EVALUATION OF TWO CLEANING METHODS
FOR REMOVAL OF ASBESTOS FIBERS FROM CARPET
John R. Kominsky
Ronald W. Freyberg
PEI Associates, Inc.
Cincinnati, Ohio 45246
Jean Chesson
Chesson Consulting
Washington, D.C. 20036
Eric J. Chatfield
Chatfield Technical Consulting Limited
Mississauga, Ontario, Canada
EPA Contract No. 68-03-4006 '
Work'.Assignment. 2-10 . \
Technical Project Monitor :
William C. Cain
Project Officer \
Thomas J. Powers
Water and Hazardous Waste Treatment Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
The information in this document has been funded wholly ornn part by
the U.S. Environmental Protection Agency under Contract 68-03-4006 to PEI
Associates, Inc. It has been subjected to the Agency's peer and administra-
tive review, and it has been approved for publication as an EPA
Mention of trade names or commercial products does not constitute endorsement
or recommendation for use.
document.
-------
FOREWORD ;
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
materials that, if improperly dealt with, can threaten both public health and
the environment. The U.S. Environmental Protection Agency (EPA)^ is charged
by Congress with protecting the Nation's land, air, and water resources.
Under a mandate of national environmental laws, the Agency strives to formu-
late and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life.
These laws direct the EPA to perform research to define our environmental
problems, measure the impacts, and search for solutions. i
r
The Risk Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development, and demonstration programs
to provide an authoritative, defensible engineering basis in support of the
policies, programs, and regulations of the EPA with respect to drinking
water, wastewater, pesticides, toxic substances, solid and hazardous wastes,
and Superfund-related activities. This publication is one of the products of
that research and provides a vital communication link between the researcher
and the user community. j
This report provides information on the decontamination effectiveness of
dry-vacuuming and wet cleaning to remove asbestos fibers from cairpet under
experimental conditions. A reduction in the amount of asbestos 'in the carpet
would suggest a possible reduction in the potential exposure to building
occupants. :
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
iii
-------
ABSTRACT
The effectiveness of dry-vacuuming and wet-cleaning for the removal of
asbestos fibers from carpet was examined and the potential for fiber re-
entrainment during carpet cleaning activities was evaluated. Routine carpet
cleaning operations were simulated by using high-efficiency particulate
air (HEPA) filtered dry vacuum cleaners and HEPA-filtered hot-water extrac-
tion cleaners on carpet artificially contaminated with asbestos fibers.
Overall, wet-cleaning with a hot-water extraction cleaner reduced the level
of asbestos contamination in the carpet by approximately 70 percent. There
was no significant evidence of either an increase or a decrease in carpet
asbestos concentration after dry-vacuuming. The level of asbestos contami-
nation had no significant effect on the difference between the asbestos
concentrations before and after cleaning. Airborne asbestos concentrations
were two to four times greater during than before the carpet cleaning activi-
ties. Neither the level of asbestos contamination in the carpet nor the type
of cleaning method used greatly affected the difference between the airborne
asbestos concentration before and during cleaning. '•
This document was submitted in fulfillment of Contract No. 68-03-4006 by
PEI Associates, Inc., for the U.S. Environmental Protection Agency's Office
of Research and Development, Risk Reduction Engineering Laboratory. This
report covers the period from January 1988 to September 1989, and work was
completed as of September 30, 1989.
iv
-------
- CONTENTS ;
Foreword | i-j-j
Abstract , -jv
Figures j Vii
Tables j V1--,•-,•
Acknowledgments : x
!„ Introduction ! i
Background ; i
Objectives i i
2., Conclusions and Recommendations : 2
Conclusions > 2
Recommendations 2
3.. Study Design i 3
Test facility 3
Experimental design ; 5
Sampling strategy , 8
Preliminary sampling and analytical performance study 10
Sample size revisions I 12
4. Materials and Methods i 15
Selection of carpet : 15
Selection of carpet cleaning equipment j 15
Sampling methodology 16
Analytical methodology ; 17
Statistical analysis • is
5. Experimental Procedures 20
Prestudy air monitoring 20
Carpet contamination : 21
Disposal of asbestos-containing material i 26
Site cleanup ; 26
Poststudy air monitoring 27
6. Quality Assurance ; 28
Sample chain of custody ' 28
Quality assurance sample analyses : 28
TEM analysis of unused sample containers , 29
Spray-application technique 30
7. Results and Discussion ! 33
Carpet samples i 33
Air samples 37
References ! 41
-------
CONTENTS (continued)
Appendices
A Transmission Electron Microscopy Results From Preliminary
Performance Experiments on the Microvac and Sonic :
1 Extraction Procedures : 42
B Chrysotile Fiber Size Distribution in the High- and !
Low-Concentration Ampules ' 46
C Carpet Asbestos Concentrations Before and After Carpet
Cleaning , 43
,D Average Asbestos Concentrations Before and After Carpet
Cleaning for Each Experiment ; 55
E Fiber Length Distributions of Asbestos in Carpet Samples
Collected Before and After Carpet Cleaning 57
vi
-------
FIGURES
Number
1
2
3
5
6
7
8
10
Layout of test facility, Experiments 1 through 16 ;
Layout of test facility, Experiments 17 through 24
Sample locations for preliminary performance experiments
on the microvacuum and sonic extraction sampling and
analytical techniques
Average asbestos concentrations in carpet samples from
preliminary performance experiments with the microvac and
sonic extraction sampling and analytical techniques
Distribution of chrysotile fiber lengths in the low and
high concentration aqueous asbestos suspension
Fiber size distributions from preliminary study of asbes-
tos dispersion by spraying
Average asbestos carpet concentrations before and after
cleaning for each cleaning method and carpet con-
tamination loading
Ninety-five percent confidence intervals for asbestos
concentration after cleaning as a proportion of the
concentration before cleaning
Cumulative percentages of asbestos particles in carpet
after cleaning, airborne asbestos particles observed
during cleaning, and asbestos fibers used to contaminate
the carpet
Average airborne asbestos concentrations before and dur-
ing carpet cleaning
vii
Page
4
7
11
14
25
32
34
36
38
40
-------
TABLES
Number
1 Experimental Design for Experiments 1 Through 16
2 Experimental Design for Experiments 17 Through 24 < 6
3 Probability of Rejecting, at the 5% Level, the Null Hypothesis
of No Difference Between Two Experimental Treatments as a
Function of the Number of Carpet Samples and the Actual
Difference in Asbestos Concentrations
(CV = 0.75, 1.0, 1.25} 9
4 Number of Carpet Samples Collected in Experiments 1 ;
Through 16 i 9
5 Number of Carpet Samples Collected in Experiments 17 ;
Through 24 ; i()
j
6 Variance Components Analysis Comparison of Performance of
Microvacuum and Sonic Extraction for Asbestos Recovery
From Carpet ; 13
7 Probability of Rejecting, at the 5% Level, the Null Hypothe-
sis of No Difference Between Two Experimental Treatments
as a Function of the Number of Carpet Samples and the-
Actual Difference in Asbestos Concentrations ',
(CV = 0.3, 0.4, 0.75) 13
8 Summary of Prestudy Airborne Asbestos Concentrations •
in Test Facility i 20
9 Summary of Results of Transmission Electron Microscopy ,
Analyses for Low- and High-Concentration Ampules ; 26
10 Results of Duplicate Sample Analyses : 29
11 Results From Preliminary Study of Asbestos Dispersion by
Spraying—Fibers and Fiber Bundles (All Lengths) ; 30
I
12 Fiber Length Distributions From the Preliminary Study of
Asbestos Dispersion by Spraying ! 31
viii
-------
TABLES (continued)
Number
13
14
15
16
Summary Statistics for Asbestos Concentrations in Carpet
Before and After Cleaning ;
Analysis of Variance Table for Difference Between Asbestos
Concentrations Before and After Cleaning
Estimated Asbestos Concentration in Carpet After Cleaning
as a Proportion of the Concentration Before Cleaning ,
Structure Morphology Distribution in Carpet Samples Col-
lected Before and After Carpet Cleaning
33
35
35
37
-------
ACKNOWLEDGMENTS j
i
This document was prepared for EPA's Office of Research and Development,
Risk Reduction Engineering Laboratory, in fulfillment of Contract! No. 68-03-
4006. Mr. Thomas J. Powers, P.E., served as the EPA Project Officer. Mr.
Powers also offered the invaluable suggestion of contaminating the carpet
with an aqueous suspension of asbestos. Mr. William C. Cain served as the
Technical Project Monitor for this project. The administrative efforts and
support given by Mr. Roger Wilmoth of EPA's Office of Research and Develop-
ment are greatly appreciated. '•
Colonel Stephen F. Kollar, Commander, U.S. Air Force, authorized the use
of a building at Wright Patterson Air Force Base to conduct this research
study. Administrative support given by Behram Shroff, Douglas Post, and
Suzette Smith of the U.S. Air Force is also acknowledged. Also appreciated
are review comments and suggestions provided by William Burch, P.E., Kin
Wong, Ph.D., Elizabeth Dutrow, and Joseph Breen, Ph.D., of EPA's Office of
Toxic Substances; and William McCarthy, Bruce A. Hollett, C.I.H.,'and Michael
Beard of EPA's Office of Research and Development. Christopher Frebis of
Computer Sciences Corporation also provided a statistical review of this
report.
John R. Kominsky, C.I.H., and Ronald W. Freyberg of PEI Associates,
Inc.; Jean Chesson, Ph.D., of Chesson Consulting; and Eric J. Chaltfield,
Ph.D., of Chatfield Technical Consulting Limited were the principal authors.
Robert S. Amick, P.E., of PEI Associates, Inc., served as senior reviewer.
