United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-85/044 Aug. 1985
SERA Project Summary
Environmental Release of
Asbestos from Commercial
Product Shaping
Dennis A. Falgout
Abstract
For the first time, the release of res-
pirable asbestos fibers has been quanti-
fied in terms of standard mechanical
forces using widely accepted method-
ology and specified QA/QC procedures.
Both fabrication of new products from
asbestos containing materials and re-
pair or removal of in-use asbestos con-
taining products contribute to the total
environmental exposure to asbestos.
There is a need to assess these materi-
als and operations with respect to the
potential severity of their fiber releases.
This research consisted of performing
several simulated industrial/commer-
cial shaping operations on several as-
bestos containing products. The rates
of fiber release, expressed as fibers per
cubic centimeter of air inside an en-
closed test chamber per gram of as-
bestos milled, were measured. The fil-
ter samples were analyzed by the
transmission electron microscope
ITEM) method. Lengths, widths, and
type of asbestos were reported for
fibers and other asbestos structures. In
addition, samples were taken for phase
contrast microscopic (PCM) analysis
during most of the experiments. The re-
sults of these analyses are compared.
Research on the release of asbestos/
substitutes resulting from commercial
product manufacture, use, and disposal
is of continuing importance. More infor-
mation about the quantities and dimen-
sions of fibers released during these ac-
tivities is required in order to develop
effective control methods to help pro-
tect the public health.
This Project Summary was devel-
oped by EPA's Water Engineering Re-
search Laboratory, Cincinnati, OH, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
The scientific community is in general
agreement that exposure to asbestos
dust increases the risk of: (1) asbesto-
sis, a fibrotic disease of the lung
whereby imbedded dust fibers are sur-
rounded by scar tissue; (2) lung cancer;
(3) mesothelioma, a cancer of the mem-
brane lining the chest and abdomen;
and (4) cancers of the gastrointestinal
tract. Prevailing opinion is that there is
no minimum dose causing the various
cancers. The environmental release of
asbestos fibers from the use and dis-
posal of numerous products may
present widespread harmful exposure
to the general public.
Presently, government agencies such
as the National Institute for Occupa-
tional Safety and Health (NIOSH) and
the Occupational Safety and Health Ad-
ministration (OSHA) are directing atten-
tion to the hazards of asbestos exposure
through proposal of more stringent reg-
ulations. Current OSHA standards limit
asbestos exposure to a time-weighted
average of 2 fibers/cm3 over an 8-hr pe-
riod, with a 15-min ceiling limit of 10
fibers/cm3. The existing standard is
based on counts of fibers 5 p. or longer
in length and having an aspect ratio
greater than 3:1 using Phase Contrast
Microscopy (PCM) to analyze samples
collected from the breathing zone.
The overall objective of this research
is to develop and verify testing proto-
cols for quantifying fiber release from
commercial asbestos products and pro-
-------
posed substitute materials during com-
mon fabricating operations, and sec-
ondly, to obtain actual fiber release data
with which to rank the release potential
of various asbestos product/commer-
cial operation combination. The initial
research project was developed by
three phases to study environmental re-
lease:
* Phase I • Preliminary assessment to
define the status and applicability of
any existing methods.
• Phase II - Development of a stand-
ard test method.
• Phase III Tests of the potential fiber
release of some representative as-
bestos products/operations.
Phase I was an assessment of previ-
ously existing laboratory procedures
used to estimate asbestos release rates
and/or exposure in the atmosphere. No
reproducible methods could be found
for generating and measuring the re-
lease of asbestos fibers during indus-
trial operations on asbestos-containing
materials. Also, no procedure that could
be considered controlled and repro-
ducible was identified. Therefore, this
research effort proceeded with the de-
sign, construction, and testing of the ap-
paratus and identification of the analyti-
cal techniques that together would
constitute such a method and test the
reprodutibility of the method. The re-
sult of Phase I was a recommendation
of test procedure for measuring the as-
bestos fiber released during commer-
cial product use. Phase II fully devel-
oped and tested the laboratory
procedure. The objective of Phase II was
to evaluate the precision of the labora-
tory procedure and to determine its sen-
sitivity to variation of fiber generating
and sampling factors. Phase III included
additional precision tests and a compi-
lation of an asbestos fiber release po-
tential index that ranked various pairs of
material operations according to their
potential for causing worker and envi-
ronmental exposure to asbestos fibers.
In addition, simultaneous samples were
taken and analyzed by the NIOSH PCM
method during 32 of the experiments.
The test methodology developed for
the project is referred to as the "glove
box" method. An apparatus was devel-
oped that allows reproducible genera*
tion of a cloud of asbestos fibers within
a confined volume. The fibers are gen-
erated by means that are physically sim-
ilar to common industrial operations.
