
desirable initially to view the method as a sum of generic components rather
than the system as a whole.  Using this approach key system components may be
identified and options reviewed and assessed.  The components may then he com-
bined, giving proper consideration to component compatibility, to generate the
most desirable complete method.
     An essential element in determining the feasibility of developing a stand-
ard method is a brief review of the state-of-the-art of the pertinent compo-
nents.  Such a review should aid in (a) determining specific areas where re-
search needs exist and where they do not;  (b) identifying and eliminating
specific component options and (c) properly focusing efforts directed toward
method development.
      For the subject task,  asbestos sampling,  the system can be viewed in four
 parts;
                    (a)   System Inlet
                    (b)   Transport Probe
                    (c)   Extraneous Material Separation
                    (d)   Collection Technique.
      Viewing the system in this piece-wise fashion not only facilitates feasi-
 bility assessment but also provide** an approach for determining in what areas
 state-of-the-art deficiencies lie and thus where development efforts should be
 concentrated.
      Table 6 gives a summary of the system components of interest.
                                      29

-------
                                     TABLE 6.   SYSTEM COMPONENTS
    Component
      Factors
    Influencing
        Options
          Deficiencies
            (Areas of
         Needed Research)
Inlet and Probe
Extraneous Material
  Separation
Collection Tech-
  nique
Sampling Environment
Relative size, concen-
  tration and physico-
  chemical properties
  of extraneous material
  and asbestos

Physical and Aerodynamic
  Properties of Fibrils,
  Fibers and Fiber
  Bundles
Stack Sampling
Fugitive Emissions
  Sampling

(1) Inertial Separation
(2) Other Mechanical
    Separation
(3) Chemical Treatment
(4) Pyrolytic Treatment

(1) Electrostatic
(2) Filter
(3) Imp.ir.ger and Filters
None, State-of-the-Art
  Sufficient
Physical Characterization
  (size distribution and
  relative concentration)
Collection Substrate Com-
  patability with
  Analytical Procedures

-------
INLET AND PROBE
          The inlet design is important to insure that proper representative
sampling is conducted.  This requires isokinetieally removing the airborne
asbestos emissions from their environment.  The characteristic sizes of the
asbestos fibers likely to be present (0.03 to 4 ym diameter) are compatible
with standard inlet designs (26-28) and isokitietic methods* (10,11).  Thus the
current state-of-the-art is adequate and no further development necessary.
          Likewise, for the probe design, current procedures are applicable
to the case of asbestos sampling (5,7 App. G).  The probe should transport the
sampled asbestos from the inlet to the collection medium or monitoring
instrument without disrupting the sample.  In many sampling instances a
heated or special noncorrosive probe is required, however the industry survey
(Table 2) reveals that for the case of asbestos emissions no extreme environ-
ments are likely to "be encountered.
     At this juncture it is also appropriate to mention that other monitoring
techniques necessarily associated with any standard sampling method such as
flow monitoring have been adequately developed for other methods and are appli-
cable to an asbestos method.  Thus no further treatment of such is given in
this report.
EXTRANEOUS MATERIAL SEPARATION
     Undesirable nonasbestos material (extraneous material) will be present
(see Table 1) in the sampling environment thus complicating the measurement
of the airborne asbestos.  Ideally one would like to remove the extraneous
material at the time of sampling to facilitate ease of subsequent analysis.
     Classically,  extraneous material has been removed by employing differences
in either physical or chemical form between the undesired material and the
material of interest.  For the case at hand, the broad size range of asbestos
present stretching from the diffusion dominated region (0.03 ym) to the
interial behavior region (4.0 urn) makes complete separation of extraneous
material from the asbestos impossible by traditional mechanical means such as
impactors or cyclones.  However if further investigation were to reveal a
     *0nly for the case of extremely long  (on the order of cm) or clustered
asbestos fibers, neither of which are likely emissions from the industries
considered, will standard particulate inlet considerations not be applicable.
                                  31

-------
large portion of the extraneous material to have a characteristic size larger
than 4 um then removal by an imertial device during sampling could he quite
beneficial.  The apparent nonhomogeneous form of the extraneous material makes
other types of separation (such as magnetic for metallic material) impractical.
Thus, at this time, it would appear that chemical or pyrolytic separation of the
nonasbestos material holds the most promise.  Such techniques are more appropri-
ately suited to analytical procedures than sampling procedures and as such are
not of concern within the scope of this study.
COLLECTION TECHHIQOES

          Because of their nonsphericity,  fibers pose a unique sampling
problem.  The physical behavior of asbestos fibers in air has been reviewed
elsewhere (28).  The short discussion which follows will be limited to con-
cerns directly related to collection and detection of asbestos.
          Generally speaking detection and analysis of asbestos may be
accomplished either by direct measurements or by collection on a substrate
coupled with subsequent analysis.  Whereas techniques of the former type
have advantages of real-time data gathering and ease of operation there are
serious drawbacks limiting their application to fiber detection.  With the
latter techniques problems may arise during handling and preparation of
samples for analyses.
Direct Detection
          Techniques classically used to directly measure particle size and/or
concentration include electrical mobility, diffusive mobility,  inertial sepa-
ration and optical analyses.  Of these only the optical techniques have been
pursued to a great extent for analysis of fibers.  Because of the nonspherical
nature of fibers and interference by nonfibrous aerosols, direct measurement
using electrical, diffusive or inertial techniques does not appear promising.
          Optical measurements of fibrous aerosols have been conducted with
some success though limitations do exist.  There is some evidence (2,29),
that fiber concentration can be measured with an optical particle counter.
However there may be errors associated with fiber orientation.   The presence
of nonfibrous particles is not considered.  Also the technique, as with other
optical techniques, is not applicable to fibers with diameters  less than
about 0.3 vm.  Lillienfeld (30) presents an optical instrument which is
                                      32

-------
designed to overcome fiber orientation problems (and thus nonfibrous aerosol
interference), but is still quite limited as to size of fiber detectable
and concentration range both maximum and minimum.
     Given the current state-of-the-art there is no instrument which is univer-
sally adequate for direct measurement of asbestos fiber aerosol.  Furthermore
development of such an instrument and subsequent incorporation into a standard
method does not appear feasible in the near term,
Fiber Collect ion
     The most prevalent and well tested methods of asbestos fiber detection
involve collecting the fibers and subsequently analyzing them by micro-
scope techniques.  The collection mechanism and substrate must be compatible
not only with the sampling situation but also with the analytical procedure.
Attention has been given to collection of fibers with various filter media,
electrostatic and thermal precipitation, impingers, and cyclones.

