EPA/560/5-88/011
sd States
renmental Protection
Agency
Office of Pesticides and
Toxic Substances
Washington DC 20460
EPA-560/5-88-011
September, 1988
Toxic Substances
Al
Asbestos Content In Bulk
Insulation Samples:
Visual Estimates and
Weight Composition
a
-------
EPA 560/5-88-011
September, 1988
ASBESTOS CONTENT IN BULK INSULATION SAMPLES:
Visual Estimates and Weight Composition
By
Ian M. Stewart
RJ Lee Group
Monroeville, PA 15146
Prepared for:
Midwest Research Insitute
Kansas City, MO 64110
EPA Contract No. 68-02-4252
Work Assignment 43
MRI Project 8861-A43
Field Studies Branch
Exposure Evaluation Division
Office of Toxic Substances
U.S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This report was prepared under contract to an agency of the United States Government.
Neither the United States Government nor any of their employees makes any warranty,
expressed or implied, or assumes any legal liability for any third party's use of or the
results of such use of any information, apparatus, product, or process disclosed in this
report, or represents that its use by such third party would not infringe on privately owned
rights. Mention of trade names or commercial products does not constitute endorsement
for use.
-------
INTRODUCTION:
In April 1973, the U.S. Environmental Protection Agency (EPA) issued the National
Emissions Standards for Hazardous Air Pollutants (NESHAP) for asbestos (38 FR 8820).
The NESHAP regulation governs the removal, demolition, and disposal of asbestos-
containing bulk wastes. An asbestos-containing product, as stated by the regulation, was
defined for the first time to be a product with greater than 1% asbestos, by weight. The
intent of the 1% limit was:
...to ban the use of materials which contain significant
quantities of asbestos, but to allow the use of materials
which would: (1) contain trace amounts of asbestos which
occur in numerous natural substances, and (2) include very
small quantities of asbestos (less than 1 percent) added to
enhance the material's effectiveness. (38 FR 8821)
It must be clearly understood that the EPA NESHAP definition of 1% by weight was not
established to be a health-based standard.
In May 1982, EPA issued a regulation which required schools to inspect and sample
suspect friable surfacing materials for their asbestos content. EPA maintained consistency
in its definition of an asbestos-containing material (ACM) by defining it as 1% by weight.
At that time, the Agency investigated the available methodologies for measurement of
asbestos fibers. The regulation included an interim methodology entitled "Interim Method
for the Determination of Asbestos in Bulk Insulation Samples" (47 FR 23376). The
polarized light microscope (PLM) protocol issued by the Agency was prepared by expert
mineralogists and has been generally accepted by the analytical community as the
appropriate analytical tool for measurement of asbestos content in bulk samples.
The interim method includes a description of its quantitation procedure. This procedure
employs a technique called "point counting" to provide a determination of the area percent
of asbestos in the sample. Based on a measurement made by point counting, the 1982 rule
states "...reliable conversion of area percent to dry weight is not currently feasible unless
the specific gravities and relative volumes of the material are known." EPA amended this
statement in a correction to the regulation in September 1982 (47 FR 38535). EPA altered
paragraph 1.7.2.4 of Appendix A of the rule by stating, "Paragraph 1.7.2.4 of Appendix A
of the rule was intended to provide for a point counting procedure or an equivalent
estimation method for determining the amount of asbestos in bulk samples." This
-------
correction, acknowledged the practical and economic limitations of the point counting
method and permitted the use of the visual estimation methodology. Visual estimation
methodology is employed by most PLM laboratories and gives results which are very
similar to a volume percentage.
In the following discussion, the validity of the assumptions that are made in extrapolating
an area/volume percentage estimation to a weight percentage estimation of the asbestos
content of insulation and other building materials will be examined. The reader should note
that this discussion considers only the expected variation from the true weight percentage as
is found when applying the visual estimate technique to determine the asbestos content in a
bulk sample. The questions of laboratory/analyst variability of such visual estimations are
not considered in this discussion.
RELATIONSHIPS BETWEEN AREA, VOLUME, AND WEIGHT PERCENTAGE
The principles of stereology are well documented (see, for example, "Quantitative
Stereology," Underwood) 1 and will not be reiterated here other than to state that in
classical stereology, with the assumption of a homogeneous distribution of phases within a
solid, there is a direct relationship between the volume fraction of a phase present in the
solid and the area fraction of that phase observed in a section taken through the solid.
That is to say,
Vp Ap
where Vp refers to the volume of the phase p present in the total volume V, and Ap
represents the area projection of that phase in a planar section of that solid of total area A.