The equipment was constructed from
readily available parts so that it could be
reproduced by other investigators. The
2
apparatus consists of a table top glove
box, a controlled, variable speed work
feeder, a remote power source coupled
to the tool by a flexible drive shaft, a fan
to provide consistent mixing within the
glove box, a filter holder, and means to
withdraw up to four samples at constant
rates.
Development of an asbestos fiber re-
lease potential index required means to
generate tn aerosol that would allow
ranking of industrial or commercial op-
erations on the various products. To this
end, the tools and the machining rates
and materials to be used mimicked, as
closely as possible, those operations
commonly employed. The intent was to
reproduce the mechanism of the com-
mercial operation, not to reproduce the
entire commercial operation. The tools
were actuated mechanically rather than
by hand to enhance the precision and
repeatability of the experiments.
The test materials were obtained di-
rectly from manufacturers insofar as
was possible. Direct contact was made
with the quality control department (or
other appropriate division) to be sure of
obtaining materials for which manufac-
turing specifications were known.
These data included the percent as-
bestos, the nature and composition of
binders and extenders, and the results
of any other physical and chemical anal-
yses that are available. The information
attainable from manufacturers was in-
adequate in some cases so the fiber re-
lease potential index computations
were based on bulk analyses. These
bulk analyses and the PCM analyses
were performed at the Mt. Sinai School
of Medicine, Environmental Science
Laboratory.
A test procedure and quality assur-
ance plan were developed. The analyti-
cal procedure of choice was the provi-
sional EPA transmission electron
microscopy (TEM) method that was de-
veloped for EMSL/RTP by the Illinois In-
stitute of Technology Research (IITRI)
under separate contract. The repro-
ducibility of the procedures was tested
during this project by replicate perfor-
mance of the same experiment {sawing
of an asbestos cement sheet) and found
to be good. Ultimately, the TEM analy-
sis was chosen over the PCM method
because of its superior capacity to pro-
vide information about the concentre*
tion of very small particles.
There is debate among asbestos re*
searchers •• to which configurations of
small asbestos particles are hazardous.
Some adhere to a strict definition of
fibers; other include other structures
such as bundles of fibers, agglomera-
tions of fibers, and fibers adhered to
small pieces of binder or other material.
The TEM data of this research include
counts of all of these structures and
fiber release potential factors calculated
only for fibers. All data have been re-
ported to facilitate alternate computa-
tions by any reader.
Procedure
To test the potential for release of
fibers from commercial use, an as-
bestos fiber generation system was de-
signed and built to simulate commercial
product shaping. The material/opera-
tion (M/0) chosen for evaluating the
techniques was sawing asbestos ce-
ment sheet. The fiber generation sys-
tem was contained in a controlled atmo-
sphere glove box, as is the sample
collection apparatus. Samples were col-
lected on Nuclepore** polycarbonate
filters and sent to the laboratory for
analysis of asbestos fibers.
Equipment
The equipment used for generating
and measuring airborne asbestos con-
sisted of the following components:
1) controlled atmosphere glove box.
2) fiber generation system, 3) air sam-
pling system. 4) glove box decontami-
nation unit, 5) carbon coating unit.
6) TEM, and 7) PCM.
Glove Box
A Labconco controlled atmosphere
glove box (Figure 1) served as the
sealed test chamber for the fiber gener-
ation and air sampling system. The inte-
rior volume of the glove box is about
0.33 m3. The glove box provides a com-
pletely sealed environment in which to
conduct the experiments. The glove box
has a 0.01 m9 interchange compartment
to prevent contamination of room air
during passage of materials into or out
of the main chamber. Two 20-cm-
diameter glove ports are located on the
front of the box, with a pair of neoprene
gloves clamped to the ports for use in
manipulating components inside the
test chamber. The glove box was also
equipped with a 70- by 48-cm safety
glass viewing panel, two 115-volt elec-
trical outlets, and one 15-watt fluores-
cent light. The glove box was made of
fiber glass reinforced polyester mete-
rial.
•Mmtion of tntto nwnn or commtreW product!
dOM not COnitilUtt •ROtaMfMM Of
lion tor UM,
-------
Fiber Generation System
The secondary manufacturing opera-
tion initially simulated in the glove box
was the sawing of asbestos cement
sheet. The fiber generation system con-
sisted of a circular saw, saw table, mate-
rial feed mechanism, and asbestos ce-
ment sheet. The design of the system
attempted to minimize the number of
components inside the box and to use
off-the-shelf components to the extent
possible.