Thermal Precipitation—
        Thermal precipitation of asbestos fibers onto a suitable medium
is a possible collection mechanism although low efficiencies for long
fibers have been noted  (29).  The most detrimental characteristics of
this technique however may be the long sampling periods typically required for
adequate collection, on the order of months for ambient concentrations (2).
This drawback may be overcome by using a larger precipitation unit however
subsequent practical problems associated with using the precipltator and
handling the samples may result.  For these reasons thermal precipitation
does not appear fco be a promising collection technique.
Electrostatic Precipitation—
         Electrostatic precipitation suffers many of the same drawbacks
 as thermal precipitation for application to the subject sampling situation,
 however a recently developed instrument (21)  has given promise to using
 an electrostatic sampler in source emissions environments.  High
 efficiences (87-100 percent) were reported for several types of appli-
cations.  Although the  currently available commercial instrument would need
to be modified  (especially with regard to flow rate determination) it would
appear to be feasible for collecting asbestos fibers.  The use of such a
system has the advantage of collecting the sample on a cylindrical tube

                                     33

-------
(aluminum in the current instrument) which could conceivably be sealed in
the field for easy transport back to the laboratory without fear of losing
or disrupting the sample.  Compatibility of the collection substrate and
preparation of samples for analysis would be key issues in determing utility
of such an electrostatic collection scheme in a standard sampling method.
Collection by Impingers—
          Impingers may also be used to collect asbestos fibers.  By collecting
the fibers in liquid many of the handling and transporation problems associated
with filters are eliminated and the collection is highly suitable for most
analytical procedures.  In addition a greater volume of sample nay be collected
with an Impinger.  However the collection efficiency is poor for submicrojaeter
fibers, and therefore for the case of asbestos with many fine fibers and
fibrils, impinger collection alone is not appropriate.  However impingers may
be useful as a precollection method to avoid undesirable heavy loading on
high efficiency filters.
Air Filtration—
          fhe use of high efficiency membrane filters has traditionally been
the desired method for collecting asbestos fibers.  However, the use of
filters for collection and subsequent preparation for analysis is not without
problems.
          When sampling  asbestos with high efficiency filters the investigator
must consider the analytical procedure in making his selection.  Clearly glass
fiber  filters are unacceptable because of the possible ambiguity which may
result in viewing the asbestos fibers among the glass fiber substrate.  Of
the Gonmon  filter materials prominantly used in the 0. S. only  cellulose
ester  membrane or polycarbonate membrane filters are realistic  choices.
          The cellulose  ester filter  (a spongelike collection substrate) has
the advantage of superior  handling  characteristics compared to  polycarbonate
filters and has  a collection efficiency of nearly 100 percent for all size
fibers at all flow  rates.  The only disadvantages are that  the  pressure  drop
across the  cellulose ester is greater  than that for  the  polycarbonate at a
given  face  velocity and  some loss of  fibers is likely to occur  during prepa-
ration for  analysis.
          The polycarbonate filter  is  less than 100 percent efficient for
certain circumstances (2)  and  is  difficult to  handle in  field applications.
However the polycarbonate is roost suited  to  the  electron microscope analytical
procedure  (1).
                                     34

-------
          The constraints associated with EM analysis are important consid-
erations when filters are used to collect asbestos.  The three general con-
straints of uniformity of particle dispersion on the EM grid, loading, and
nonasbestos interference were discussed In Section 5.  With respect to
filtration, filter loading Is a real problem.  Complications will result
if the loading of the asbestos material is either too light (leading to
statistically invalid conclusions) or two heavy (making counting, sizing
and subsequent data analysis impractical for even the most patient micros-
copist).  They can be reduced during the collection phase by adjusting
sampling rates and times to achieve a loading in the optimal range.  This,
however requires a good deal of knowledge about the sampling environment.
If light or heavy samples are obtained, the only solution then lies in
concentrating or diluting the sataples as required during preparation.
These procedures are time consuming and add possibilities of further error
in the data.
Inertlal Collection—
          Cyclones are commonly used to collect particulate matter in
emission sources (31).  A cyclone cannot be used to collect fibrils or
the smallest fibers because of their small inertia.  Cyclones could be
used to collect fibers and fiber bundles if in the collection process,
the fibers and bundles were not broken into fine fibers and fibrils on
the cyclone walls.  Cyclones do not appear to be suitable for collecting
asbestos over the required asbestos size range.
                                   35

-------
                                SECTION 7
                         FEASIBILITY ASSESSMENT

     The feasibility of developing a method for asbestos sampling depends on:
           (a)  Developing an appropriate collection technlque(s),
           (b)  Determining  the ability to remove extraneous
                material,
      and  (c)  Evaluating the applicability of current inlet and
                probe technology to the selected collection technique.
After review of each of these elements it has been concluded that the develop-
ment of a standard method for sampling asbestos is feasible.  This study has
not uncovered any limiting industry anamolies or unsurmountable technical prob-
lems.  The sampling method would require different inlet and probe configura-
tions for fugitive and source sampling, respectively, however the same collec-
tion technique(s) should be applicable.
COLLECTION TECHNIQUES
     As indicated in previous sections the practical options for the collection
technique component are limited to either (a) electrostatic precipiation or (b)
membrane filter collection by cellulose ester membrane or polycarbonate
filters.  These  techniques may be  supplemented by  precollection with
an impinger  to reduce loading.  Table 7 summarizes the collection optivns with
the corresponding concerns associated with each and the advantages of each.
     Though a significant effort would be required to develop a standard method
with one of these techniques, such a development seems quite feasible.  This
statement of feasibility is supported by the field experience of Battelle and
others.  Ultimately however the feasibility of an actual method can only be
demonstrated through a development program.
ECrFANEQUS MATERIAL REMOVAL
     Past experience of analysts would indicate that it is feasible to separate
asbestos and nonasbestos material in the laboratory.  However the usefulness of