It should be noted that, for the classical rules of stereology to apply in a transmission
sample, the section through the sample should be no thicker than the thickness or diameter
of the smallest component.
The point counting method has been criticized as a technique for observing ACM because it
does not take into consideration the fact that the asbestos fibers present may be
comparatively thin in the Z direction relative to the other components present. Thus, if the
volume percentage of asbestos present is extrapolated from the projected area obtained by
the point counting technique, the volume percent of asbestos present will generally be
Underwood, E.E., Quantitative Stereology, Addison-Wesley Publishing Company, (1970)
-------
overestimated. As a result, the majority of laboratories analyzing ACM have adopted a
visual estimate which allows a certain amount of latitude on the part of the rnicroscopist to
compensate for this thickness factor when examining samples on the microscope slide. In
most instances, the visual estimation of asbestos content is made on a stereomicroscope
with which the rnicroscopist may more readily estimate the third dimension. Therefore,
these estimates may be more readily extrapolated to a volume percentage than those from
the point count method. This technique is essentially that which is proposed in the Interim
American Society for Testing and Materials (ASTM) Method. Currently, this method is
being considered for adoption by the National Institute of Standards and Technology
(formerly the National Bureau of Standards) as part of its National Voluntary Laboratory
Accreditation Program for the determination of bulk asbestos in samples. This procedure
will provide a measurement of the asbestos in the sample which may be easily extrapolated
to a volume measurement
CURRENTLY ACCEPTED EXPERIMENTAL METHOD
The currently accepted and most generally used methodology for the identification of
asbestos in building materials is compatible with both the EPA interim method and the
proposed ASTM method. Identification of the asbestos type present using polarized light
microscopy follows accepted mineralogical practices. The quantification of the asbestos
content by visual estimation which is used is acceptable under the amendment to the 1982
Regulation published in the Federal Register and is substantially the same as that
recommended in the ASTM method. It can be seen that there is continuity of approach and
direct correlation between existing data and that which may be produced under the ASTM
procedure.
While the visual estimation procedure is generally called the polarized light microscopy
method, the microscopist, in fact, uses a combination of a low magnification stereo-
microscope for preliminary examination and estimation of the percentage of each fiber type,
followed by a detailed examination, using the polarized light microscope, of individual
fibers removed from the bulk material. The procedure has been outlined in a draft to
ASTM Committee D22.05 dated January 14, 1988—"Standard Method of Testing for
Asbestos-Containing Materials by Polarized Light Microscopy."
-------
The method calls for bulk samples of building materials to be first examined with a low
power binocular microscope. By use of such a microscope, the following observations can
be made.
(1) The fibers can be detected.
(2) The homogeneity of the material can be determined.
(3) A preliminary identification of the fibers present can be made.
(4) An estimate of fiber content by volume can be made.
(5) Fibers may be separated from the matrix for more detailed
analysis of subsamples with the polarized light microscope.
The method has been used, essentially in its present form, by the majority of the
participants in the EPA Bulk Sample Analysis Round Robin program. These results
indicate generally good reproducibility and good accuracy in assessing the volume
percentage of an asbestos mineral present in an insulating material. The accuracy of such
an analysis does not differ very greatly from the expected inhomogeneity (or homogeneity)
of the material being analyzed (manufacturers' specifications generally show a range of
composition for any one product which frequently was additionally modified at the point of
application). In the ASTM technique, quantification of asbestos content is discussed in the
following terms: "A quantitative estimate of the amount of asbestos present is most readily
obtained by visual comparison of the bulk sample in slide preparations to other slide
preparations and bulk samples with known amounts of asbestos present in them." The
document goes on to state that estimates of the quantity of asbestos obtained by the method
are neither volume nor weight percent estimates, but are based on estimating the projected
area, from observation, of the distribution of particles over the two dimensional surface of
the glass slide, and on an observation of bulk material, and that a basis for correcting to a
weight or volume percent has not been established. It is this latter aspect which will be
discussed more fully in this document. The ASTM method, however, provides for the
percentage to be first assessed from the bulk material as observed on the stereomicroscope;
it would seem, therefore, that this percentage is a closer approximation to a volume
percentage rather than a projected area one. In addition the ASTM document states,
"However, the error introduced by assuming that the estimates are equivalent to weight
percent is probably within the precision of the visual estimate technique."