A small saw table (Figure 2) was fabri-
cated to support the saw and power
shaft, as well as the asbestos cement
sheet. The table is 30 cm long, 13 cm
wide, and 18 cm high, and is bolted to
the floor of the glove box. Two spring
clips and metal bar hold the materials
firmly in place as it is fed to the saw. The
material feed apparatus is shown in Fig-
ure 3.
Air Sampling System
Air samples were collected by pass-
ing a known volume of air through a
polycarbonate membrane filter. Real
time monitoring of asbestos fibers was
conducted using the GCA Fibrous
Aerosol Monitor (FAM) (Model FAM-1).
The FAM is designed to automatically
count airborne fibers for sample times
of 1, 10, 100, and 1000 min and display
the count and resulting concentration
on a digital display.
Glove Box Decontamination
Unit
A Dayton vacuum (Model No. 27564)
and Dayton asbestos filtering system
(Model No. 6X724) were used to decon-
taminate the glove box. The vacuum
line runs from the glove box to the vac-
uum unit and the filtering system, which
was located outside the building. The
vacuum is rated at about 90 ft3/min. The
asbestos filter system meet OSHA
standards for vacuuming asbestos and
consists of a HEPA cartridge filter to
back up the primary collection bag. The
decontamination unit was designed to
remove asbestos in the box without
contaminating room air during the
cleaning cycle.
Disposal polyvinyl gloves are used to
transfer used asbestos cement sheet
from the glove box to sealed plastic
bags. Whenever the glove box was
opened for washing, a personal respira-
tor with a NIOSH-approved filter car-
tridge was worn in addition to the dis-
posal gloves. Disposable towels were
placed in sealed plastic bags after use in
washing the glove box interior.
Figure 1. Controlled atmosphere glove box.
Figure 2. Sawing/grinding table.
-------
Figure 3
i.il It'i'tl ,i;»/i.»Mf MS
Carbon Coating Unit
Carbon coatino, of the polycarbonate
fillets was performed usini) ;i Thermi
omcs (Model TL1 10) Vacuum Evapora
lor rented from the Department of
Anatomy of the George Washington
University ei Washington. D.C.
Transmission Electron Micro-
scope
Fiber counting and si/iny was per
formed tisitui IITRI's 100 KV TEM. The
filters were prepared in a clean room
adiacent to 'he TEM room. The Mteis
were transferred to ,111 election micro
scope iEM> (jrid. and the filter was dis
solved in .1 modified Jaffi* Wu:k Washer
The EM t|ud was viewed under a Huo
tescent view'i'i) screen inscribed with
iiraduations So estimate the lemjth and
width of fibrous p.iMii (cs
Laboratory Procedures
The laboratory procedures for i|enei
atini) and mtMsuriiu) airborne asbestos
« ortsist (.if sevii steps 1 * sample piepa
ration. '/\ lihi'i i|enei,ihon, \\\ sample
i nlli'i lion •!' i|lnvi- ho* decontiimina
ticrt Si i a'lxin I'lMt'iK) (>l transfer to F.M
t|nd. and /) TEM examination and data
collection.
Sample Preparation
Air samples were taken on 37 mm
diameter, 0.4-(i pore si/e polycarbonate
fibers. The shiny, smooth side was used
as the particle capture surface The filter
was supported by a cellulose pad in a
3/ mm plastic filter holder A piece of
t.ipe. which also served as a iahel. was
placed on the filter cartridge so that it
foimrd .in air tu|ht seal lietween the
bottom half ami middle run) of the plas
tic filter holder
Fit>or Generation
ft>r tin? niituit set of runs, the sawmi)
of asbestos cement sheet was the
method for the <|cnetatu>n of fibers The
asbestos cement sheet w.is fed mtu tl'e
saw wheel at .1 constant rate tiy a van
.iliie speed motor The leruith of mate
(Ml i ut and the time teqiined for the mil
were recorded on the data sheet Prior
to the cut. the fan in the fiont left comet
of the i|love hex was switched on to cir
i ul,ite the an inside the hox duruui the
i ut The Ian operate!I duriiu) thet ut. but
was switched off at the end of the cut
because larije cement particles from the
bottom of the hox were reentramed if
the fan wive left on
The theoretical settlim) rate data are
in dose agreement with actual settling
data obtained under working condi
dons Fibers 1 t< to fi p in lencjth with an
aspect ratio of roughly b:l are a com
mon material dispeised from overhead
insulation in huildmos The seitlmu, ve
locities (or fibers ii i>. ? i>. ar\d 1 |i in
lenijlh. with a !j.1 aspect rat'o an»k
about 10 mm A 10 mm sample was
then taken with the (AM to determine if
the hox was suffii lentlv clean 'o pro
ceed with another experiment The en
tenon of a f AM readmit of less than 0 It)
f «:c for ,i 10 mm average w.is selected
tor the indication of a (lean iilove box it
the I AM toadimi exceeded 0 10 f cc. the
>|love box was levacinimed and of
washed usini) water. p,ipet towels and
dispos.ible i|loves A respiraloi was
worn diiiirui all these ipeiations
-------
Carbon Coating
The carbon coating of the polycar-
bonate filters was performed at the
George Washington University, about
10 miles from the ES laboratory. The
polycarbonate filters remained in the
plastic filter holder at all times, so there
was no handling of the filters prior to
the application of the carbon film to the
filter.