Marty Phillips and Jerry Day of PEI Associates, Inc., performed the technical
and copy edits, respectively. ;
-------
SECTION 1
INTRODUCTION
BACKGROUND
Buildings that contain friable asbestos-containing materials (ACM) may
present unique exposure problems for custodial workers. Under certain condi-
tions, asbestos fibers can be released from fireproofing, acoustical plaster,
and other surfacing material. The episodic release of asbestos fibers from
aging and deteriorating ACM relates to a myriad of factors, such as the
condition and amount of asbestos present, the accessibility of the material,
activity within the area, vibration, temperature, humidity, airflow, use
patterns, etc. A major concern is the extent to which carpet and furnishings
may be serving as reservoirs of asbestos fibers and what happens >to these
fibers during normal custodial cleaning operations. i
The Asbestos Hazard Emergency Response Act (AHERA) requires that all
carpeting in areas of school buildings in which asbestos-containirig materials
are present be cleaned with either a high-efficiency particulate iair (HEPA)-
filtered vacuum cleaner or a hot-water extraction cleaner ("steam cleaner").
Little quantitative information is available on how effectively these cleaners
remove asbestos fibers from carpet or on the potential for airborne asbestos
fibers to become reentrained during these carpet cleaning activities.
This report presents an evaluation of the concentrations of asbestos
fibers in the carpet before and after cleaning by each of the two^ cleaning
methods and a summary of the air monitoring results obtained duriing cleaning.
A complete description of the air monitoring portion of the study; is pre-
sented in a separate EPA report.1 ;
OBJECTIVES
A series of controlled experiments in an unoccupied building were
performed to evaluate the effectiveness of a HEPA-filtered vacuum cleaner and
a HEPA-filtered hot-water extraction cleaner in the removal of asbestos from
carpet. A secondary objective was to investigate the potential for the
reentrainment of asbestos fibers during carpet-cleaning activities.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
The following are the principal conclusions reached during this study:
0 Wet cleaning significantly reduced the asbestos concentration in
the carpet by approximately 70 percent. There was no significant.
change in carpet asbestos concentration after dry-vacuuming.
0 Both dry vacuuming and wet cleaning of carpet resulted ;in a statis-
tically significant increase in the airborne asbestos concentration
in the area. Airborne asbestos concentrations were two to four
times greater during than before the carpet cleaning activities.
0 Airborne asbestos particles reentrained during carpet-cleaning
activities were predominantly smaller than the residual particles
in the carpet.
0 Use of a microvacuuming technique on the carpet tended to recover
significantly less asbestos than the bulk-carpet sonic extraction
technique.
RECOMMENDATIONS
The study conclusions led to the following recommendations:
0 Further research should be conducted to examine the performance
different HEPA-filtered dry and wet carpet cleaners, e._
ance as a function of horsepower, static water lift, and
air volume and velocity. Further study also should be
examine other cleaning methodologies, e.g., repeated carpet
of
g., perform-
operating
conducted to
cleaning.
Further research is needed to confirm the possible reentrainment
asbestos fibers during actual operating conditions and to
exposure to custodial workers performing these activities
ings containing friable asbestos-containing materials.
of
to determine
in build-
-------
SECTION 3 ;
STUDY DESIGN
TEST FACILITY !
This study was conducted in an unoccupied building at Wright-Patterson
Air Force Base in Dayton, Ohio. Two rooms, each containing approximately 500
square feet of floor space, were constructed in a large bay of the building.
Figure 1 presents the layout of the test facility. The rooms were
constructed of 2-in. x 4-in. lumber with studs spaced on 24-in. centers and
3/4-in. plywood floors. The ceiling, floor, and walls were double-covered
with 6-mil polyethylene sheeting. (The interior layer of polyethylene sheet-
ing was encapsulated and replaced after each experiment.) Where;the joining
of separate sheets of polyethylene was necessary, the sheets were overlapped
at least 12 in. and joined with an unbroken line of adhesive to prohibit air
movement. Three-inch-wide tape was then used for further sealing of the
joint on both the inside and outside of the plastic sheeting. !
Entry from one room to another was through a triple-curtained doorway
consisting of two overlapping sheets of 6-mil polyethylene placed over a
framed doorway. Each sheet was secured along the top of the doorway, and the
vertical edge of one sheet was secured along one side of the doorway and the
vertical edge of the other sheet was secured along the opposite side of the
doorway. . :
Determination of room size (approximately 29 ft x 17 ft x 7.5 ft) was
based on the minimum amount of time required to vacuum or wet-clean the room
and to attain an adequate volume of sample air to achieve a specified analy-
tical sensitivity. A 52-inch, ceiling-mounted, axial-flow, propeller fan was
installed in each room to facilitate air movement and to minimize temperature
stratification.
Separate decontamination facilities for workers and waste materials were
connected to the experimental areas. The worker decontamination'facility
consisted of the following three totally enclosed chambers: I
1) An equipment-change room with triple-curtained doorways—one to the
work area and one to the shower room.
2) A shower room with triple-curtained doorways—one to the equipment
change room and one to the clean change room. The one :shower
installed in this room was constructed so that all water was col-
lected and pumped through a three-stage filtration system. The
-------
o
2
LLJ
Q_
<
CO
cc
CTi
3
O
c:
01
•^-
s-
OJ
o.
X
UJ
o
ns
to
O)
O
>>
fO
O)
CD
u_
-------
three-stage filtration system consisted of a 400-micrometer, nylon-
mesh, filter-bag prefilter; a 50-micrometer, filter-bag second-
stage filter; and a 5-micrometer final-stage filter, filtrate was
disposed of as asbestos-contaminated waste. Water was, drained from
the filtration system exit into a sanitary sewerage system.
3) A clean change room with triple-curtained doorways—one to the
shower room and one to the noncontaminated areas of the building.
Air Filtration ;
High-efficiency particulate air (HEPA) filtration systems were used to
reduce the airborne asbestos concentrations to background levels; after each
experiment. These units were operated during both preparation and decontam-
ination of the test rooms. The air filtration units did not operate during
the carpet-cleaning phase of each experiment. i
I
One HEPA filtration system was dedicated to each test room ([Figure 1).
Each unit provided approximately eight air changes every 15-minute period.
The negative pressure inside the test rooms ranged from -0.08 to -0.06 in. of
water. All exhaust air passed through a HEPA filter and was discharged to
the outdoors (i.e., outside the building). All makeup air was obtained from
outside the building through a window located on the side of the;building
opposite the exhaust for the HEPA filtration systems. •
EXPERIMENTAL DESIGN
i
Experiments 1 Through 16 \
I
Two carpet-cleaning methods—dry vacuuming with a HEPA-filtered vacuum
and wet cleaning.with a HEPA-filtered hot-water extraction cleaner—were
evaluated on carpet artificially contaminated at levels of approximately 100
million and 1 billion asbestos structures per square foot (s/ft2). Each
combination of cleaning method and contamination level was replicated four
times. Four different (same model) HEPA-filtered vacuums and foiir different
(same model) HEPA-filtered hot-water extraction units were used in this study
so the results would not be influenced by the peculiarities of a'single unit.
Each machine was used only once per combination of cleaning method and con-
tamination level. This experimental design, which yielded a total of 16
experiments, is summarized in Table 1. i
TABLE 1. EXPERIMENTAL DESIGN FOR EXPERIMENTS 1 THROUGH 16
Approximate Cleaning method and experiment
contamination ——;
level, s/ft2 Wet cleaning Dry vacuuming
100 million 1, 4, 5, 8 2, 3, 6, 7 :
i
1 billion 9, 12, 13, 16 10, 11, 14, !l5
-------
Two experiments were conducted during each day of the study.; Each
combination of cleaning method and contamination level was tested! twice in
each test room. A single experiment consisted of contaminating a: new piece
of carpet (approximately 500 square feet) with asbestos fibers, collecting
work-area air samples, collecting microvacuum and bulk carpet samples, dry-
vacuuming or wet-cleaning the carpet while concurrently collecting a second
set of work-area air samples, collecting a second set of microvacuum and bulk
carpet samples, removing the carpet, and decontaminating the test room. Each
test room was decontaminated by encapsulating the carpet and the polyethylene
sheeting on the ceiling and walls prior to their removal. These materials
were removed and replaced after each experiment.
Experiments 17 Through 24
Eight additional experiments were conducted to evaluate the differences
in asbestos retention characteristics of new carpet versus carpet that has
been wet-cleaned. These experiments were designed for comparison! with
Experiments 1 through 16. :
Experimental procedures for Experiments 17 through 24 were identical to
those in the first 16, except for one difference; prior to contamination, the
carpet was dry-vacuumed, wet-cleaned, and then dry-vacuumed again!when dry.
These experiments were conducted to examine differences in the asbestos fiber
retention characteristics of new carpet versus new carpet which had been wet
cleaned. These experiments were conducted in the same test area used for
Experiments 1 through 16; however, the two 500-ft2 test rooms were converted
to four 160-ft2 test rooms, each with dimensions of approximately^ ft x 20
ft. Figure 2 shows the modifications to the two test rooms. ,
Each of the two cleaning methods was tested at two carpet contamination
levels (100 million and 1 billion s/ft2). Each cleaning method was tested
twice in two different rooms. The same four HEPA-filtered dry vacuums and
hot-water extraction cleaners were used. Each machine was used ohly once for
each combination of cleaning method and contamination level. This experi-
mental design, which yielded a total of eight experiments, is summarized in
Table 2.
TABLE 2. EXPERIMENTAL
Approximate
contamination
level, s/ft2
100 million
1 billion
DESIGN FOR EXPERIMENTS
Cleaning method and
Wet cleaning
17, 19
21, 23
17 .THROUGH! 24
experiment !
Dry vacuuming
18, 20
22, 24 !
A single experiment consisted of dry-vacuuming, wet-cleaning; and dry-
vacuuming again a new piece of carpet in a previously cleaned room; contami-
nating the carpet with asbestos fibers; collecting microvacuum and bulk
carpet samples; dry-vacuuming or wet-cleaning the carpet; collecting a second
set of microvacuum and bulk carpet samples; removing the carpet; and decon-
taminating the test room. Each test room was decontaminated by encapsulating
-------
•si-
CSJ
CD
CD
D.