-------
CORRELATION OF WEIGHT PERCENTAGE WITH VOLUME PERCENTAGE
To correlate the weight fraction of the phase to its area or volume fraction, it is necessary,
as is pointed out in the The EPA Test Method, that the specific gravities and relative volume
fractions of all the phases present in the material are known.1
In any multicomponent system consisting of n components, the weight percent of
component i is given by the following formula:
PixVjx 100
I Pi x
where Pi is the specific gravity of the ith component and Vi is the volume of the ith
component From this formula, it is clear that if the volume percent and the density of each
individual element in a bulk insulation sample is known, it would be possible to obtain a
weight percentage for any particular component and specifically for those components
which are classed as asbestos. To determine this information experimentally would,
however, be extremely time consuming, requiring the separate identification of each
component in the matrix, determining its specific gravity from reference tables, and
applying these factors in the formula.
An alternative conversion is therefore suggested in which an average density is assumed for
the nonasbestos matrix. In this model, the weight percentage, Wa, of a particular asbestos
type present at a volume percentage of Va and having a density of Pa present in a matrix of
density Pm is given by the formula
- PaXV.xlOO -
(100-Va)xPm + (VaxPa)
The density value ascribed to the nonasbestos matrix should be selected taking into
consideration the major constituents of the matrix but, for a large range of commonly
encountered inorganic matrices, a value of 2.5 g/cm3 may be assumed..
Interim Method for The Determination of Asbestos in Bulk Insulation Samples
EPA 600/MA-82-020, December, 1982.
-------
PRACTICAL APPLICATION
These formulas will be applied to a range of samples. In applying formula 1 to determine
actual weight percentages, published values for the several components were used. To
determine the weight percentages using the model described by formula 2, a matrix density
of 2.5 g/cm^ was assumed.
Sample 1 Acoustical Material
Sample 1 is a sample of an acoustical material taken from an actual ceiling treatment.
Component Vol% Wt% (Actual) Wt% (Model)
Chrysotile 15.0 15.12 15.51
Glass Fiber 60.0 60.47
Carbonate 10.0 10.85
Cement 3.0 3.26
Clay 10.0 8.53
Gypsum 2.0 1.78
(Appendix 1 shows in detail how these weight percentages are calculated.)
Sample 2 Round Robin Sample from Independent QC Ring
Sample 2 is from an independent round robin sample series in which four laboratories
participated. Reported values for amosite content were 30%, 30-40%, 45%, and 15-20%.
The results from the second laboratory were taken using the midpoint of the reported
compositional range (the midpoint of the reported range for sample two was selected as
most probably representing the actual composition, lying between the reported values of
one and three, with four regarded as an outlier).
Component Vol% Wt% (Actual) Wt% (Model)
Amosite 35.0 38.82 41.55
Carbonate 35.0 32.94
Cement 30.0 28.24
Sample 3 Sample A EPA Bulk Sample Analysis Round Robin No. 16
Sample 3 is sample A from the EPA Bulk Sample Analysis Round Robin series, Round
number 16. .
Component Vol% * Wt% (Actual) Wt% (Model)
Amosite 3.0 4.04 3.92
Glass 87.0 92.29
Cellulose 10.0 3.67
Volume percentage data for samples 3,4,5 and 6 are averages taken from EPA Round Robin
reports and would not normally be reported to this level of significance.
-------
Sample 4 Sample D EPA Bulk Sample Analysis Round Robin No. 16
Sample 4 is Sample D from the EPA Bulk Sample Analysis Round Robin series, Round
Number 16.
Component Vol% Wt% (Actual) Wt% (Model)
Chrysotile 3.0 3.53 3.12
Clay 97.0 96.47
Sample 5 Sample D EPA Bulk Sample Analysis Round Robin No. 17
Sample 5 is Sample D from the EPA Bulk Sample Analysis Round Robin series, Round
Number 17.
Component Vol% Wt% (Actual) Wt% (Model)
Chrysotile 2.9 2.56 3.01
Amosite 30.7 34.40 36.90
Cement 66.3 63.04
Sample 6 Sample A EPA Bulk Sample Analysis Round Robin No. 17
Sample 6 is Sample A from the EPA Bulk Sample Analysis Round Robin series, Round
Number 17.
Component Vol% Wt% (Actual) Wt% (Model)
Crocidolite 97.0 97.52 97.78
Cement 3.0 2.48
It is clear from these data that, for most samples, the weight percentage of the asbestos
content is not substantially different from the volume percentage which is normally reported
and is within the expected variation both of the analytical procedure and the sample
homogeneity. A close estimate of the weight percentage can be derived from a simple
model which assumes an average matrix density of 2.5 g/cm3.
Plots of the difference between observed volume percentage and calculated weight
percentage for chrysotile, density 2.6 g/cm3. (Figure 1) and crocidolite, density 3.4 g/cm3.