Transfer to EM Grid
The collected panicles from the car-
bon coated polycarbonate filter were
transferred to an electron microscope
grid. The transfer was accomplished in
a modified Jaffe Wick Washer. Briefly,
the Jaffe Wick Washer is a petri dish
containing a substrate to support the
EM grid and carbon coated polycarbon-
ate filter. Solvent is added to cause dis-
solution of the polycarbonate mem-
brane with a minimum loss or
dislocation of the particles. The result is
a membrane-free EM grid with particles
embedded in the carbon coating.
TEM Examination
The EM grid was examined in the
TEM at a magnification of 250X to as-
sess the quality of the EM grid. Since
asbestos fibers were found isolated as
well as with each other or with other
particles in varying configurations, the
fibrous particles were characterized as
asbestos structures of the following
types:
• A fiber was defined as a particle
with an aspect ratio of 3:1 or greater
with substantially parallel sides.
• A bundle was a particle composed
of fibers in a parallel arrangement
with each fiber closer than one fiber
diameter.
• A cluster was a particle with fibers
in a random arrangement such that
all fibers were intermixed and no
single fiber was isolated from the
group.
• A matrix was a fiber or fibers with
one end free and the other end em-
bedded or hidden by a particle.
Figure 4 demonstrates the different
types of asbestos structures.
The asbestos fiber count was given in
term* of the number of asbestos struc-
tures that were identified. Thus, a clus-
ter was counted as one asbestos struc-
ture, even though there were numerous
individual fibers comprising the cluster.
Similarly, a bundle was counted as one
asbestos structure, even though the
bundle was composed of several
(though not always distinguishable)
fibers.
Width and length measurements
were obtained for individual fibers, and
a cylindrical shape was assumed for
volume calculations. Bundles and clus-
ters were sized by estimating their
width and length. A summation of indi-
vidual diameters was used to obtain
total width and an average length for
the total length. A laminar sheet shape
was assumed with the average diame-
ter of the individual fiber as the thick-
ness. Matrices were sized by summa-
tion of the best estimate of individual
fiber components. A laminar or sheet
structure was assumed for volume cal-
culation.
The selected area electron diffraction
ISAED) pattern was obtained for the
fiber portion of each structure by use of
the field limiting aperture. Electron
diffraction patterns from single fibers of
asbestos minerals fall into distinct
groups. TEM and SAEO patterns ob-
tained with standard samples were
used as guides to fiber identification.
From the visual examination of the elec-
tron diffraction pattern, the structure
Count as 1 Fiber
Count as 2 Fibers
Count as 3 Fibers
/\
Count as 1 Bundle
Count as f Cluster
-XV£L
Count as 1 Matrix
r
tlgun 4. Types ff asbestos structures
was classified as belonging to one of
the following categories:
• Chrysotile
• Amphibole group
> Ambiguous
• No identification
Data Reduction
The basic quantities to be calculated
are air flow rate, fiber number concen-
tration, and fiber mass concentration.
Some means for assuring compara-
bility among diverse M/O combinations
was needed. This was done by weigh-
ing the material before and after each
experiment to determine the amount of
material actually machined. These
weights were measured on a laboratory
balance having a 160-g capacity and a
sensitivity of 0.1 mg. The weight loss.
together with the percent asbestos (as
per phase contrast microscope and
x-ray fluorescence analysis) in the ma-
terial, provides a factor by which the re-
sults were normalized. This factor was
merely the mass of asbestos machined
from the piece of material. The concen-
tration of asbestos fiber measured was
divided by the mass of asbestos ma-
chined so that the units of the concen-
tration measured were:
Fibers/cm3
(Gram MilledHFraction of Asbestos)
The mass of asbestos removed from
each product by each operation was
held constant so that the sawing.
drilling, and sanding experiments have
a common basis. The fiber concentra-
tions in the aerosol generated during
these experiments are then a true re-
lease pc'ential index. In addition, a nor-
malization on the volume of material
milled will be developed. Other factors
affecting worker exposure (such as the
length of time of the operation on a ma-
terial, the mass of asbestos machined
away during the operation, and the ef-
fectiveness of any control devices) can
be tested later so that the index values
can be used to project potential expo-
sures.