X
-p
4-
O
3
O
ro
CM
s-
-------
the carpet and the polyethylene sheeting on the ceiling and walls prior to
their removal. These materials were removed and replaced after each experi-
ment,, j
SAMPLING STRATEGY j
Experiments 1 Through 16 i
Carpet Samples -- '
Bulk carpet and microvacuum samples were collected to establish the pre-
and post-cleaning carpet contamination levels. Six samples were Collected
before and six after cleaning during each experiment. i
• • "i •
Power calculations, based on computer simulations, were made to deter-
mine the number of samples to be collected before and after cleaning during
each experiment. For the purpose of these calculations, the number of
experimental replicates was fixed at four. Because little information was
available on which to base a sample size determination for carpet sampling,
statistical assumptions were based on information from the analysis of air
samples. i
Inasmuch as measured concentrations were expected to be relatively large
(i.e.,, based on fiber counts of 10 or more), individual measurements from a
given carpet were assumed to be lognormally distributed with a coefficient of
variation between 0.75 and 1.25. The power calculations were based on trans-
forming each measurement with the log scale and taking an average to give a
single measurement for each carpet. A two-sample t-test was then used to
compare various sets of four measurements (e.g., before and after cleaning).
Table 3 shows the probability of rejecting, at the 5 percent level, the
null hypotheses of no difference between experimental treatments ;for various
combinations of sample size, "true" differences between treatments, and coef-
ficients of variation. The probabilities are overestimates because sources
of variability other than sampling and analysis of the carpet were not con-
sidered. Variability between different carpets, experimental chambers,
cleaning equipment, etc., was not included. Increasing the number of carpet
samples, however, would not reduce variability introduced by these other
sources. Assuming the other sources of variability are small relative to
sampling and analysis variability, Table 3 still provided a useful guide for
determining sample size. '
Assuming a coefficient of variation of 1.0, six samples taken before
cleaning and six samples taken after cleaning gives a probability of approxi-
mately 0.84 of obtaining a statistically significant difference when one
concentration is half the other (0.5 in Table 3). Detection of more subtle
differences in concentration would be unlikely even if the sample size were
increased to eight. The chance of detecting a proportional difference of 0.5
decreases rapidly with sample sizes less than six; however, proportional
differences of less than 0.33 are detected with high probability with as few
as three samples. The number of carpet samples collected is shown in
Table 4. i
8
-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
TABLE 3. PROBABILITY OF REJECTING, AT THE 5% LEVEL, THE NULL HYPOTHESIS
OF NO DIFFERENCE BETWEEN TWO EXPERIMENTAL TREATMENTS AS A FUNCTION OF THE
NUMBER OF CARPET SAMPLES AND THE ACTUAL DIFFERENCE IN ASBESTOS CONCENTRATIONS
(CV = 0.75, 1.0, 1.25)
Asbestos level after cleaning Number of carpet samples
as a proportion of asbestos '•
level before cleaning 3 5 6 ; 8
CV = 0.75 j
0.75 0.27 0.34 0.36 0.52
0.5 0.75 0.88 0.95 0.98
0.33 0.98 1.00 1.00 1.00
0.25 1.00 1.00 1.00 1.00
0.1 1.00 1.00 1.00 1.00
CV = 1.0 ;
0.75 0.22 0.26 0.28 0.32
0.5 0.60 0.75 0.84 0.90
0.33 0.87 0.98 0.99 1.00
0.25 0.98 1.00 l.OQ 1.00
0.1 1.00 1.00 1.00 1.00
CV = 1.25 :
0.75
0.5
0.33
0.25
0.1
0.16
0.49
0.77
0.91
1.00
0.19
0.64
0.94
0.99
1.00
0.21
0.71
0.97
1.00
1.00
0.26
0.82
0.99
1.00
1.00
For example, 0.25 means that an initial concentration of 100 million fibers
per square foot before cleaning is reduced to 25 million fibers per square
foot after cleaning. I
TABLE 4. NUMBER OF CARPET SAMPLES COLLECTED IN !
EXPERIMENTS 1 THROUGH 16
Number of samples
Type Before cleaning After cleaning Field blanks
Microvacuum 192 192 24
Bulk carpet 192 192 ;'-
Total samples 384 384 !24
-------
The carpet was divided into 400 1-ft2 areas (a 16-ft by 25-|ft grid) by
using a string grid system. The carpet was then stratified into;three pairs
of equally sized sections. One bulk carpet sampling location and one micro-
vacuum sampling location were selected at random within each of the six
sections. This sampling strategy assured representative samples from the
entire piece of carpet.
Air Samples ~ :
Work-area air samples were collected to establish airborne asbestos
concentrations before and during cleaning. For each experiment,:three air
samples were collected before and three during cleaning. A total of 96 air
samples were collected. |
Experiments 17 Through 24 i
!
Bulk carpet and microvacuum samples were again collected to;establish
the pre- and post-cleaning carpet contamination levels. During each experi-
ment, four samples were collected before and four after carpet cleaning. The
number of carpet samples collected in Experiments 17 through 24 is shown in
Table 5.
TABLE 5. NUMBER OF CARPET SAMPLES COLLECTED IN •
EXPERIMENTS 17 THROUGH 24
Number of samples ;
Type Before cleaning After cleaning Field blanks
Microvacuum 32 32 4
Bulk carpet 32 32 4
Total samples 64 64 i 8
The carpet was divided into 160 1-ft2 areas (an 8-ft by 20-ft grid) by
using a string grid system. The carpet was then stratified into 'fourths.
One bulk carpet sampling location and one microvacuum sampling location were
selected at random within each of the four sections. This sampling strategy
assured representative samples from the entire piece of carpet. ;
PRELIMINARY SAMPLING AND ANALYTICAL PERFORMANCE STUDY i
Preliminary experiments were conducted to document the performance of
the microvacuum sampling and sonic extraction techniques for the recovery of
asbestos from carpet. The precision and level of recovery of asbestos by
these two methods were determined by contaminating an 8-inch by 24-inch strip
of carpet with approximately 1 billion s/ft2 and then collecting samples for
analysis by both techniques. Six microvacuum samples were collected from
10-cm by 10-cm sections of the contaminated carpet. Bulk samples; for analy-
sis by sonic extraction were collected from 2-inch by 2-inch sections of the
contaminated carpet. Sample locations, which were randomly chosen from the
contaminated carpet, are shown in Figure 3.
10 ;
-------
OJ
m
o>
o
c
CM
1
2
i
2
to
i
LU
CM
2"
CO
1
LU
1C
1
LU
CM
1
LU
1
2
(0
i
2
i
LU
CO
2
i
LU
r i—
CD re
Q. 0
X •!-
O) 4-i
o "re
E E
re re
s- ~a
0 E
-------
Each carpet sample was analyzed in triplicate to assess the iprecision of
each,method. Individual sample results are presented in Appendix A.
The data were analyzed by standard analysis of variance (ANQVA) tech-
niques. Data from each method were analyzed separately by a one-!way ANOVA
with a random effects model. These results are summarized in Tab;ie 6. For
each method, the between-sample variation contributed to most of Ithe varia-
tion., which suggests that the variation between different locations in the
carpet was greater than the variation between different preparatiions of the
same sample. These results indicate that increasing the number of carpet
samples would have a greater impact on the precision of both methods than
would increasing the number of replicate analyses of the same sample. The
calculated coefficient of variation (CV) for the microvacuum technique (166
percent) was four times larger than the CV for the sonic extraction procedure
(43 percent). Figure 4 shows the mean recoveries from each method. Micro-
vacuuming the carpet recovered significantly less asbestos than the bulk-
carpet sonic extraction procedure. The mean asbestos recovery obtained with
the microvacuum technique was 23 million s/ft2, whereas approximately 794
million s/ft2 was obtained with the sonic extraction technique. Based on the
superior precision and performance of the sonic extraction technique for
asbestos recovery from carpet, only the sonic extraction method was used to
analyze the carpet samples; all microvacuum samples collected during this
research study were archived for future consideration.
SAMPLE SIZE REVISIONS I
The preliminary experiments conducted to assess the performance of the
sonic extraction technique for asbestos recovery from carpet provided useful
information on the variability associated with this analytical technique that
was not available when the sampling strategy was being developed.: The calcu-
lated coefficient of variation associated with this method was 43 percent.
The original sample size calculations for this study assumed a CV; of 100
percent. Table 7 shows the results of new calculations for a different range
of CVs based on the results of the performance study. Assuming a| CV of 40
percent, three samples collected before cleaning and three samples collected
after cleaning give a probability of approximately 0.99 of obtaining a
statistically significant difference when one concentration is half the
other. Therefore, rather than analyze all six sets of samples collected
before and after cleaning, three sets of samples were randomly selected from
each of the 24 experiments to be analyzed. This provided a total; of 144
estimates of carpet contamination (72 estimates before cleaning and 72 esti-
mates after cleaning). I
The use of these preliminary results to modify the number of! samples
needed to achieve statistical significance greatly reduced analytical costs
and turnaround time during this study.