                                     36

-------
TAME 7.   SWKARV OP GOUSCTIOR OFTIOMS
Collection technique Arena at Concern
A. Electrostatic I. Adaptability of current comraercinl
Collector deaign to controlled flow design
2* CoBpatabillty off aluminum aab~
, •trace with analytical procedures
ft. FtlCer 1, Loading constraints la enviroRUwmt
(CuiluXofte of interest
•oter mcabir«nc)
2. Pvesauee drop aerons filter and
3. Preparation preecdurfia for su!>~
•nquene: EH analyaia

-------
as well as the feasibility of a separation during sampling can only be
assessed after more thorough data characterizing the industry emissions
is obtained and reviewed.
APPLICABILITY OF ISLET AND PROBE TECHNOLOGY
          The applicability of inlet and probe technology to the selected
techniques is feasible.  The demonstrated versatility of the developed
technology would lead to the conclusion that the applicability to an
asbestos sampling method is simply an engineering task for both fugitive
and source sampling applications.
                                   38

-------
                                SECTION 8
                     APPROACH TO SYSTEM DEVELOPMENT

          We recommend proceeding with a system development program on a com-
ponent-wise basis.  This would include mutually pursuing the R&D efforts out-
lined bslow for collection technique and extraneous material separation.  Sub-
sequently the most promising techniques should be incorporated with each
other and current state-of-the-art inlet and probe designs to form a sampling
system(s).  The integration efforts must be assessed under the constraints
discussed above.  Consideration must also be given to the practicality (with
regard to both engineering and utility aspects) of the system.  Finally, the
complete system must be laboratory checked and field demonstrated.  Figure 2
shows a flow chart for such a research program.
RESEARCH AND DEVELOPMENT FOR EXTRANEOUS MATERIAL REMOVAL
          Achievement of'the ability to separate extraneous material from
asbestos without altering the asbestos size distribution would be a major
breakthrough.  Separation during sampling does not appear promising.  Further
research on separation techniques should begin with the laboratory techniques
of ashing, Bonification, and two phase liquid separation to determine how
each of these techniques affects the collected asbestos size distribution.

RESEARCH AND DEVELOPMENT FOR COLLECTION TECHNIQUES
Electrostatic Collector
     A feasible asbestos sampling method might incorporate an electrostatic
collector to remove the asbestos from the air.  A device designed and tested
to perform such a task in certain circumstances is currently commercially
available (21) at a reasonable price giving promise to the possibility that
such a collection device could practically be incorporated into a standardized
method.  To do so however will require an appropriate research effort.
     The necessary research program would include several tasks:  (a) suitability
and/or adaptability of the isokinetic electrostatic sampler (21)  to the
                                     39

-------
             R&D
*•
o
Collection
Techniques

R&D *
Extraneous
Material
Separation

STATE-OF-THE-
ART
Probe &
Inlet Design






Component Compatibility*
Check

r — :
i
i
1
MrmTT?Tr>ATTriK(


^P^*w »•••« «
'

System Constraint
and Utility Assessment
t
Laboratory Operability*
t
Field Demonstration*
          Primarily Experiment Tasks
                                                                                    1
                                                                             Final Review and

                                                                                Assessment
                        Figure 2.  Sampling method development program flow chart.

-------
specific situations of concern; (b) compatibility of method with analytical
procedures; (c) compatibility of method with other system components and
(d) field test, calibration and verification of the method as part of the
final system.
     (a)  Suitability and/or adaptability of the sampler to airborne asbestos
collection.  This task would primarily involve investigation of (i) flow
monitoring possibilities of the current device, and (ii) loading and disruption
characteristics.  The current device is designed to sample isokinetically
with no provisions for flow monitoring.  As part of a standardized method,
it will be necessary to monitor the flow volume in order to subsequently
determine concentration levels.  This will require external monitoring of
the air flow around the sampler or adaptation of the sampler to a forced
flow sampling train through which gas flow can be monitored using traditional
methods.  The electrostatic principle has shown promise with regard to high-
collection efficiency in an emissions environment, however a complete R&D
effort should include as part of this initial task a laboratory determination
of the loading limits in the specific case of asbestos collection and a
specification of efficiency as a function of loading for the appropriate
electrostatic device.
     (b)  Compatibility with analytical procedure.  Concurrently with the first task
a study should be pursued to determine (i) the problems associated with
extracting the sample from the collector and preparing the corresponding
samples suitable for analysis, and (ii) the bias such a procedure generates
(e.g., agglomeration and clustering of fibers or break-up may occur during
collection and handling thus leading to unrepresentative results).
      (c)  Compatibility with other components.  This task would assess the
constraints placed on other selected components by the selection of electro-
static collection.  This task would include a laboratory demonstration of a
complete system employing electrostatic collection.
      (d)  Field tests.  As a final step in system development field tests
should be conducted using the laboratory proven system(s) and intercomparisons
(if appropriate) of the systems performance made.
Filters
     As pointed out by Spurny and StSber  (32) a need exists  to standardize the
filter type (if Indeed a filter collection is to be used) employed for asbestos

                                     41

-------
collection.  Ho experimental data are available to quantify concerns of
the field-worthiness, high efficiency and small handling losses of cellulose
ester membrane filters.  Therefore, since we are faced with a choice of two
filter media, cellulose ester or polycarbonate, a comparative research effort
should be undertaken to document their respective merits for each phase of
the required task; (i) collection efficiency (including loading constraints),
(ii) handling losses and (ill) preparation losses and biasing.  Such a task
would be experimental in nature aided to a certain degree by the past work
of Gentry, Spumy and others (2,32,33,34).  The selection of the filter medium
could then be made on a sound scientific basis.
          After the appropriate medium has been selected the research effort
should continue, as with the electrostatic collector, to include component
compatibility and field tests.
Impingers
          The use of Impingers will be appropriate only If loading is of
concern.  Thus further H&D regarding the usefulness of impingers for asbestos
collection should be deferred at this point.
                                    42

-------
                                  REFERENCES
 1.   Samudra, A.  ¥.,  C.  F.  Harwood, and J. D.  Stockham.   Electron Microscope
     Measurement of Airborne Asbestos Concentration—A Provisional Methodology
     Manual.  IFA-600/2-77-178, Revised 1978,  U.S.  Environmental Protection
     Agency,

 2.   Spumy, K.  R.,  J. W. Gentry, and W. StBber.   Sampling and Analysis of
     Fibrous Aerosol Particles.  In:  Fundamentals of Aerosol Science,
     D.  T. Shaw,  ed.   J. Wiley & Sons, New York,  1978.  pp. 257-324.