Personnel Protection
At the conclusion of each experimen-
tal run, the operator removed the filters
from the main glove box chamber and
placed them into the smaller chamber.
This smaller chamber (0.01 m3) was
sealed off from the contaminated larger
chamber during the experiment. During
removal of the filters from the smaller
-------
compartment, and during the subse-
quent vacuuming of the larger cham-
ber, the worker wore a mask (MSA-Type
S filter or equivalent) for his protection
from fugitive particles. The air in the
room was tested for fibers periodically
with the FAM.
Results and Discussion
The experimental design of this study
has been based upon achieving the fol-
lowing objectives:
• Determining the precision of the en-
tire fiber release analytical system
(composed of fiber generation sys-
tem, air sampling system, carbon
coating unit, and TEM particle
counting methods).
• Comparing TEM results to PCM re-
sults.
• Collecting data for asbestos fiber re-
lease potential index.
The first of these objectives is of pri-
mary concern because determining the
precision of the analytical system must
precede all subsequent efforts to evalu-
ate asbestos containing products in the
laboratory. The second objective is im-
portant because broad application of
the method for testing will require
knowledge of how the methods may be
compared. Collecting data for the fiber
release potential index was done to de-
termine if the resulting values were sig-
nificantly different and if the various ex-
perimental parameters could be
measured accurately.
System Development and
Testing
Approximately 20 preliminary tests
were conducted in Phase I of the project
to establish the values for several of the
test variables. After these values were
established it was determined that fiber
loading test results could be repeated.
Phase II consisted of nine reproducibil-
ity tests using the cut off wheel on as-
bestos cement sheet to establish the
TEM sampling criteria. Phase III (con-
sisting of five material/operation tests
of eight runs each) was designed to de-
termine whether a fiber release poten-
tial index could be developed and if so,
what was the range of values. During
this effort, two additional tests were run
to determine the effect of inverting the
sample filters. In addition, two tests
were run with the filters located at the
top and bottom of the glove box to de-
termine if there was stratification of the
fibers. A test matrix of the three phases
of the project is shown in Table 1.
Table 1. Test Matrix
Phase Description
No. of
Tests Operation
Material
Analysis
I
Preliminary experiments
20 Cut off wheel Asbestos Gravimetric
cement sheet & SEM
II
III
III
III
III
III
Reproducibility tests
Fiber- environmental
release index tests
• Inverted samples
• Stratification tests
Fiber environmental
release index tests
Fiber environmental
release index tests
Fiber environmental
release index tests
Fiber environmental
release index tests
9
8
2
2
8
8
8
8
Cut off wheel
Sawing
Sawing
Sawing
Sawing
Grinding
Drilling
Drilling
Asbestos cement
sheet
Asbestos cement
sheet
Asbestos cement
sheet
Asbestos cement
sheet
Millboard
Brake lining
Asbestos cement
sheet
Millboard
TEM
TEM
TEM
TEM
TEM, PCM
TEM, PCM
TEM, PCM
TEM, PCM
Comparison of TEM to PCM
Results
Millipore filter samples were col-
lected for PCM analysis during all M/0
runs except those for sawing of as-
bestos cement sheet with the cutoff
wheel. The results of the PCM analyses
were compared to the TEM analyses of
the Nuclepore® filter samples that were
collected simultaneously for four of the
M/0 experiments. The samples taken
for PCM analysis during the brake
shoe/grinding experiments and the in-
verted Millipore filter samples have
been saved but not analyzed because of
budgetary restrictions.
It was first attempted to correlate the
PCM result to the total asbestos fiber
result of the TEM analysis. Even though
the TEM is able to discern far smaller
and therefore far more fibers than the
PCM, it was thought that there might be
some multiplier which could be applied
to the PCM result to adjust for its lesser
sensitivity. The correlation between the
two measurements is decent for the
millboard/saw experiment. The results
for the other experiments are poor: the
slopes are negative and the correlation
coefficients near zero.
Correlation was tried for all structures
(except matrices) found by TEM to the
PCM result. The correlations were sim-
lar to those obtained for the asbestos
fibers. An attempt was made to corre-
late the total NIOSH fibers (fibers longer
than 5 |x with an aspect ratio >3.0) actu-
ally counted by the TEM analyst to the
PCM result. Such fibers were seen on
only 2 of the 32 filters analyzed by TEM,
and the resulting correlation was uni-
formly poor. Next, all asbestos struc-
tures (again except matrices) not having
a length greater than 5 IJL and an aspect
ratio <3.o reported by the TEM were
considered. Again, no correlation was
found.