12
-------
TABLE 6. VARIANCE COMPONENTS ANALYSIS COMPARING PERFORMANCE OF'MICROVACUUM
AND SONIC EXTRACTION FOR ASBESTOS RECOVERY FROM CARPET
Variance components ;
Method
Sonic Extraction
Microvacuum
Between
samples
103,913
1,145
Within
sample
10,900
305
Total
114,813
1,449
Overall mean,
million s/ft2
794 ;
23 !
cv, %
43
166
TABLE 7. PROBABILITY OF REJECTING, AT THE 5% LEVEL, THE NULL HYPOTHESIS
OF NO DIFFERENCE BETWEEN TWO EXPERIMENTAL TREATMENTS AS A FUNCTION OF THE
NUMBER OF CARPET SAMPLES AND THE ACTUAL DIFFERENCE IN ASBESTOS CONCENTRATIONS
(CV = 0.3, 0.4, 0.75) ;
Asbestos level after cleaning Number of carpet samples
as a proportion of asbestos '•
level before cleaning 2 316
; CV = 0.3 :
0.75 0.54 0.68 ;0.90
0.5 : 0.98 1.00 1.00
0.33 1.00 1.00 .1.00
0.25 i 1.00 1.00 1.00
0.1 1.00 1.00 1.00
CV = 0.4 i
0.7.5 0.36 0.49 10.73
0.5 0.93 0.99 1.00
0.33 1.00 1.00 1.00
0.25 1.00 1.00 1.00
0.1 1.00 1.00 1.00
CV = 0.75
0.75
0.5
0.33
0.25
0.1
0.18
: 0.58
0.91
0.96
1.00
0.23
0.75
0.98
1.00
1.00
0.36
D.93
1.00
1.00
11.00
For example, 0.25 means that an initial concentration of 100 million fibers
per square foot before cleaning is reduced to 25 million fibers: per square
foot after cleaning. ;
i
: 'is :
-------
14
-------
SECTION 4 :
MATERIALS AND METHODS j
!
A survey was made of 14 General Service Administration (6SA), field
offices in 11 States distributed across the United States to determine the
type of carpet, HEPA-filtered vacuum, and HEPA-filtered hot-water extraction
unit to use in this study. Building managers were asked to identify 1) the
specific type and manufacturer of carpet used in 6SA buildings, 21) the manu-
facturer and model of HEPA-filtered vacuum cleaner commonly used, and 3) the
manufacturer and model of HEPA-filtered hot-water extraction unit routinely
used in their buildings. \
None of the 6SA offices routinely wet-cleaned their carpet. When wet-
cleaning was necessary, contractors were hired to perform the work. There-
fore, six trade associations (the American Institute of Maintenance, the
Building Service Contractors Association, the International Maintenance
Institute, the Environmental Management Association, the International Sani-
tary Supply Association, and the Vacuum Cleaner Manufacturers Association)
were surveyed to obtain their recommendations on a HEPA-filtered hot-water
extraction cleaner. j
I
t
SELECTION OF CARPET j
Eight GSA offices indicated a preference for the same manufacturer and
type of carpet. The selected carpet was first-grade, 100 percent:nylon, with
0.25-inch cut pile, 28 ounces of yarn per square foot, and dual vinyl back-
ing. The carpet was manufactured in roll sizes of 4.5 by 90 ft. :
SELECTION OF CARPET CLEANING EQUIPMENT !
HEPA-Filtered Vacuum !
The HEPA-filtered vacuum selected for this study was the model most
frequently mentioned in the GSA survey. The unit had an airflow capacity of
87 cubic feet per minute, a suction power of 200 watts, and 75 inches static
water lift. (Water lift is the maximum amount of force a vacuum can exert
throughout the system if the end of the vacuum hose is completely I closed
off.) This unit was also equipped with a motor-driven carpet nozzle with a
rotating brush.
15
-------
Hot-Mater Extraction Cleaner .
Three of the trade associations surveyed recommended the same hot-water
extraction unit. The selected cleaner was equipped with a HEPA-fi1tered
power head with a moisture-proof, continuous-duty, 2-horsepower Vacuum motor
that develops a 100-inch static water!ift. This unit was also equipped with
an extractor tool that uses a motor-driven cylindrical nylon-bristle brush, 4
inches in diameter by 14 inches long, to agitate and scrub the carpet during
the extraction process. !
SAMPLING METHODOLOGY ;
Bulk Carpet Samples
Carpet samples were collected before and after cleaning by using a
100-cm2 (4-in.2) template and a utility razor knife. Each carpet sample was
cut in half, providing a duplicate sample for archiving. Each piece of
carpet was placed in a separate labeled container. Wide-mouth polyethylene
jars with polypropylene screw caps were used to contain the carpet samples.
The template and utility razor were thoroughly cleaned prior to sample col-
lection to reduce the possibility of cross-sample contamination. '.
Microvacuum Samples \
Microvacuum samples were collected by vacuuming a 100-cm2 area of carpet
with a membrane filter air-sampling cassette and a vacuum pump. iThe sampling
assembly consisted of a 25-mm-diameter, 0.45-um pore-size, mixed ;cellulose
ester membrane filter with a 5-ym pore-size mixed cellulose ester backup
diffusing filter and cellulose ester support pad contained in a three-piece
cassette. The cassette was connected to an electric-powered sampling pump
with flexible tub.ing. The pump and cassette assembly was calibrated to 10
liters per minute. The 100-cm2 area was vacuumed by dragging the filter
cassette across the carpet to agitate the carpet pile. The carpet was va-
cuumed for 30 seconds in one direction, and another 30 seconds in a direction
90 degrees to the first. After 1 minute of vacuuming, the pump was turned
off and the filter cassette was labeled and sealed.
Air Samples ,
i
Air samples were collected on open-face, 25-mm-diameter, 0.45-ym-
pore-size, mixed cellulose ester membrane filters with a 5-pm pore-size,
mixed cellulose ester backup diffusing filter, and cellulose ester support
pad contained in a three-piece cassette. The filter cassettes were
positioned approximately 5 feet above the floor with the filter face at
approximately a 45-degree angle toward the floor. The filter assembly was
attached to an electric-powered vacuum pump operating at a flow rate of
approximately 10 liters per minute. In each test room, the air samplers were
positioned in a triangular pattern (Figure 1). Air samples were collected
for a minimum of 65 minutes before and during carpet cleaning to achieve a
minimum air volume of approximately 650 liters. The sampling pumps were
calibrated both before and after sampling with a precision rotameter.
i
16 ;
-------
ANALYTICAL METHODOLOGY :
Bulk Carpet Samples ;
A sonication procedure developed by McCrone Environmental Services,
Inc., was used to extract asbestos particles from the bulk carpet; samples for
subsequent analysis by transmission electron microscopy (TEN), the labora-
tory preparation procedure is as follows: '.
1) The carpet sample was placed carpet-side down in a 1000-ml beaker
containing 100 ml of a 0.1 percent solution (by volume)! of Aerosol
OT (a commercial surfactant) made with deionized particle-free
water. !
2) The beaker containing the carpet sample and Aerosol OT solution was
ultrasonicated three times, 10 minutes each time. After each
sonication, the solution was drained into the 500-ml polyethylene
screw cap sample container and another 100 ml of fresh Aerosol OT
solution was added for the next sonication. j
3) The carpet sample was then removed from the beaker. The beaker was
rinsed with 100 ml of deionized particle-free water. The rinse
from the beaker was added to the sample container. The^ carpet
sample was dried and stored. !
4) The resulting suspension was shaken vigorously to disperse the
fibers evenly and then allowed to sit for 2 minutes while the large
or heavy particles settled or rose. A measured volume of this
suspension was extracted with a disposable graduated pipette from i
to \ inch below the water surface. The aliquot was then filtered
onto a 0.22-ym-pore-size mixed cellulose ester filter backed by a
0.45-um-pore-size mixed cellulose ester filter. Three jneasured
aliquots of different volumes were generally sufficient! to attain a
good fiber loading on a filter. :
5) An optional step for removal of very large nonasbestos structures
from the sample solution before filtration was occasionally in-
cluded in the preparation procedure. This involved passing the
solution through a coarse-mesh stainless steel or plastic screen
prior to filtration. The screen was thoroughly cleaned'or replaced
before each sample. :
6) When filtration was complete, the 0.22-um filter was carefully re--
moved from the funnel assembly and placed in a Gelman "Analyslide"
dish. The filter was dried in a closed container with a dessicant.
(Some of the filters were dried by placing the dish holding the
filter on a hot plate at low temperature.) i
i
7) Portions of the filter were prepared for TEM analysis in accordance
with the NIOSH 7402 preparation procedure. At least two 200-mesh
TEM grids from different areas of the filter were prepared for each
sample. ;
17
-------
Asbestos structures were counted and identified in accordance with the
EPA provisional method, Level II.2 Only asbestos structures were counted
because the carpet samples often contained a significant number of clay
fibers and other nonasbestos structures. McCrone Environmental Services,
Inc., performed the TEM analyses on the carpet samples under separate con-
tract with EPA's Risk Reduction Engineering Laboratory in Cincinnati.
Microvacuum Samples
i
The mixed cellulose ester filters used to collect the microv:acuum carpet
samples were analyzed by TEM. These samples were prepared according to the
analytical laboratory's Standard Operating Procedure for dust sample collec-
tion. Counting and identification of the asbestos structures were performed
in accordance with EPA provisional method, Level II. McCrone Environmental
Services, Inc., performed the TEM analyses on the microvacuum samples under
separate contract with EPA's Risk Reduction Engineering Laboratory in
Cincinnati.
Air Samples ;
The mixed cellulose ester filters were analyzed by transmission electron
microscopy (TEM). These filters were prepared and analyzed in accordance
with the nonmandatory TEM method as described in the Asbestos Hazard Emer-
gency Response Act (AHERA) final rule (52 CFR 41821). Battelle Laboratories,
Columbus Division, performed the TEM analyses on the field samples under
separate contract with EPA's Risk Reduction Engineering Laboratory (RREL) in
Cincinnati. !
STATISTICAL ANALYSIS I
Carpet Samples
A single estimated concentration was obtained before and after cleaning
during each experiment by taking the arithmetic mean of the individual esti-
mates. This gave 24 pairs of concentrations, one for each experiment. The
natural logarithm of each of the 48 concentrations was used for subsequent
statistical analyses. This is equivalent to assuming that the data follow a
lognormal distribution. The lognormal distribution is commonly assumed for
measurements of asbestos and other air pollutants. '•
The geometric mean and a 95 percent confidence interval were calculated
for each contamination level and cleaning method. A three-way analysis of
variance (ANOVA)3 with contamination level (low, high), cleaning jnethod (wet,
dry)., and experimental set (1 to 16, 17 to 24) as the three experimental
factors was performed on the difference (on the log scale) between the con-
centration before cleaning and the concentration after cleaning. ; (The dif-
ference on the log scale is equivalent to the ratio on the original scale.)