 3.   Control Techniques for Asbestos Air Pollutants.   Publication No. AP-117,
     U.S.  Government Printing Office, Washington, D.C.,  February 1973,  83 pp.

 4.   Proceedings of a Workshop on Asbestos:  Definitions and Measurement Methods.
     July 18-22,  1977, Gaithersburg, Maryland.  NBS Special Publication 506,
     U.S.  Government Printing Office, Washington, D.C.,  November 1978.   490 pp.

 5.   Standard Reference Methods for Source Testing:  Measurement of  Emissions
     of  Asbestos from Asbestos Mining and Milling Operations.  EPA l-AP-75-1,
     Air Pollution Control Directorate, Environmental Protection Service, Canada,
     December 1976.

 6.   U.S.  Environmental Protection Agency.  Standards of Performance for New
     Stationary Sources.  Federal Register, 43(37), February 23, 1978.  pp.
     7584-7596.

 7.   Clayton Environmental Consultants, Inc.  Iron Ore Beneficiation—Emission
     Test Report—Reserve Mining Company, Silver Bay, Minnesota. EMB Report
     78-IQB-5.  U.S.  Environmental Protection Agency, Research Triangle Park,
     North Carolina,  1979.

 8.   Clayton Environmental Consultants, Inc.  Source Testing Study at an Iron
     Ore Beneficiation Facility.  EMB 76-10B-2, U.S.  Environmental Protection
     Agency, Research Triangle Park, North Carolina,  1976.

 9.   U.S.  Environmental Protection Agency.  Standards of Performance for New
     Stationary Sources.  Federal Register, 43(160),  August 18, 1977.   pp.
     41776-41782.

10.   U.S.  Environmental Protection Agency.  Standards of Performance for New
     Stationary Sources.  Federal Register, 43(160),  August 18, 1977.
     pp. 41755-41758.
                                      43

-------
11.  U.S. Environmental Protection Agency.  Standards of Performance for New
     Stationary Sources.  Federal Register, 43(160), August 18, 1977, pp.
     41758-41768.

12.  Harwood, C. P., D. K. Qestreich, P. Siebert, and J. B. StocWiam.
     Asbestos Emissions from Baghouse Controlled Sources.  Am. Ind. Hyg.
     Assoc. J., 36:595-603, 1975.

13.  Yamate, George.  Evaluation of Provisional Method for Measurement of
     Asbestos, IITRI Report No. C-547Q-7, 30 November 1980.

14.  Kallka, P. W., R. E. Reason, and P. T. Bartlett.  Development of
     Procedures for the Measurement of Fugitive Emissions.  IPA-600/2-76-284,
     U.S. Environmental Protection Agency, December 1976.

15.  Cowherd, C., Jr.  Measurement of Fugitive Participate,  In:  Second
     Symposium on Fugitive Emissions:  Measurement and Control, May 1977,
     Houston, fexas, EPA-6QO/7-77-148, U.S. Environmental Protection Agency,
     December 1977.  pp. 47-62.

16.  Souka, A., R. Marek, and L. Gnan.  A New Approach to Roof Monitor
     Particulate Sampling.  Air Pollut. Control Assoc. J., 25s397-398, 1975.

17.  Joint ACGIH-AIHA Aerosol Hazards Evaluation Committee.  Background
     Documentation on Evaluation of Occupational Exposure to Airborne Asbestos,
     Amer. Ind. Hyg. Assoc. J., 36:91-103, 1975.

18.  Stinson, M. K., C. F. Harwood, and P. Ase.  Asbestos Waste Emission
     Control.  In:  Second Symposium on Fugitive Emissions:  Measurement and
     Control, May 1977, Houston, fexas, EPA-600/7-77-148, U.S. Environmental
     Protection Agency, December 1977.  pp. 187-203.

19.  Henry, W. M., R, E. Heffelfinger, C. W. Melton, and D. L. Kiefer.
     Development of a Rapid Survey Method of Sampling and Analysis for
     Asbestos in Ambient Air.  Battelle-Columbus Final Report.  Contract
     CPA 22-69-110.  Report APTB-0965, 1972.  34 pp.

20.  Steen, B.  A New Simple Isokinetic Sampler for the Determination of
     Particle Flux.  Atmos. Environ., 11:623-627, 1977.

21.  Steen, B., P. B. Ready, and G. J. Sem.  A Sampler for Direct Measure-
     ment of Particle Flux, TSI Quarterly, 7:3-9, 1981.

22.  Samudra, A. V.,  F. C. Bock, C. F. Harwood, and J. D. Stockham.
     Evaluating and Optimizing Electron Microscope Methods for Charac-
     terizing Airborne Asbestos.  EPA-600/2-78-038.  U.S. Environmental
     Protection Agency.  June 1978.

23.  Bishop, K., S. Ring, R. Suchanek, and D.  Gray.  Preparation Losses
     and Size Alterations for Fibrous Mineral Samples.  In Scanning Electron
     Microscopy, 1:207-212, 1978.

-------
24.  Melton, C. W.» S. J. Anderson, C. F, Dye, ¥, E. Chase, and
     R. E, Heffelfinger.  Development of a Rapid Analytical Method for
     Determining Asbestos in Water.  Battelle-Columbus Final Report.
     Contract No. 68-03-2199.  Report EPA-60D/4-78-066, 1978.  67 pp.