(Figure 2) are shown calculated using this model. The maximum deviation between the
numerical values of weight and volume percentage occurs near the 50% mark and, in the
worst case (crocidolite), is less than 10%.
Exceptions will be found in samples whose matrices have significantly higher or lower
densities than the asbestos observed. Figure 3 presents the extreme case of crocidolite
(density 3.4 g/cm3) in a matrix of cellulose with an estimated average density of 0.9 g/cm3.
-------
The magnitude of the discrepancy in the critical region near 1% is shown in figure 4. If
only the volume percentage estimate is used, mass percentages as high as 3% would be
reported as below the definition of ACM. In this case, a conversion to weight percentage is
necessary if the weight percentage is not to be grossly underestimated.
SAMPLE TREATMENT
Some samples, for example floor tiles, roofing felts, and some cementitious products, may
require special treatment (ashing, solvent or acid extraction) to separate the asbestos from
other materials in order to facilitate analysis. In such cases, the resulting weight loss of the
sample due to treatment must be recorded and any volume to weight percentage correction
applied to the remaining material must be further corrected to take this weight loss into
consideration. For example, if 30% asbestos is detected in a sample after processing which
resulted in a 25% weight loss, then the corrected asbestos content is 0.75 x 30 =22.5%
CONCLUSIONS AND RECOMMENDATIONS
An assessment has been made of the validity of extrapolating to a weight percentage the
area or volume percentage of asbestos present in a sample as determined by polarized light
microscopy. A model has been presented which can be applied to area or volume
percentage data to give a more accurate estimation of the weight percentage. With the
exception of asbestos-containing materials having a substantial density differential between
matrix and asbestos, generally low density cellulosic orperlitic matrices, the magnitude of
this correction is smaller than the expected variability imposed by both the analytical
variation and the inhomogeneity of the sample. As a result, the weight percentage of
asbestos present can generally be equated with the observed area or volume percentage.
The following recommendations are made:
1) For samples whose approximate average matrix density is close to that of the
asbestos species observed (within 0.5 g/cm3), assume equivalence of weight
and area or volume percentage.
2) For samples whose approximate average matrix density differs from that of
the asbestos species present by more than 0.5 g/cm3, convert the observed
area or volume percentage to weight percentage using formula 2, using a
matrix density consistent with the principal matrix components.
-------
TABLE I
Calculated relationship between weight percentage and volume percentage of chrysotile
(density 2.6 g/cm3) in matrix of average density of 2.5 g/cm3.
DIFFERENTIAL
VOLUME % WEIGHT % (WEIGHT %-VOLUME %)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
0.00
5.17
10.36
15.51
20.63
25.74
30.83
35.90
40.94
45.97
50.98
55.97
60.94
65.89
70.82
75.73
80.62
85.49
90.35
95.18
100.00
0.00
0.19.
0.36
0.51
0.63
0.74
0.83
0.90
0.94
0.97
0.98
0.97
0.94
0.89
0.82
0.73
0.62
0.49
0.35
0.18
00.00
These values were used to produce Figure 1.
-------
TABLE II
Calculated relationship between weight percentage and volume percentage of crocidolite
(density 3.4 g/cm3) in a matrix of average density 2.5 g/cm3.
DIFFERENTIAL
VOLUME % WEIGHT % WEIGHT %-VOLUME %
0 . 0.00 0.00
5 6.68 1.68
10 13.13 3.13
15 19.35 4.35
20 25.37 5.37
25 31.19 6.19
30 36.82 6.82
35 42.27 7.27
40 47.55 7.55
45 52.67 7.67
50 57.63 7.63
55 62.44 7.44
60 67.11 7.11
65 71.64 6.64
70 76.04 6.04
75 80.31 5.31
80 84.47 4.47
85 88.51 3.51
90 22.45 2.40
95 96.27 1.27
100 100.00 0.00
These values were plotted to produce the curve of Figure 2.
10
-------
TABLE HI
Calculated relationship between weight percentage and volume percentage of crocidolite
(density 3.4 g/cm3) in a matrix of average density 0.9 g/cm3.
DIFFERENTIAL
VOLUME % WEIGHT % WEIGHT %-VOLUME%
0 0.00 0.00
5 16.59 11.59
10 29.57 19.57
15 40.00 25.00
20 48.57 28.57
25 55.74 30.74
30 61.82 31.82
35 67.04 32.04
40 71.58 31.58
45 75.56 30.56
50 79.07 29.07
55 82.20 27.20
60 85.00 25.00
65 87.52 22.52
70 89.81 19.81
75 91.89 16.89
80 93.79 13.79
85 95.54 10.54
90 97.14 7.14
95 98.63 3.63
100 100.00 0.00
These values were {dotted to produce the curve of Figure 3.