A 95 percent confidence interval for the difference in concentration before
and after cleaning was calculated by using the error mean square of the
18
-------
analysis of variance. Results were transformed back to the original scale
for reporting purposes.
Air Samples :
Airborne asbestos concentrations were determined before and;during
carpet cleaning to study the effect of the cleaning method and contamination
loading on fiber reentrainment during carpet cleaning. Three work-area
samples were collected before and during carpet cleaning for each experiment.
A single estimate of the airborne asbestos concentrations beforejand during
cleaning was then determined by averaging the three respective work-area
samples. The natural logarithm of each of the concentrations was used for
subsequent statistical analyses. This is equivalent to assuming;that the
data follow a lognormal distribution. A two-factor ANOVA with cleaning
method (wet, dry) and contamination level (low, high) as the experimental
factors was performed on the difference (on the log scale) between the
concentration before cleaning and the concentration during cleaning.
19
-------
SECTION 5
EXPERIMENTAL PROCEDURES
PRESTUDY AIR MONITORING
Before construction of the contamination enclosure system,
air samples
were collected to determine a baseline airborne asbestos concentration inside
the test facility. Seven interior air samples and two field blanks were
collected in accordance with sampling procedures described in Section 4. The
air samples were collected for a period of approximately 200 minutes to
achieve a minimum air volume of 1260 liters for each sample. These samples
were analyzed in accordance with the nonmandatory TEM method, as described in
the AHERA final rule. • ,
The average airborne asbestos concentration for the seven samples col-
lected was 0.0031 s/cm3. The TEM analysis of the seven samples yielded a
total of 6 asbestos structures (4 chrysotile and 2 amphibole). One chryso-
tile fiber was detected on each field blank. Table 8 summarizes; these re-
sults, i
TABLE 8. SUMMARY OF PRESTUDY AIRBORNE ASBESTOS '
CONCENTRATIONS IN TEST FACILITY
Sample
001
002
003
004
005
006
007
Field blank
Field blank
Number of
structures
observed
1
0
2
0
1
1
1
1
1
Concentration,
s/cm3
0.0028
<0.0039
0.0077
<0.0038
0.0039
0.0039
0.0038
20
-------
CARPET CONTAMINATION ;
l
Selected levels of carpet contamination for this study were based on
field data reported by Wilmoth et al.1* Asbestos concentrations ranging from
approximately 8000 s/ft2 to 2 billion s/ft2 were detected in the contaminated
carpet by use of a microvac technique. Bulk sample sonication of: the samples
revealed levels ranging from 30 million to 4 billion s/ft2. Based on these
preliminary results, the target experimental asbestos contamination levels of
approximately 100 million and 1 billion s/ft2 were believed to represent
carpet contamination likely to be present in buildings where asbestos-con-
taining materials are present. I
The carpet was contaminated with a spray-applied dispersion of Union
International Centre le Centre Calidria chrysotile asbestos in distilled
water. The asbestos was dispersed uniformly on the carpet by use; of a manual
pesticide sprayer equipped with a stainless steel container. :
Preparation of Concentrated Aqueous Suspensions of Chrysotile ;
Aqueous suspensions of chrysotile are not stable for long periods unless
they are specially prepared.5 Even small amounts of high-molecular-weight
organic materials, such as those generated by bacteria, result in the desta-
bilization of chrysotile suspensions and the attachment of fibers; to the
walls of the container. This process can be reversed only by carrying out
oxidation of the organic materials with ozone and ultraviolet light treat-
ment.5 If precautions are taken to exclude all organic materials! and to
prevent bacterial growth, however, chrysotile suspensions can be prepared
that remain stable for several years. This can be achieved by sterilizing
all containers used in the preparation!, using freshly distilled water for the
dispersion process, and storing the preparation in flame-sealed glass ampules
that are autoclaved immediately after sealing.
For this project, the decision was made to prepare sealed ampules of
fiber dispersions so that the contents of one ampule dispersed in; 6 liters of
freshly distilled water would provide the concentration of suspension re-
quired for artificial contamination of one 500-ft2 sample of carpiet. Calcu-
lations of the amount of chrysotile required were based on the assumption
that all of the fibers needed to contaminate one carpet sample wo|uld be
contained in a volume of 50 ml sealed in one ampule.
For the higher of the two concentrations used, the fiber concentration
required in each ampule was calculated as follows: :
Higher contamination level required
Number9of fibers required to contaminate
500 f r
Fiber concentration required for this
number of fibers to be in a volume of 50 ml
= 109 fibers/ft2
= 6.5 x 1011 fibers
1.3 x 1013 fibers/liter
21
-------
The lower of the two concentrations used was a factor of 10:lower than
this. As a way of ensuring an exact factor of 10 ratio between the two
concentrations, the lower-concentration dispersion was prepared by diluting
an aliquot of the high-concentration dispersion. '
Because the original suspension was to be prepared by dispersing a known
weight of chrysotile in water, knowledge of what numerical concentration of
fibers would result from this dispersion was required. Previous'work on
preparation of ampules indicated that a suspension of purified Calidria
chrysotile in water with a mass concentration of 1 yg/liter yielded a numer-
ical fiber concentration of approximately 200 million fibers per:liter.
Based on this conversion, the weight of chrysotile is calculated:as follows:
Weight required = 1.3 x 1013 x 10"6/(2 x 108) g/liter :
= 65 mg/liter :
i
Therefore, the preparation of 1.5 liters of a suspension with this concentra-
tion requires 97.5 mg of chrysotile. i
i
The calculation for determining the mass of chrysotile required is based
on data from very dilute suspensions. Initial experiments indicated that
some difficulty could arise in obtaining complete dispersal of tfjie chrysotile
at the high concentrations in this program; if some aggregation were to
occur, the numerical structure count would be somewhat lower than that re-
quired. For this reason, the suspensions were prepared to have a higher mass
concentration than that indicated in the preceding calculation. ;
Before the fiber suspensions were prepared, the 50-ml ampules were
thoroughly cleaned. Each ampule was filled to the top with freshly distilled
water and placed in an ultrasonic bath for a period of 15 minutes; the water
was then removed by suction. This process was repeated twice before the
ampules were considered ready for filling. i
The higher-concentration chrysotile suspension was preparedifirst. All
water used for preparation of these dispersions was freshly distilled (within
8 hours of preparation). A weight of 409.5 mg of purified Calidna chryso-
tile was placed in an agate mortar and lightly ground with a small volume of
water by use of a pestle. More freshly distilled water was added gradually
until a creamy liquid was obtained. Up to 400 ml of this liquidjwas made up
in a disposable polypropylene beaker, and the beaker was placed in an ultra-
sonic bath for approximately 30 minutes. Up to 1500 ml of the chrysotile
suspension was then made up with water in a 1-gallon polyethylene bottle.
The bottle was placed in an ultrasonic bath for approximately 30 |minutes,
during which time the bottle was removed several times and shaken vigorously.
The lower-concentration suspension (a volume of 150 ml) was made :up to 1500
ml with water in another 1-gallon polyethylene bottle. The two suspensions
had concentrations of 273 and 27.3 mg/liter, respectively. i
A disposable polyethylene funnel was used to place a volume 'of 50 ml of
suspension in each of the ampules. This left adequate space in the ampule to
22
-------
permit efficient shaking of the contents. The filled ampules were immedi-
ately flame-sealed and then autoclaved for 30 minutes at a temperature of
121°C to sterilize the contents. After the ampules cooled, they Iwere labeled
in the order of their filling. ;
!
Preparation of Asbestos Dispersion :
The following steps were followed precisely in the preparation of the
asbestos dispersions used to contaminate the carpet: |
i
1. All water used for dilution of the ampules of chrysotile suspension
was freshly distilled from a glass still. :
._ . ..... ! . .
2. Before the ampule was opened, it was shaken vigorously :for 1 minute
and then placed in an ultrasonic bath for 30 minutes. 'During the
ultrasonic treatment, the ampule was removed every 5 minutes and
again shaken vigorously for 1 minute. ;
3. A new 32-ounce glass bottle was washed with several changes of
freshly distilled water. The ampule was then opened, ctnd the
entire contents were emptied into 450 ml of freshly distilled water
in the glass bottle. For the high-concentration ampules only, the
pH was adjusted to approximately 4.0 by adding 300 to 400 yl of
glacial acetic acid. The bottle was capped, shaken vigorously, and
then placed in an ultrasonic bath for 15 minutes. No surface-
active agents were added. ',
4. The pesticide sprayer was sterilized and cleaned by rinsing it with
a 10 to 15 percent solution of Clorox for approximately 15 minutes.
The sprayer, including the interior of the outlet pipe^ was then
thoroughly washed with several changes of freshly distilled water.
I
5. The sprayer was filled with 5.5 liters of freshly distilled water,
and the contents of the bottle were added. The sprayer was then
shaken before the carpet was sprayed.
The sprayer was not allowed to dry before it was washed after each
experiment because chrysotile is much more difficult to remove frfom the
interior surfaces when it has dried. !
To ensure that no bacterial growth had occurred in the sprayer between
uses, the inside of the sprayer and the outlet pipe were treated |with a 10 to
15 percent solution of Clorox to remove any bacteria and their byproducts.
Any bacterial growth would scavenge fibers from the suspension and cause
fibers to become attached to the wall of the container. The container and
outlet pipe were then rinsed with isopropyl alcohol.