25.  Speil, S. and J. P. Leineweber.  Asbestos Minerals in Modern
     Technology.  Environ. Res., 2:166-208, 1969.

26.  Davies, C. N.  The Entry of Aerosols into Sampling Tubes and Heads.
     Brit, J. Appl. Phys., 1 (2):921, 1968.

27.  Knapp, K, T.  The Effects of Nozzle Losses on Impactor Sampling.
     Proc. Advances in Particle Sampling and Measurement,  Jan. 1980.
     EPA-600/9-80-004.

28.  Stb'ber, W.  A Note on the Aerodynamic Diameter and the Mobility of Non-
     Spherical Aerosol Particles.  Aerosol Science,"2:453-456, 1971.

29.  Addingley, G. G.  Asbestos Dust and Its Measurements.  Am. Occup. Hyg.,
      9:73-82.  1966.

30.  Lilllenfeld, P.  GCA Fibrous Aerosol Monitor Model FAM-1 Users Manual.
     GCA, July 1978.

31.  Smith, W. B.» P. R. Cavanaugh, and R. R. Wilson.  Technical Manual:
     A Survey of Equipment and Methods for Partieulate Sampling in
     Industrial Process Streams.  EPA-600/7-78-043, U.S. Environmental
     Protection Agency, March 1978.  281 pp.

32.  Spurny, K. R., and W. Stober.  Some Chemical Characteristics of Ambient
     Air Fibers,  Proc. 3rd Internet. Colloq. on Dust Measuring Technique and
     Strategy, June 10-12, 1980.  AIA., November 1980.

33.  Gentry, J. W.  University of Maryland, personal communication,  March 1981.

34.  Spurny, K. R., J. P.  Lodge, E.  R. Frank, and D,  C.  Sheesley.  Aerosol
     Filtration by Means of Nuclepore Filters,  Structural  and Filtration
     Properties.  Environ. Sci.  & Tech., 3 (5):453-464, 1969.
                                   45

-------
                              APPENDIX A
                     DESCRIPTION OF TOE INDUSTRIES

ASBESTOS ORE PRODUCTION
          Asbestos production can be divided into the two phases mining
and milling.  Mining operations include (1) drilling to place explosives,
(2) blasting, (3) surface scraping, (4) sorting, (5) screening, (6) conveying,
(7) shoveling, (8) transport by truck, and (9) dumping (A-l).
Asbestos Mining
      ma  Two open pit mining operations were visited in California.
At the first mine blasting is conducted every other day.  Slurry explosives
are used to reduce dust generation.  The explosives are detonated in 6
inch diameter, 33 inch deep holes.  Moisture and asbestos content are
about 10 and 3.8 percent respectively.
          The ore is  transported from  the mine  to the ore preparation area
where it is fed into a series of crushers and screens.   Initially a law crusher
reduces the size of the  ore to 5 inches in combination with  screens.  A
water spray is used at the discharge of the jaw crusher to reduce  the dust.
The ore then  travels  to  a cone crusher where  it is reduced to 1-1/8 inches
and finally an impact crusher in which it is  reduced to 0.5  inches.  At  this
point ore over 0.5 inches in diameter  is discarded.  The remaining ore is
transported via a 1500-ft covered  conveyor to the mill site.  Three baghouses
are used to control the  dust in the ore preparation stage.   Dust collected
by the baghouses  is mixed with water in a screw conveyor to  form a paste
which is conveyed to  a nearby disposal site.
          At  the  second  mine moisture  and asbestos concentrations  are 16-18
and 60 percent.  The  ore is screened to one inch at the mine before the  60
mile  transit  to  the mill.

-------
Asbestos Milling Operations
          Milling operations consist of extracting asbestos from the ore,
cleaning, and grading the asbestos.  Asbestos is separated from the rock
by means of crushing the ore to liberate the asbestos and then extracting
the asbestos by aspiration over vibrating screens.
          In the first milling operation, which was visited, the moisture
content of the ore is reduced to less than two percent by either a vertical
dryer or a nearly horizontal rotary kiln.  The exhaust temperature from the
dryers is about 250 F.  The emissions from the dryers are controlled by
baghouses.  After the ore is dried it is conveyed to an enclosed storage
area kept under negative pressure.
          The finely crushed and dried ore is conveyed from the storage
area to the mill where the asbestos fibers are separated from the coexisting
rock by means of a series of vibrating screens, fiberizers, and shaker
screens.  The screens are fitted with aspiration hoods that entrain the
asbestos into an air stream which  then flows through cyclone collectors.   The
cyclones grade the fibers into three classes:  short, medium, and long.
The rock is expelled to an exterior tailings dump.
          The asbestos fibers are machine packaged in a hooded area.  The
smaller fibers are compressed into dense bundles, while the longer fibers
are blown into containers and loosely packed to mininize fiber damage.
Two types of bags are used—multi-ply paper and reinforced plastic bags.
          Emissions from the milling operation are controlled by baghouses.
The baghouse catch is transported to a belt conveyor via enclosed screw
conveyors and chutes.  The belt conveyor deposits the material into a
mixing screw conveyor which discharges wetted waste onto belt conveyors
for transport and disposal.  A dust suppressant chemical is added at the
mixing screw conveyor.
                                      47

-------
          A wet milling process is used at the second milling operation
which was visited.  As the ore enters the mill, it is slurried by spraying
with water as it passes over a 1/4 inch screen.  The fine fraction is then
passed through a cyclone separator and through a series of fiber opening
and separation stages.  The iron oxides are removed by magnetic separators.
          the separated asbestos slurry is filtered through a large filter
press system containing six banks, each about 30 feet long.  Shriver filter
presses with 48 x 48 inch plates operating at 80 psi pressure remove the
water.
          The filtered asbestos is then extruded through 1/2 inch orifices
and conveyed into a dryer.  A knife blade cuts the pellets into about one
inch lengths.  A rotary dryer with concurrent air flow is used to dry the
pellets.  The dryer operates at 1200 F, with an exhaust temperature of
approximately 250 F.  Some of the pellets are broken in mills to release
the fibers, others are shipped in pellet form.
          The primary control is via three baghouses, each of which
possesses an exhaust duct of sufficient length to permit appropriate
sampling.  The exhaust gas from the dryer is hot (^250 F) and contains
moisture, necessitating the use of insulated baghouses to avoid condensation
as the gases cool.  The bags are the pulse-air cleaning type.
          The waste from the baghouses and the tailings from the milling
operation are conveyed to the dump site in a wet condition.  When the
waste reaches a pre-determined level, it is covered with a foot or more
of top soil and seeded with rye.