In the comparison of TEM to PCM re-
sults, the number and percentages of
asbestos fibers, all structures, and as-
bestos bundles and clusters that project
to be longer than 5 JJL were examined.
The inescapable conclusion is that al-
most none of the structures generated
and measured during these tests are
longer than 5 p, and that restricting the
analysis to those that exceed 5 (x is tan-
tamount to deciding to ignore 99% of all
the asbestos fibers generated during
the machining operation.
The difficulty in correlating TEM anal-
ysis with PCM results appears to be
over the particles with diameters less
than 0.4 p, and lengths smaller than 5 JA.
The graphic plots of the TEM data for
Filter 230 Saw AC Sheet, Filter 362 Drill
AC Sheet, Filter 314 Grind Brakes, and
Filter 282 Saw Millboard (Figures 5,6,7,
-------
.1
.01 .02 .04 .06.08.1 .2 .4 .6 .8 1.0
Fiber Diameter I Microns)
4 6 8 10
Figure 5. 230 Saw AC Sheet.
and 8) illustrate the paniculate sizes as
seen by the TEM and the PCM vs TEM
comparison difficulties.
Only 14 of the 480 structures reported
by TEM for the 8 sawing asbestos ce-
ment sheet with the toothed blade were
larger than 0.3 IJL in diameter. When only
asbestos fibers are considered, the situ-
ation is even worse. The TEM has suffi-
cient resolution so that any structure
larger than 0.125 p, in diameter is classi-
fied as a bundle or cluster. There are by
definition almost no TEM fibers larger
than 0.125 \L. This observation led to the
attempt to correlate the PCM result to
the concentration of clusters and bun-
dles longer than 5 n, with aspect ratios
<3.0. There is no correlation between
the fraction of particles longer than 5 p.
and the fraction having a diameter
greater than 0.3 \L. We have arbitrarily
chosen 0.3 JJL as the smallest size particle
visible with the PCM. It may be that
0.2 \L particles are visible with the PCM.
ilt may be that 0.2 (JL particles are visible
to a skilled microscopist, but that would
not change the conclusion that the vast
majority of particles reported by TEM
are not visible with a light microscope.
Figure 9 illustrates the PCM operator's
difficulty. This electron photomicro-
graph reveals several structures that are
obviously distinct single fibers and two
larger bundles of fibers. The larger of
these two would be seen by PCM and
labeled a fiber.
It should be noted that 0.0625 \i, is an
approximation of the fiber diameter that
results from the rounding to the nearest
1 /16 it. Precision measurements by other
researchers indicate that this smallest
diameter, which is regarded as compris-
ing an asbestos fiber, lies between 0.03
JJL and 0.07 n,.
The total number of fibers counted
during the PCM analyses was some-
what low; only the filters collected dur-
ing the millboard sawing experiments
exceeded the 10 counts/100 fields crite-
ria established by NIOSH. This, in itself,
is an interesting result. Ten minutes
after cessation of a shaping operation
the PCM analysis shows low levels of
fibers when in fact the concentration of
asbestos fibers is in the hundreds of
fibers per cubic centimeter. Projections
of the number of concentration of clus-
ters and bundles, which are the struc-
tures most likely to be identified as
fibers by a light microscopist, indicate
the values of their concentrations to be
over twice the concentrations measured
by PCM. If structures identified as fibers
(almost all of which have such small di-
ameters that they are not visible by
PCM) are included, the error increases.
Further, other researchers have re-
ported that if TEM counts are extended
to thousands of particles (from the hun-
dred or so counted in these analyses)
more large (>5 n) structures are found
than is expected based upon the
log-normal distribution of the shorter
fibers. This again implies that the PCM
method, in addition to not detecting
over 99% of the total asbestos particles
in the aerosol, has underestimated the
concentrations of structures it ought to
have measured by a significant margin.
Asbestos Fiber Release Poten-
tial Index
After gaining assurance that the tech-
nique was reproducible, the experimen-
tal effort turned to development of a
fiber release potential index. This index
was to be a quantitative measure of the
propensity of asbestos containing ma-
terials to release fiberous particles dur-
ing their subjection to various industrial
or commercial machining operations.
Sawing (with two different types of saw
blades), drilling, and grinding were se-
lected as the operations to be tested.
Asbestos cement sheet millboard and
brake shoes were the material chosen.