Concentrations of Suspensions
Several of the ampules were used to make precise measurements of the
fiber concentrations and to determine the fiber size distributions. To
23 ;
-------
measure these very high fiber concentrations required a total dilution factor
of 1 in 25,000 for the low-concentration ampules and 1 in 250,000; for the
high-concentration ampules. This was achieved by successive dilutions in
freshly distilled water. For the low-concentration ampules, the contents of
one ampule were first dispersed in 500 ml. In the second dilution, 10 ml
was diluted to 500 ml, and 10 ml of this second dilution was then: diluted to
500 ml. Three filters were prepared from this final suspension in accordance
with the EPA Analytical Method for Determination of Asbestos Fibers in
Water.6 For the high-concentration ampules, the final suspension; was diluted
by a further factor of 10 before the filters were prepared. ;
The dilution factors and the volumes of suspension filtered were select-
ed to yield fiber counts of approximately 40 per grid opening. One fiber
count incorporating approximately 600 asbestos structures was made for each
of the two concentrations.
The high-concentration ampules yielded asbestos structure counts signi-
ficantly lower than those obtained during the initial tests on the suspension
at the time the ampules were prepared. This effect was investigated and
found to have been caused by a rise in pH of the suspension after packing and
autoclaving. The increase in the pH was probably due to some leaching of the
chrysotile during the autoclave treatment, which caused destabilization of
the dispersion and aggregation of the fibers into bundles and clusters. The
effect was found to be reversible by adjusting the pH of the dispersion to
approximately 4.0 with acetic acid at the time of the first dilution. The
measurements on the high-concentration ampules were repeated; another ampule
was used and the pH was adjusted during preparation of the first dilution.
The aggregation effect did not occur in the low-concentration ampules; there-
fore, no pH adjustment was required when these ampules were used.;
Table 9 shows the results of the fiber concentration measurements made
on the low- and high-concentration ampules. The analysis of the laboratory
dilution was continued for approximately 600 chrysotile structures to provide
a precise concentration value and a size distribution with a sufficient
number of structures in each size classification. Appendix B contains the
size distributions for the measurements made on the low- and high-concentra-
tion ampules. Figure 5 shows the fiber size distribution in the low- and
high-concentration ampules.
Application of Dispersion to Carpet \
A meticulously cleaned hand-pumped garden sprayer was used to apply the
asbestos dispersion to the carpet. A fixed number of pumps was used for each
batch to provide consistent spray pressure. The desired controlled spray was
experimentally determined by trial and error before the tests with asbestos
began. The pressure was kept within the desired range by adding a fixed
number of pump strokes after each fixed area was sprayed in a predetermined
pattern by following a grid work of string placed over the carpet;before the
24
-------
160-
100.0
Low-Concentration
Suspension
Total Fibers 619
1.0 10.0
Fiber Length, micrometers
High-Concentration
Suspension
Total Fibers i601
1.0 10.0
Fiber Length, micrometers
100.0
Figure 5. Distribution of chrysotile fiber lengths in the low and high concentration aqueous asbestos
suspensions.
25
-------
beginning of each experiment. The tank was periodically agitated to help
keep the asbestos fibers suspended. Dehumidifiers were placed in the room
overnight to aid in drying the carpet. The following day a 200-pound steel
lawn roller was rolled over the carpet surface to simulate the effects of
normal foot traffic in working the asbestos into the carpet. i
TABLE 9. SUMMARY OF RESULTS OF TRANSMISSION ELECTRON MICROSCOPY
ANALYSES FOR LOW- AND HIGH-CONCENTRATION AMPULES '
12
Structure concentration, 10 structures/liter
Sample
description
Equivalent No. of
95% con- volume struc-
fidence Analytical sampled, tures
Fiber type Mean interval sensitivity yl ! counted
Low-concentra-
tion ampule
Chrysotile 2.2 2.0-2.5
High-concentra- Chrysotile 25
tion ampule
22-27
0.0036
0.0409
0.400
0.040
619
601
Carpet Cleaning Technique j
The carpet was vacuumed or wet-cleaned for a period of approximately 65
minutes to allow the collection of a sufficient volume of air samples.to
obtain an analytical sensitivity of 0.005 s/cm3 of air. The carp.et was
cleaned in two directions, the second at a 90-degree angle to the first.
DISPOSAL OF ASBESTOS-CONTAINING MATERIAL i
i
Asbestos-contaminated materials, including carpeting, polyethylene, pro-
tective clothing, etc., were placed in disposable 6-mil polyethylene bags and
labeled according to EPA regulations. When filled, the disposal bags were
sealed, sponged clean, and moved from the test room to the primary waste-
loadout work area (Figure 1). The disposal bags were then sponge'd a second
time,, taken through the equipment-change area, and placed in the 'shower
chamber for a thorough washing. The cleaned disposal bags were t^aken into
the clean chamber, loaded into a fiberboard drum, labeled with an EPA-
approved asbestos warning label, and transported to a disposal site approved
by the Ohio EPA. :
SITE CLEANUP
Prior to removal of the primary polyethylene barrier (i.e., 'the first
barrier installed to isolate the work area, including test rooms), the sur-
face was thoroughly wet-wiped with amended water. The HEPA filtration system
continued to operate during site cleanup. :
26 !
-------
All debris and waste resulting from the experiments were removed from
the building. All the drummed waste was removed from the site and disposed
of in a landfill approved by the Ohio EPA.
POSTSTUDY AIR MONITORING
After removal of the polyethylene sheeting from the floor, ceiling, and
walls, air samples were collected to determine the airborne asbestos concen-
trations inside the building. Four interior air samples were collected in
accordance with the sampling procedures described in Section 5. iThese sam-
ples were collected for a period of approximately 180 minutes to lachieve a
minimum air volume of approximately 1800 liters for each sample. ' These
samples were analyzed in accordance with the nonmandatory TEM method as
described in the AHERA Final Rule. No asbestos was detected in any of these
samples.
27
-------
SECTION 6 j
QUALITY ASSURANCE :
The Quality Assurance Project Plan (QAPP) contains complete'details of
the quality assurance procedures followed during this research project. The
procedures used for this study are summarized in the following subsections.
SAMPLE CHAIN OF CUSTODY j
Sample chain-of-custody procedures were an integral part of'both sam-
pling and analytical activities during this study. They were followed for
all air and bulk samples collected. The applied field custody procedures
documented each sample from the time of its collection until its jreceipt by
the analytical laboratory. Internal laboratory records then documented the
custody of the sample through its final disposition. |
Standard sample custody (traceability) procedures were used. Each
sample was labeled with a unique project identification number, which was
recorded in the field log book along with other information specified.by the
V^M r r «,
QUALITY ASSURANCE SAMPLE ANALYSES I
Specific quality assurance procedures for ensuring the accuracy and pre-
cision of the TEM analyses of carpet samples included the use of laboratory
blanks and duplicate counting. !
Laboratory Blanks j
A sample blank was prepared and analyzed for every 10 carpet samples
analyzed. Each blank was prepared in a manner identical to that used for the
carpet samples, although no carpet segment was actually used. Thjese blanks
served as a quality control check on contamination from the solutions, glass-
ware, filters, and handling procedures. Analysis of 10 TEM grid openings per
blank showed all laboratory blanks to be free of asbestos fiber contami-
nation. ;
Duplicate Sample Analyses ;
Duplicate sample analysis provides a means of quantifying any analytical
variability introduced by the preparation procedure and refers to:the analy-
sis of a second preparation of the sample by the same microscopist. Thirteen
samples were randomly selected for duplicate analysis.
28 !
-------
The coefficient of variation for duplicate analyses was estimated by
assuming a lognormal distribution for data on the original scale land
estimating the variance on the log scale. For a random variable iX with a
lognormal distribution, the relationship between the coefficient iof variation
(CV) of X and the variance (a2) of Y = log X is given by j
CV =
The variance was estimated by the error mean square obtained from a one-way
ANOVA of log concentration with sample ID as the experimental factor.
The error mean square for the ANOVA on the 13 duplicate QC samples is
0.066, which corresponds with an estimated coefficient of variation of 0.26.
This compares with a coefficient of variation of 0.13 estimated in the preci
sion study conducted during the design stage of the experiment. iBecause the
precision study included only one carpet contamination level (100 million
s/ft2) and no vacuuming treatment, a higher coefficient of variation for the
experimental data is not unexpected. i
Table 10 presents the results of the duplicate analyses. ;
TABLE 10. RESULTS OF DUPLICATE SAMPLE ANALYSES !
Original
Duplicate
Sample
02-B017B
03-B026B
05-B051B
07-B073B
08-B090B
10-B110B
11-B125B
13-B147B
15-B171B
18B-1B
19B-1B
21B-2A
23B-1B
N
20
14
50
17
12
116
147
215
113
19
18
9
65
s/ft2
27,512,900
20,210,093
70,328,163
30,134,309
34,284,979
303,907,233
594,511,529
923,308,634
412,908,443
51,103,713
56,008,403
157,855,263
218,304,359
N
39
15
50
27
9
135
131
204
114
14
30
16
85
s/ft?
i
53,650,155
21,653,671
70,328,163
47,860,372
25,713,735
530,527,712
529,802,791
876,069,588
476,071,429
37,655,367
93,347,339
248,000,000
285,474,931
TEM ANALYSIS OF UNUSED SAMPLE CONTAINERS
Eleven unused, wide-mouth, polyethylene, screw-cap sample containers
were analyzed for background asbestos contamination. Laboratory preparation
was identical to that used for carpet samples, except no carpet segment was
used. All 11 unused sample containers were found to be free of asbestos
fiber contamination.
29
-------
SPRAY-APPLICATION TECHNIQUE
To confirm the validity of the spraying technique, an additional experi-
ment was conducted with a pesticide sprayer identical to those used to apply
the chrysotile to the carpet samples. An ampule of low-concentration suspen-
sion was diluted to 500 ml and then further diluted to 6 liters i:n the pesti-
cide sprayer by using freshly distilled water. The sprayer was thoroughly
shaken, and the contents were sprayed out into several containers;. Three
500-ml samples of the spray were collected, one at the beginning of the
spraying, one when approximately 50 percent of the contents had b|een dis-
charged, and one just before the end of the spraying. These three samples
were analyzed to establish that the concentration and size distribution of
the fibers did not change during the spraying period. The results are pre-
sented in Table 11. These results indicate no significant loss of fibers
during the transfer of the diluted liquid suspension through the ;sprayer's
hose and nozzle. ;
TABLE 11. RESULTS FROM PRELIMINARY STUDY OF ASBESTOS DISPERSION BY
SPRAYING—FIBERS AND FIBER BUNDLES (ALL LENGTHS)
T-r-rrnr—.-.---. _ TL J-_,_L, ,_-.-_.,:_ - - - - ,--,..--, -. -. ,-_. -L..-. -,_ -. -_..-. --.- -. ,. - . - -, -^ -- -.-__-,-..--,__,__.- _
Structure concentration, !