TACONITE PRODUCTION

          Taconite  production was considered in this program because
amphibole fibers  contained in the ore are released from tbe ore as it
is crushed and  further processed.
          Taconite  production activities can be broadly classified into
four  areas:  mining,  beneficiation,  agglomeration, and handling of taconite
and  tailings.   Each of the activities  is briefly described below.
                                   48

-------
                  CONCENTRATOR  PLANT
                                                                                      O ••"'"'
                                                                                        tO p>«LkLfl]>»
                                                                       Ftirea PLANT  a1—s^r*"
                                   PELLET STORA.se AREA
Figure A-l.  Frocess flow diagram  for taconite production.

-------
          Mining operations include drilling, blasting, removing, and hauling
rock.  Three principal types of drills are used;  jet piercers, rotary drills,
and percussion drills.  Figure A-l is a diagram of the specific process
visited under the current program.  In this mine a jet piercer is used.
Emissions from the mining process are uncontrolled,

Beneficiation
          Iron ore beneficiation includes crushing, grinding, and concentrating
operations.  Crushing circuits usually consist of primary, secondary, and
tertiary crushers combined with screens to separate the desired fraction.
          Grinding circuits consist of rod mill-ball mill combinations
or autogeneous mills.  Generally water is added at this point and the
material is handled as a slurry through the concentrator.  In both cases
the material is subjected  to size classification by screens or cyclones
and concentration by magnetic separators, gravity separators, flotation,
or some combination of these methods.
          Typically the rod mill discharge is pumped to magnetic separators
referred to as cobbers.  The nonmagnetic material is discarded to the
tailings.  Some plants classify this material to separate the coarse
tailings from the fine tailings.  These plants generally use the coarse
tailings for dike construction or road building material whereas the fine
material is pumped directly to the tailings thickeners.  The magnetic
fraction is pumped to primary ball mills and ground further; the ball mill
discharge is then pumped to classifying cyclones with  the coarse material
(underflow) returning to the ball mill and the fine material (overflow) pumped
to rougher magnetic separators.  The rougher magnetic  separators discard the
nonmagnetic fraction to  the tailings  and  the magnetic  fraction is ground to  a
finer  size in secondary ball mills.  The secondary ball mills discharge to
classifying cyclones with  the coarse material returning to the rougher
magnetic separators or secondary ball mills.  The finer material is
pumped to either a dewatering device such as hydroseparators,  thickeners,
siphonsizers, etc., or magnetic separators.  In some plants there are
cleaner and finisher magnetic separators.  In essence, some plants use
two  stages of magnetic separation; others use three or four stages of
magnetic separation.  In addition to the above flow schemes, some plants

                                     50

-------
use gravity separation devices such as jigs, spirals, heavy media separators,
flotation, or various combinations of these techniques.  Also some plants are
using autogeneous grinding circuits instead of the conventional methods
described above.
          The final product from the beneficiation section is a filtered
iron concentrate containing approximately 9 percent moisture, 60 percent
iron, and 2 to 7 percent silica with a size ranging from 80 percent minus
0.0445 mm (325 mesh) to 90 percent minus 0.0127 mm (500 mesh).  This material
is transferred to large bins which feed the agglomeration section.
          At the site visited in this program beneficiation operations take
place at the fine crusher, dry cobbing, a* d concentrator plants.  When the
ore cars arrive at the plant, they are automatically fed into a rotary
dumpster at the car dumper plant as shown in Figure A-l.  The dumpster dumps
two 85 long ton railroad cars simultaneously without uncoupling.  The process
is capable of dumping 8,000 long tons per hour of -4 in. taconite.  The ore
is dumped into a large hopper below the railroad tracks and processed through
a pan feeder on the way to the fine crusher plant.  Dust is generated by the
dumping operation and the pan feeder.  Dust control from this operation is
achieved by drawing the air in the hoppers and pan feeder through a baghouse
located near the roof of the plant.
          From the car dumper plant the ore is conveyed to the fine crusher
plant where it is further crushed and screened before entering the dry
dobbing plant.  The coarse tailings are separated from the ore at this
location.  Up to this point in the process all operations have been "dry"
and dust control is achieved by drawing the surrounding air through bag
houses.
          When the ore enters the concentrator plant, water is added for
the first time.  The slurry then passes through a series of rod mills,
magnetic separators, sump concentrators, and primary and secondary
hydroseparators.  Fine tailings are removed during each operation.
          The ore slurry is pumped  to  the filtering plant where the final
tailings are removed.  This is accomplished via several large concentrate
thickener tanks, slurry tanks and a large disc type filter.  Bust is
controlled in the concentrator plant and filter plant by cyclone collectors.

-------
Agglomeration
          Agglomeration operations in the taconite industry produce sinter
and pellets.  Sintering causes the fine particles to bond together into
porous agglomerates which are strong enough to diminish dusting problems
but porous enough to permit good gas dispersion through a bed or the material
in a furnace.  Pelletiziag operations form balls in the diameter range
0.95 to 1.27 cm.
          As shown in Figure A-l, in the pelletizing plant Bentonite clay
is nixed with the ore in a balling drum.  The material is then screened
and fed into a pelletizing machine.  The pelletizer is gas fired and
operates at 1700 F.  As the pellets pass from the pelletizer, they are
screened (vibrating type) and conveyed to an outside pellet storage area.
The air from the pelletizing machine is exhausted into a wet electrostatic
precipitator unit.
Handling of Taconite and Tailings
          Taconite pellets are conveyed to large storage piles near the
plant to await transport.  Dusting from these piles and from ventilator
stacks of loading silos constitute potential fugitive emission sources.
          Tailings are conveyed to a disposal site by rail and water
slurry pipelines.  At the site visited in this program, tailings are
dumped into a man-made lake to minimize fugitive emissions from erosion.