The results (Table 2) appear to form
the basis for a system for ranking vari-
ous M/0 pairs. There are other differ-
ences between the various M/0 pairs in
terms of the types of structures that
they are prone to produce. For example,
brake shoe grinding produced signifi-
cantly more bundles and clusters than
any other operation when all structures
are considered (Table 3) and when only
asbestos structures are considered
(Table 4). Sawing asbestos cement
sheet with the toothed saw blade pro-
duced more matrix particles than any
other operation. Interestingly, the per-
cent of all structures that were identified
as asbestos is more or less constant and
-------
is not well correlated with the percent
asbestos in the material being milled.
This probably irrplies that most of the
non-asbestos particles generated dur-
ing the millinc ope* ations are larger and
settle from the aerosol during the 10-
minute waiting period. This means that
the test procedure to * large extent seg-
regates the important (asbestos related)
portions of the dust created during
milling from the less important genera-
tion of extraneous dust.
Conclusions
The literature survey carried out m tiro
early stages of this project revealed no
procedure, which coulc be considered
controlled and reproducible, b.v'sted for
generating an asbestos aerosol oy oper-
ations similar to commercial/irdustnal
machining of non-friable asbestos bear-
ing products. A technique was devel-
oped for generating such an aerosol.
The technique is timple and rela*' .-oly
inexpensive, and is sufficiently fie. ibte
to be adapted to mimic a variety of in-
dustrial milling operations on a variety
of products.
TEM was selected for analysis of the
sampled parties because of its ability
to discern extremely small particles and
to differentiate asbestos from other
fibers. The reproducibility of the tech-
nique of the combined generation/ana-
lytical sysiem was found to be excellent.
Relative standard deviations for repeti-
tive performances of an experiment
were typically in the 40% to 80% range.
An index that rate* the propensity of
six industrial/commercial operations on
three asbestos bearing materials was
developed. The urv.a of tho indbx are:
A»be»to» Fiber*
(Grim Aubwtos Mac- iliu'd) (Mi Sampled, cm'}
The experimental procedure maintains
the amount of asbestos machined rela-
tively constant so that errors of scale are
eliminated and so thai the filter samples
collected are all (ceded properly for
TEM c-oalvsi*. Normalizing the data on
the grama of asbestos actually ma-
chined during an experiment removes
the residual variance within an experi-
mental sat and the residual variance
among the various experiments The
various material op«r*tiofts tested are
listed in decreasing order of their
propensity to generatw *»besto> fibers
in TabU B.
Stratification of particle* within the
aerosol was tested by sampling «lmul-
/UU
ao
60
40
20
10
•s *
\
J '
10
.8
6
,4
y
•
D.
X -
• -
—
-1 —
60ft
20
IOC
8 to
tier
und
lust
Fern
--
••-
---
4-
t
let
er*
X
-1
1 '
4 x
•
X
— •••-•
l~ —
—
'":.'..
e
X
i
•
t — —
PCM fiber*
(1
K
D
X
A«
JJT
+ *
t
ft
/
y
—
/
. >
A
]/
4- "
:::
1
j>
"ft
/
/
1
J
•+••
-.L
01 .02 .04 .06,08 1 .2 .4 .6 .81.0 2 4 6 8 10
Fiber Oiemeter (Microns)
Figure 0. 362 DrillA/C Sheet.
taneously at high and low elevations
within the glove box. The aerosol was
four.d to be homogeneous.
Deferences in concentrations mea-
su-ed by upward-facing as opposed to
downward-facing filters were tested by
sampling simultaneously with upright
and inverted filters. No significant dif-
ference was found.
Fiber and structure lengths were
found to be log-normally distributed.
Fiber and structure diameters do not fit
tha log-normal distribution as well be-
cause of the large number of 0.0626 p.
(approximately) diameter fibers.
Correlation between TEM results and
PCM result* was attempted. Samples
were takcm with Mllllpore filters during
32 of the exr*i Iments, representing
four different fu'O pairs, No co, relation
of the PCM anatyftia of these Mars with
thfi TEM analvitfi of the Nucltpore* fil-
ters that were exposed simultaneously
could be fot" u. It was determined that
only approximately 1% of the structures
identified by TEM were longer than 5 »*..
None of the structures identified as be-
ing fibers by the TEM have diameters
sufficiently large to be seen by the PCM.
We ware unable to identify any subset
of the TEM data that would correlate
with the PCM data.
Recommendations
The results of thu study have sug-
gested several avenues of future re-
search to define and
-------
80
60
4O
20
to
a
6
4
2
1 0
g
4
2
~ • -t
-O-l
. X -4
• .