11
-------
TABLE IV
Calculated relationship between weight percentage and volume percentage of crocidolite
(density 3.4 g/cm3) in a matrix of average density 0.9 g/cm3 over the range 0 to 2
_ 1 frt
volume%.
DIFFERENTIAL
VOLUME % WEIGHT % WEIGHT %-VOLUME %
0 0 0.00 0.00
0.1 0.38 0.28
0.2 0.75 0.55
0.3 1.12 0.82
0.4 1.49 1.09
0.5 1.86 1.36
0.6 2.23 1.63
0.7 2.59 1.89
0.8 2.96 2.16
0.9 3.32 2.42
1.0 3.68 2.68
1.1 4.03 2.93
1.2 4.39 3.19
1.3 4.74 3.44
1.4 5.09 3.69
1.5 5.44 3.94
1.6 5.79 4.19
1.7 6.13 4.43
1.8 6.48 4.68
1.9 6.82 4.92
2.0 7.16 5.16
These values were plotted to produce the curve of Figure 4.
12
-------
o
O)
to
c
9
O
10
9-
8-
7-
6-
5-
4-
«
3-
«
2"
1-
Mass - Volume Percent Differential
Chrysotile In Matrix Of S.G. = 2.5
10 20 30 40 50 60 70 80 90 100
Observed Volume Percentage
Figure 1.
13
-------
Mass - Volume Percent Differential
Crocidolite In Matrix Of S.G. = 2.5
o
B.
£
c
w
fl>
«»-
Q
10
9-
«
8-
7
6-
5-
4-
3-
2-
1-
o-
10 20 30 40 50 60 70 80 90 100
Observed Volume Percentage
Figure 2.
14
-------
o
O)
O
O
Mass - Volume Percentage Differential
Crocidolite In Matrix Of S.G. = 0.9
0 10 20 30 40 50 60 70 80 90 100
Observed Volume Percentage
Rgure 3.
15
-------
Crocidolite In Cellulose Matrix
6-
5-
o
Q.
O)
"5
3-
2-
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Observed Volume Percentage
i&
-------
APPENDIX I
Examples of Calculations
a) Actual Weight Percentages
Sample #1
Wt. %
Density Rel. Wt. x 100
Compound Vol. % (g/cm3) Relative Weight Total Rel. Wt.
Chrysotile 15.0 2.6 15x2.6 = 39.0 15.12
Glass Fiber 60.0 2.6 60x2.6=156.0 60.47
Carbonate 10.0 2.8 10x2.8 = 28.0 10.85
Cement 3.0 2.8 3.0 x 2.8 = 8.4 3.26
day 10.0 2.2 10.0x2.2 = 22.0 8.53
Gypsum 2.0 2.3 2.0x2.3 = 4.6 1.78
TOTALS 100.0 -- 258 100.01
b) Approximate weight percentages based on a model with assumed 2.5 g/cm3 density.
Sample #1
Vol. % Rel. Wt. Approx. Wt. %
Chrysotile 15.0 15x2.6 = 39.0 15.51
Non-asbestos matrix 85.0 85 x 2.5 = 212.5
TOTALS 100.0 251.5
Sample #5
Sample 5 contains both chrysotile and amosite. The approximate weight percentage is
calculated separately for each asbestos type as follows:
Vol. % Rel. Wt. Approx. Wt. %
Chrysotile (density 2.6 g/cm3) 2.9 x 2.9x2.6= 7.54 3.01
Non-chrysotile matrix 97.1 94.1 x 2.5 = 242.75
Chrysotile totals 100.0 250.29
Amosite (density 3.3 g/cm3) 30.7 101.31 36.90
matrix 69.3 173.25
Amosite totals 100.0 274.56
17
-------
50272-101
REPORT DOCUMENTATION
PAGE
1. REPORT NO.
EPA 560/5-88-011
3. Recipient's Accession No.
4. Title and Subtitle
Asbestos Content in Bulk Insulation Samples:
and Weight Composition
5. Report Date
Visual Estimates
September 1988
7. Author(s)
Ian M. Stewart
8. Performing Organization Rept. No.
9. Performing Organization Name and Address
a. RJ Lee Group, Monroeville, PA 15146
b. Midwest Research Institute, Kansas City, MO 64110
10. Protect/Task/Work Unit No.
11. Contract(C) or Grant(G) No.
------- |