ASBESTOS-CEMENT PRODUCTS
          The largest single use of asbestos fibers in the U.S. occurs in
the manufacture of asbestos-cement (AC) products of which the AC pipe
industry is the largest segment.  A flow chart for the production of AC
pipe is shown in Figure A-2.
          Two AC pipe plants were visited.  The first plant contains two
production lines.  The AC pipe is made from a blend of asbestos, Portland
cement, and sand.
          The asbestos used is primarily chrysotile and is a blend from two
different sources—South Africa and Canada.  No U.S. asbestos is used in
the process.  The asbestos is shipped to the plant in 100 Ib plastic bags.
Upon arrival each bag is inspected and repairs are made immediately if

                                   52

-------
                                                FLOW CHART OF A-C PIPE MANUFACTURE
1. Production Una Una—raw materials enter the lino as follows;
      |a) aibMtos—from the willow where the fiber is separated into
         individual strands and thoroughly mixed
      (b) cement—directly from receiving hoppers
      (c) silica—'from grinding mill
2. Electronic Scatot—lor precise weighing; accurate control for uniform
   results
3, Wet Minor—blends raw materials thoroughly
4. Convoying Trough—water carries stock 10 wet mix vat
5, Wot Mix Vat—thorough dispersion of reinforcing fibers
0, Swsari Cylinder Mold—picks up slurry and deposits on moving felt
7, Vacuum Box—excess water removed
8. Fell deposit! (lock on Mandrel—wall thickness «jyiH up under pressure
   to proper size
 9.  Mandrel (with pips)—removed from machine; next mandrel positioned
10. Looigner—frees pipe from mandrel; prevents distortion
11, Slow-down Conveyor—provides pro-cure nmo; initial set
12. Mandrels—removed and pipe stencilled for identification
13, Air Cure Room—strict controt of time, temperature and humidity
14. Autoclaves—high pressure steam curing imparts maximum strength and
    excellent chemical stability
1 B, Lathes—trim and machine ends to enact dimensions
16, Testing Equipment—checks for adherence to rigid specifications
     (a) flexure testing machine
     (b) inspection
     (c) hydrostatic looter
     (d| crush tester (laboratory)
17, Materials Handling Equipment—transfers pipe to shipping area
                                   Figure A-2.   Process flow diagram for A-C  pipe  production.

-------
required.  The bags are opened in a "white room" and dumped onto a conveyer
(under a hood exhaust) which transports it to a willows machine.  The wil-
lows machine breaks open the asbestos fibers and fluffs them in order to
achieve better mixing.  The asbestos is then conveyed to a large holding
bin from which it is dry mixed with the cement and silica.  The blending
formula varies with the type of pipe being manufactured.  The empty plastic
bags (which contained the asbestos) are placed into a larger plastic bag,
sealed, and labeled "asbestos hazard".  The larger plastic bags are subse-
quently autoclaved shrinking them into small bundles.  The bundles are then
disposed of at the waste disposal site.
          After the raw materials are dry mixed, a homogeneous slurry is
formed by the addition of water.  The slurry is delivered to cylinder vats
for deposition onto horizontal screen cylinder molds.  A thin layer of
asbestos-cement is formed on an endless felt conveyer.  After partial
drying the sheet is wound around a mandrel into pipestock of the desired
thickness.  The pipe section wrapped around the mandrel is removed from
the machine and then freed (loosened) from the mandrel by an electrolytic
loosener.  A one-hour precure time is provided by a very slow moving
conveyer before the mandrel is removed.
          After the mandrel is removed the pipe is stenciled for identifi-
cation and transported to a temperature-humidity controlled air-cure room
where it remains for approximately 12 hours.  Final curing is achieved in
one of seven high-pressure autoclaves.  The autoclaves operate at 120
psi, under live steam at 340 F.  The process takes about 15 hours—four
hours to reach steady-state, eight hours soak time, and three hours to
cool.
          After autoclaving, the pipe is fed into an automated lathe
and both ends are simultaneously machined to ensure proper mating with
connectors.  This operation takes approximately 15 seconds, and produces
large quantities of dry, asbestos-containing waste.  Rubber gaskets and
couplings are added at this point, and the pipe subjected to a series of
tests on the following machines:  flex testing machine, hydrostatic
tester (500 to 750 psi), and crush tester.  Pipes passing the above
tests are transferred to the storage and shipping area.
          All machining operations at the plant are hooded, and the
exhaust gases vented to a central baghouse.  In addition, each machine

                                    54

-------
 is supplied with a three inch vacu-m line which is used to thoroughly clean
 the machines at the end of each shift.  The baghouse waste is sprayed with
 water and collected in metal dumpsters.  Each dumpster holds approximately
 4,000 Ibs of waste and they collect between 16 and 18 loads every 24 hours.
 The waste is trucked to a landfill disposal site located on the premises.
           At the second AC pipe plant, both chrysotile and crocidolite are
 used in manufacturing AC pipe.  The chrysotile which is used comes from
 California and Quebec.  The crocidolite is imported from Soath Africa.  The
 asbestos is shipped to the plant in plastic bags.  The bags are opened in  a
 "white room", approximately 8 ft wide by 10 ft long.  The asbestos is dumped
 onto a conveyer, under an exhaust hood, and transported to a holding tank
 where it is mixed with cement and silica.  The blending formula varies with
 the type of pipe being made.  The blend is obtained by dry mixing, and then
 fed to a large tank where a slurry is formed by the addition of water.  The
 slurry is distributed via a 13 ft wide trough into two vats where two thin
 layers of felt are formed and simultaneously wrapped around a mandrel.  When
 the desired thickness is obtained, the pipes pass progressively through two
 curing ovens.  The first oven has a temperature gradient of approximately  350 F
 (front) to 250 F (back).  The second oven is controlled at 140 F.  From the
 curing ovens, the pipe is loaded onto carts and placed into one of three large
 autoclaves.  Autoclaving takes approximately 12 hours—one hour up,  one hour
 down, with a 10 hour soaking period.
           After autoclaving, the pipe is fed into a lathe (automated
 system) and both ends are simultaneously beveled.  This operation takes
 approximately 15 seconds.  The pipe is then hydrotested at 525 psi,  a
 rubber gasket is added and the pipe is ready for shipment.
           In addition to the main process line, there are several smaller
 process areas which are primarily made up of coupling lathes and cut-off
 saws.  Above the pipe forming line there are two large roof exhaust fans.
 These exhaust fans, plus a 3 ft, square exhaust duct between the curing
 ovens, are primarily heat removal systems.
           The smaller process areas, scattered throughout the plant, each
 contain a hood exhaust system.  The exhaust gases are vented, via a
series of ducts, to two dust collectors (baghouse type with shaker cleaning).
The baghouses are automated and incorporate a 2 hour and 15 minutes cleaning
frequency.  The waste is collected in plastic bags, sealed, labeled as
                                    55