1012 structures/liter
Volume in sprayer
at time of sample
collection, liters
6
Fiber type
Chrysotile
Mean
2.33
95% con-
fidence
interval
1.87-2.79
i
Analytical :
sensitivity j
0.0118
Number of
structures
counted
198
(Beginning of spray)
4 Chrysotile
(50% point of spray)
2.18 1.54-2.82 0.0118
*w
(End of spray)
Chrysotile 2.38 1.90-2.85 0.0118
185
202
The size distributions for these samples are listed in Table 12 and il-
lustrated in Figure 6. Because the distributions are all approximate loga-
rithmicnormal, the size range intervals for calculation of the distribution
must be spaced logarithmically. Another requirement for the choice of size
intervals is that they allow for a sufficient number of size classes while
still retaining a statistically valid number of fibers in each cl'ass.
Interpretation is also facilitated if each size class repeats at ciecade
intervals. A ratio of 1.468 from one class to the next satisfies: all of
these requirements. The other constraint is that the length distribution
should include the minimum fiber length of 0.5 um at the first interval
point. The decade repeating automatically ensures that the other; significant
fiber length of 5 ym occurs as an interval point. >
!
!
No significant change in the fiber size distribution was evident during
the transfer of the diluted liquid suspension.
30
-------
TABLE 12.
Particle
size range, ym
0.23-0.34
0.34-0.50
0.50-0.73
0.73-1.08
1.08-1.58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
FIBER LENGTH DISTRIBUTIONS
ASBESTOS DISPERSION
FROM THE PRELIMINARY
BY SPRAYING
Number of fibers, fiber bundles (cumulative
Beginning of spray
0 (0)
0 (0)
28 (14.14)
48 (38.38)
34 (55.56)
30 (70.71)
34 (87.88)
18 (96.97)
4 (98.99)
1 (99.49)
1 (100.00)
0 (100.00)
0 (100.00)
0 (100.00)
50% point of spray
0 (0)
0 (0)
33 (17.84)
55 (47.57)
28 (62.70)
20 (73.51)
17 (82.70)
14 (90.27)
10 (95.68)
5 (98.38)
3 (100.00)
0 (100.00)
0 (100.00)
0 (100.00)
31
STUDY OF
percentage)
End of spray
0 (0)
0 (0)
24 (11.88)
43 (33.17)
45 (55.45)
28 (69.31)
22 (80.20)
19 (89.60)
13 (96.04)
5 (98.51)
1. (99.01)
1 (99.50)
0 (99.50)
1(100.00)
-------
1 *
EC CC
£L Q.
CO CO
U- U. >T
O O <
O H Q-
2 Z CO
2 O £
2 O. O
O ^S Q
UJ O 2
CD in HI
0>
10
CO
O
CD
>
i_
O
CO
xi
O
CO
i_
0)
CD
JQ
E
w,
V.
0).
•*-»
I
O
i_
O
I
<0!
DC;
o:
^i
CO,
Q)',
O
to
a.
CO
o
CO
10
CM
o
CM
k\\\\\\\\\m\\\\\\\\\\%\^^^^^
LO
O)
I
Q.
CO
CO
Q.
CO
CO
1
0)
T3
i
. CO
Q.
8
o
N
CO
CD
^3
il
CD
I
D>
il
32
-------
SECTION 7
RESULTS AND DISCUSSION
CARPET SAMPLES !
I
I
Figure 7 illustrates the average (geometric mean) concentrations of
asbestos structures in the carpet before and after cleaning. The 95 percent
confidence intervals for the geometric mean concentrations are given in Table
13. Individual estimates of carpet contamination are listed in Appendix C.
For each experiment, a single estimated concentration was obtained before and
after cleaning by taking the arithmetic mean of the three individual esti-
mates. This gave 24 pairs of concentrations, one for each experiment. These
estimates are presented in Appendix D. ;
TABLE 13. SUMMARY STATISTICS FOR ASBESTOS CONCENTRATIONS
IN CARPET BEFORE AND AFTER CLEANING ',
Approximate
contamination HEPA-filtered
level, s/ft2 cleaner
100 million
1 billion
a Each data
Before
Hot-water extraction
Dry vacuum
After
Hot-water extraction
Dry vacuum
Before
Hot-water extraction
Dry vacuum
After
Hot-water extraction
Dry vacuum
Number
of data
points
cleaning
6
6
cleaning
6
6
cleaning
6
6
cleaning
6
6
point represents the average of three
Geometric mean.;
million s/ft2
!
62 I
47 i
18 ;
56 !
589 |
535
i
;
196
447
carpet samples.
95 percent
confidence
interval
(39, 101)
(37, 59)
(8, 43)
(38, 83)
(397, 873)
(356, 803)
(85, 449)
(240, 832)
33
-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
s s
(SUOJHILU) }OOj
o
o
o
o
CO
CM
O
O
ssjnjonjjs sojsaqsv
34
D)
CD
D)
C
"E
I
O (O
t- c
CJ -r-
^a E
OS
O O
•r- O
4->
ra •)-»
s- a>
+J Q.
s: s-
c:
o ~a
o c
to
4-3
a; "o
ci- o
s.. ^i
ctS 4->
U d)
o en
OJ E
_C1 CO
« r—
o
01
O5.E
ro O
S- 03
CD CU
I—
S-
-------
Results of a three-factor ANOVA indicated no significant difference
between the results from Experiments 1 through 16 and Experiments 17 through
24 (p=0.7). The difference between the two sets of experiments Was that the
carpet in Experiments 17 through 24 was first dry-vacuumed, then iwet-cleaned,
and then dry-vacuumed again prior to contamination. Because no significant
difference was evident in the asbestos-retention characteristics |of the new
carpet versus new carpet that had first been wet-cleaned, the data from all
24 experiments were treated equivalently and reanalyzed by using ja two-factor
ANOVA. |
Results of the two-factor ANOVA are presented in Table 14. The type of
cleaning method had a significant effect (p<0.001) on the difference between
the asbestos concentrations before and after cleaning. The level of asbestos
contamination in the carpet had no significant effect (p=0.622). | The esti-
mated asbestos concentration in the carpet after cleaning, expressed as a
proportion of the asbestos concentration before cleaning, is given in Table
15 and illustrated in Figure 8 together with 95 percent confidence intervals.
TABLE 14. ANALYSIS OF VARIANCE TABLE FOR DIFFERENCE BETWEEN
ASBESTOS CONCENTRATIONS BEFORE AND AFTER CLEANING !
Source of variation
Contamination level
Cleaning method
Interaction
Error
Degrees
of freedom
1
1
1
20
Sum of
squares
0.074
8.174
0.362
5.872
F value
0.251
27.840
1.232
!P value
; 0.622
,<0.001
: 0.280
i
TABLE 15'. ESTIMATED ASBESTOS CONCENTRATION IN CARPET AFTER
CLEANING AS A PROPORTION OF THE CONCENTRATION BEFORE CLEANING
Contami-
nation
loading
Low
High
HEPA-filtered vacuum
Hot-water extraction
Dry-vacuum
Hot-water extraction
Dry-vacuum
Concentration after
cleaning as a pro-
portion of concentra-
tion before cleaning
0.29
1.19
0.33
0.84
95 percent
confidence
interval
(0.16, 0.51)
(0.68, 2.11)
(0.19, 0.59)
(0.47, 1.48)
The asbestos concentration in the carpet after wet cleaning was approxi-
mately 0.3 of the asbestos concentration before cleaning in both the high and
low contamination levels. The upper 95 percent confidence limit (Table 15)
at each contamination level is less than 1, which indicates this pis a statis-
tically significant reduction. '
35 ;
-------
O)
c
'c
ca
Q>
O
0)
0
"£
CO
o
O)
C
'ir
(0
__
O
j[>
H-
t:
o
C3
J;;
c:
0)
o
cz
o
O
co
a>
Q)
.Q
CO
,.!
o
c:
r
o
CL
o
u,
Q.
IT
cv
1 I I '•
i
? ;
•,
i t -
: ! ; :
t ;
' : I '.
\ \ i
i ( ;
! i ; '
; ;
!
i
• 9
i
1
i
i
1
1
i
j •
; i i
> CM IO i- 10 c
i - o
E
3
o =
> -^w
' CO J.
>> c >>.
0 1 *
*- o
o E
« ^ "°
0) ^ -r-
O ra
i x
0)
5
E
^ fc
O C£CM
> -- £
>. ^° ^0
Q 'E c
ll
c *~
o c
« 00
*> 50
O O 1~
j. -j
CD
5
>
to
cu
J3
V5 Q)
re j:
4->
QJ
-E M-
•»-> o
s- c
o o
<+- T-
4-J
trt S-
i— O
n3 Q.