— •- -
"^3 Fit
SBu
1 Ck
4 Ma
•n
mill
i*te
trix
9$
f$
i i
1 r
• "
, i
o
t
Si
X y
* 8
„
>
\
SK-
ir
-• •
... .
o
X
X
-*-
...
K
<
X
-_ X
-••
p
Q]
CM Ft
3
f
£
Dots
./
. A*
tfv.-
i-C'n
Jr °*i
S
Jb-
8
„ - — ,
/
---a
j
/x
X
/
/
€—
^\
/
tf
/
/
i— A. -
/
r
.'
r
1
Figure 7.
.Oj> .04 .06.08.1 .2 .4 .6 .81.0
Fiber Diameter (Microns)
314 Grind Brake*.
4 6 8 tO
ronment as it has been reported
that reducing tool speed signifi-
cantly reduces fiber release.
3. The efficiency of various vacuum
and wetting devices in reducing
the amount of asbestos released
into the environment should be in-
vestigated. This would include the
abatement procedures and final
disposal evaluation criteria for as-
bestiform materials.
4. The effects of agglomeration and
settling on the fine fiber concentra-
tion in the aerosol should be inves-
tigated. Agglomeration may signif-
icantly affect the concentration
and size distribution of suspended
fibers and the inhalable paniculate
fiber fraction thereof.
S. Work on development of a correla-
tion between the TEM and PCM
methods should continue in order
to determine whether data col-
lected by the latter may be corre-
lated with health effects data. In-
formation on the scanning
electron microscope should also
be included.
6. Means should be developed for re-
lating the data provided by this
technique to worker and general
public environmental exposures.
The full report was submitted in fulfill-
ment of Contract No. 68-03-3040 by
Engineering-Science under the spon-
sorship of the U.S. Environmental Pro-
tection Aqency.
-------
100
80
60
40
20
10
^ 8
CO
1 6
t 4
I 2
1.0
.8
.6
.4
.2
.1
- • -76 Fibers
_ D - 5 Bundle.
X - 16 Cluster
• - 7 Matrix
5
S
• 4
"
.
•
.
t
.
:•
!•
»•
!•
c*
::;
M
^'
1
: >
D
fe
"
y
x
'
X
PC
-------
Table 2. Asbestos Fiber Release Potential Index
Asbestos Fibers perccpergm Asbestos MUM
Materiel!
Operation
Number
of Tests
Average
(f/cc/gm)
Standard
Deviation
Relative
Standard
Deviation
AC sheet
Cut-off wheel
Millboard
Saw
Brakes
Grind
AC sheet
Saw
AC sheet
"oWi
Millboard
9
8
8
12
a
8
1838.3
048.6
465.1
305.4
282.9
105.2
338.6
435.2
217.3
383.7
222.6
32.0
18.4
67.3
46.7
119.1
78.7
30.5
Wtt
Table & Distribution of Structure Types Generated During Asbestos Release Experiments
ITEM Analysis)
Operation
fibers %
Bundles %
Clusters %
Matrices %
Cut-off AC sheet
Saw millboard
Grind brakes
Saw AC sheet
Drill AC sheet
Drill millboard
76.3
77.2
63.2
62.4
82.5
93.4
9.S
5,2
14.2
12.4
3.8
0.0
6.4
12.3
18.8
5.5
6.1
2.5
7.8
5.3
3.8
18.7
7.5
4.1
Table 4. Distribution of Asbestos Structure Types Generated During Asbestos Release
Experiments (TEM Analysis}
Asbestos Structure Distribution
Matertall
Operation
Percent
Afbettot
Fibers
Percent
Bundles
Percent
Asbestos
Clusters
Percent
Asbestos
Matrices
Percent
Asbestos
§»w
S2.8
64.6
45.0
48.3
81.6
68.5
6.7
4.7
10.8
7.8
3.4
0.0
4.7
10.2
13.9
4.8
4.3
f.f
5.4
3.4
2.8
13.6
8.4
4.5
83.0
72.4
7S.2
77.7
74.2
11
-------
Table 5. Environmental Release
Material
Operation
Environmental
Release Index
Value
Asbestos cement sheet
Millboard
Brake shoes
Asbestos cement sheet
Asbestos cement sheet
Millboard
Saw (cut-off wheel)
Saw (toothed blade)
Grind
Saw (toothed blade)
Drill
Drill
1838
647
465
305
283
105
Dennis A. Falgout is with Engineering-Science, Fairfax, VA 22030.
Thomas J. Powers is the EPA Project Officer (see below).
The complete report, entitled "Environmental Release of Asbestos from Com-
mercial Product Shaping," (Order No. PB 85-188 878/AS; Cost: $29.50,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati. OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PA
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-85/044
------- |