-------
 asbestos hazard and stored.   A major portion of  the waste  is  recycled back
 into the system.   However, when  the amount of waste exceeds the storage
 capacity it is transported to a  waste disposal site and  covered with top soil.
 The water and solid-waste are both recycled back to the  process whenever
 possible.  Excess waste is bagged and disposed of at  the waste-site.
           Three potential sites  exist at  this plant for  sampling controlled
 asbestos emissions.   These sites are (a)  the baghouse exhaust ducts, (b)
 the exhaust duct between the curing ovens,  and (e)  the roof top exhaust
 fans.  There are two types of fugitive emissions which could  contain asbestos:
 (a)  the  autoclave exhaust and (b)  ambient airflow through  openings in the
 building structure.

 FRICTION  PRODUCTS
          Major categories of  production  processes  for asbestos friction
 products  are:  (a) dry-mixed  and wet-mixed, molded  brake linings,
 (b) wet-mixed,  two-roll  forming brake  linings and clutch facings,  and
 (c) woven, wire-reinforced brake  linings  and  clutch facings (1)-   The
 production processes  can be  segmented  into  the following general operations:
 mixing,  forming and  processing, curing, and  finishing.
          As implied by the name, mixing of input streams for the
first production process is accomplished either by dry mixing of asbestos,
friction material (e.g., aluminum oxide), and bonding agent or wet mixing
in blenders.  Input streams for wet nixing in the second process include
asbestos, friction material, and solvents.  Mixing is accomplished in a
blender.  The input streams for the third process Include wire-reinforced
woven tape or cloth, asbestos, and a friction material bath.   Mixing is
accomplished by running the continuous strip of tape or cloth through a
bath.
Forming and Processing
          Operations encountered in the molding process include preforming
in a press, cutting and grinding into flat blanks, steam heating to
soften the resin, and bending in presses.  Operations encountered in
the second production process utilizing roll-forming include forming a
continuous sheet of material in the two-roll mill (similar in concept
                                  56

-------
to paper production), chopping or punching, drying, and bending when
required.  Operations encountered in the third production process include
drying the roll after it has passed through the bath, pressing, and cutting.
          Curing is accomplished by baking ovens in all three production
processes.
Finishing
          Finishing operations vary according to the process and product.
fhese operations generally include sanding, grinding, drilling, dusting,
inspecting, and branding.

Emissions
          Potential emissions of asbestos are most likely to occur in the dry
mixing operations and in the finishing operations.  Control technology for
these operations frequently involves fabric filtration with baghouses.  The
finishing operations in  the friction plant visited in this program consist
of  cutting, grinding, sanding, drilling, and dusting.  Emissions are con-
trolled by passing the surrounding air through baghouses.  Nonasbestos
material includes resin, graphite, and carbon black.  The waste from the
baghouses is  transported to a pelletizer by means of a screw conveyor.
The waste is  mixed with  cement and water in the pelletizer, and the resulting
pellets are used as landfill.

ASBESTOS FELT AND PAPER
          In  general, asbestos paper is produced by first mixing asbestos,
binder, pulp  and water into stock for subsequent handling.  Typical binders
are starch, glue, water  glass, resins, latex, and gypsum (A-l).  A thin
uniform layer of stock is deposited onto a screen and subsequently dried
and pressed between rolls.  The continuous sheet then passes over heated
rolls and calender rolls.  The paper is cut to size as it is wound onto  a
spindle.  Potential emissions arise from the handling of asbestos as it
enters the process.  Baghouses are used to control these emissions,
          A block diagram of the asbestos roofing felt plant which was
visited is v- -t of the industry survey is  shown in Figure A-3,  The manu-
facturing process closely resembles the general process described above.
                                   57

-------
    Asbestos    Water
 Other
Material
                                                                                              Dryers
U1
00
                  Winding
                    and
                 Transport
                                        Transport
Winding
  and
Slitting
                                   FIGURE A-3.  BLOCK DIAGBAM Of ASBESTOS FELT
                                                MANUFACTURING OPERATIONS

-------
At the slitting operation the full reel of felt is cut down to exact
size rolls (by width) with slitter knives.  The dust and excess trim is
sucked into pipes and blown to a cyclone.  The trim separated by the
cyclone is returned to the beaters to reenter the process.  The dust
escaping the cyclone goes to a baghouse.  The dust that is captured in
the baghouse is also returned to the beaters.
          Emissions from the beaters are controlled by a wet impinger.
The impinger collects dust in the moist air stream by injecting a water
spray into the air stream.  The dust particles are incorporated into the
larger water droplets through collisions.  These water droplets are then
removed from the air stream by impaction.  Material captured by the
impinger is returned to the beaters.  The exhaust from the impinger unit
is saturated with water vapor.
          The final operation in the process involves saturating the
continuous roll of felt with asphalt in a hot asphalt bath.  Emissions
from this operation pass through a filter to remove organic vapors.  The
filter is a continuous roll of fiberglass passing through the exhaust
duct of the saturator.  The filter is not designed as a particulate emission
control device.
                                   59

-------
                              REFERENCES
A-l.  Control Techniques for Asbestos Air Pollutants.  Publication Ho.
      AP-117, U.S. Government Printing Office, Washington,  D.C.,
      February 1973, 83 pp.
                                60

-------