> o
S_
CX
O) re •
o en
E en c
ID c •!-
-O -r- C
•r- c re
M- re ID
C OJ •—
O r— (J
o o
OJ
4-> S- i-
C QJ O
O) +j q_
O 4- O)
i- re .0
QJ
Q- c: c:
o o
0) -r- T-
> 4-> +->
••- re re
t- t- s-
I 4-> 4->
>5 E E
+J Q) OJ
Q) O O
C C C
•r- O O
•z. o o
00
cu
CD
36
-------
The asbestos concentration in the carpet after dry-vacuuming was 1.2
times the concentration before cleaning for the low-contamination treatment
and 0.8 times the concentration before vacuuming for the high-contamination
treatment. The 95 percent confidence intervals for both estimated include
1, which indicates the data do not provide statistically significant evidence
of either an increase or a decrease in asbestos concentration after dry
vacuuming. ;
i
Asbestos Fiber Distributions in Carpet i
The TEM analysis of the 144 carpet samples before and after cleaning
yielded a total of 8101 asbestos structures. Of this total, 8080; (99.7%)
were chrysotile and 21 (0.3%) were amphibole. The presence of amphibole
asbestos fibers in the carpet was probably due to conditions existing prior
to the study. Prestudy air monitoring identified two amphibole asbestos
fibers in seven air samples collected. The structure morphology distribution
for the particles in the carpet samples is summarized in Table 16L
TABLE 16. STRUCTURE MORPHOLOGY DISTRIBUTION IN CARPET SAMPLES
COLLECTED BEFORE AND AFTER CARPET CLEANING ;
Structure
type
Chrysotile
Amphibole
Number of
bundles
1763
2
Number of
clusters
66
0
Number of
fibers
5893
18
Number of
matrices
358
1
; Total
8080
; 21
Total
1765
66
5911
359
8101
Appendix E presents the structure-length distributions of asbestos
particles found in the carpet before and after cleaning. Figure 9
illustrates the cumulative percentage of fibers, for varying fiber lengths,
observed 1) in the air during carpet cleaning activities, 2) in the carpet
after dry-vacuuming and wet-cleaning, and 3) in the asbestos suspension used
to contaminate the carpet. For carpet contaminated with 100 million s/ft2,
a higher percentage of larger residual particles were consistently observed
in the carpet after dry-vacuuming than after wet-cleaning. Fiber;lengths of
the residual asbestos in the carpet after dry-vacuuming and wet-cleaning
carpet contaminated with 1 billion s/ft2 were comparable. The reason for the
difference in results between the two contamination levels is unknown.
i
|
AIR SAMPLES \
Airborne asbestos concentrations were determined before and during car-
pet cleaning in Experiments 1 through 16 to study the effect of the cleaning
method and contamination loading on fiber reentrainment during carpet clean-
ing. For each experiment, three work-area samples were collected:before and
37
-------
100
% of Fibers
Low Carpet Contamination
100 million s/ft
Airborne
Carpet - Dry Vacuum
Carpet - Wet Clean
Asbestos Suspension
100
>2.3 >3.4
Fiber Length, micrometers
>5.0
>7.3
>.5
>2.3 >3.4 >5.0
Fiber Length, micrometers
>7.3
>10.8
High Carpet Contamination,
1 billion s/ft2
Airborne
Carpet - Dry Vacuum
Carpet - Wet Clean
Asbestos Suspension
>10.8
Figure 9. Cumulative percentages of asbestos particles in carpet
after cleaning, airborne asbestos particles observed
during cleaning, and asbestos fibers used to contaminate
the carpet.
38
-------
during the carpet cleaning. Figure 10 presents the average airborne asbestos
concentrations measured before and during cleaning for each cleaning
method and carpet contamination loading. The samples collected before clean-
ing were obtained after the carpet was contaminated to determine the baseline
concentration in the test room. i
The type of cleaning method had no significant effect (p=0.5;8) on the
difference between the airborne asbestos concentrations before arid during
cleaning. Similarly, the level of asbestos contamination in the carpet had
no significant effect on fiber reentrainment (p=0.09). Overall, however, the
mean airborne asbestos concentration during carpet cleaning was significantly
higher during carpet cleaning than just prior to cleaning (p<0.001). A 95
percent confidence interval for the mean airborne asbestos concentration
during carpet cleaning as a proportion of the airborne concentration before
cleaning showed that the mean airborne asbestos concentration wasi between two
and four times greater during carpet cleaning. ;
i
Figure 9 also illustrates that asbestos fibers in the air during carpet
cleaning activities tended to be smaller in length than the asbestos fibers
remaining in the carpet after cleaning. For example, overall approximately
17 percent of the asbestos fibers found in the carpet were less than 1.0 ytn
in length; whereas approximately 85 percent of the fibers observed in the air
were less than 1.0 ym. ;
39
-------
i
=5
-o
T3
C
«o
£
.0
l*-
cu
c
o
to
OJ
o
o
o
in
o
O)
s- c
O t-
(O
0)
O
O)
to re
40
-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
REFERENCES
1. Kominsky, J. R., and R. W. Freyberg. Asbestos Fiber Reentrainment
During Dry Vacuuming and Wet Cleaning of Asbestos-Contaminated Carpet.
U.S. Environmental Protection Agency, Risk Reduction Engineering
Laboratory, Cincinnati, Ohio. Contract Number 68-03-4006, Final Report.
1989.
2. Yamate, 6., S. C. Agarwal, and R. D. Gibbons. Methodology f0r the
Measurement of Airborne Asbestos by Electron Microscopy. Draft Report.
U.S. Environmental Protection Agency, Office of Toxic Substances,
Washington, D.C. EPA Contract No. 68-02-3266. 1984. ;
3. Neter, J., W. Wasserman, and M. H. Kutner. Applied Linear Statistical
Models. 2nd Ed. Richard D. Irwin, Inc., Homewood, Illinois, 1985.
4. Wilmoth, R., T. J. Powers, and J. R. Millette. Observations' in Studies
Useful to Asbestos O&M Activities. Presented at the National Asbestos
Canal Conference in Atlanta, Georgia, February 1988. ;
5. Chatfield, E. J., and M. J. Dillon. Analytical Method for Determination
of Asbestos Fibers in Water. PB 83-260-471. U.S. Environmental Re-
search Laboratory, Athens, Georgia. Contract 68-03-2717. National
Technical Information Service, Springfield, Virginia. 1983.:
6. Chatfield, E. J., M. J. Dillon, and W. R. Stott. Development of Im-
proved Analytical Techniques for Determination of Asbestos in Water Sam-
ples. PB 83-261-471. U.S. Environmental Research Laboratory, Athens,
Georgia. 1983.
41
-------
APPENDIX A
TRANSMISSION ELECTRON MICROSCOPY RESULTS FROM
PRELIMINARY PERFORMANCE EXPERIMENTS ON
THE MICROVAC AND SONIC EXTRACTION PROCEDURES
42
-------
Carpet
Section
E-l
E-l
E-l
E-2
E-2
E-2
E-3
E-3
E-3
E-4
E-4
E-4
E-5
E-5
E-5
E-6
E-6
E-6
Asbestos Structures
Total Number
SONIC EXTRACTION
109
103
104
100
106
101
117
100
107
108
100
103
119
107
104
100
100
106
Per Square Foot
6.55 X 108
8.26 X 108
6.64 X 108
4.78 X 108
4.56 X 108
4.28 X 108
1.36 X 109
1.16 X 109
1.24 X 109
6.41 X 108
5.93 X 108
4.89 X 108
1.33 X 109
9.57 X 108
1.16 X 109
5.42 X 108
6.62 X 108
6.53 X 108
43
-------
Carpet
Section
M-l
M-l
M-l
M-2
M-2
M~3
M-3
M-3
M-4
M-4
M-4
M-5
M-5
M-5
M-6
M-6
Asbestos
Total Number
MICROVAC
106
33
12
11
6
9
20
24
7
5
9
114
128
92
24
23
Structures
Per Square Foot
3.41 X 107
2.10 X 107
1.52 X 107
4.54 X 106
2.48 X 106
3.72 X 106
8.26 X 106
9.94 X 106
2.89 X 106
. 2.06 X 106
3.72 X 106
4.71 X 107
1.06 X 108
1.27 X 108
9.94 X 106
9.48 X 106
44
-------
APPENDIX B
CHRYSOTILE FIBER SIZE DISTRIBUTION
IN THE HIGH- AND LOW-CONCENTRATION AMPULES
45
-------
APPENDIX B
CHRYSOTILE FIBER SIZE DISTRIBUTION
IN THE HIGH- AND LOW-CONCENTRATION AMPULES
TABLE B-l. FIBER LENGTH DISTRIBUTION IN THE
LOW CONCENTRATION AMPULE
Particle
size range, ym
0.23
0.34
0.50
0.73
1.08
1.58
2.32
3.41
5.00
7.34
10.77
15.81
23.21
34.06
50.00
73.40
107.70
158.10
232.10
- 0.34
- 0.54
- 0.73
- 1.08
1.58
- 2.32
- 3.41
- 5.00
- 7.34
- 10.77
- 15.81
- 23.21
- 34.06
- 50.00
- 73.40
- 107.70
- 158.10
- 232.10
- 340.60
Number
of fibers
counted
0
0
107
147
106
90
69
57
26
11
5
0
1
0
0
0
0
0
0
Cumulative
fiber count
0
0
107
254
360
450
519
576
602
613
618
618
619
619
619
619
619
619
619
Percent
of total
0.00
0.00
17.29
.23.75
17.12
14.54
11.15
9.21
4.20
1.78
0.81
0.00
0.16
0.00
0.00
0.00
0.00
0.00
0.00
Cumulative
percent
0.00
0.00
i 17.29
: 41.03
'• 58.16
' 72.70
1 83.84
! 93.05
; 97.25
: 99.03
' 99.84
i 99.84
100.00
100.00
100.00
100.00
100.00
100.00
100.00
46
-------
TABLE B,-2. FIBER LENGTH DISTRIBUTION IN THE
HIGH CONCENTRATION AMPULE
Particle
size range, ym
0.23
0.34
0.50
0.73
1.08
1.58
2.32
3.41
5.00
7.34
10.77
15.81
23.21
34.06
50.00
73.40
107.70
158.10
232.10
- 0.34
- 0.54
- 0.73
- 1.08
- 1.58
- 2.32
- 3.41
- 5.00
- 7.34
- 10.77
- 15.81
- 23.21
- 34.06
- 50.00
- 73.40
- 107.70
- 158.10
- 232.10
- 340.60
Number
of fibers
counted
0
0
101
135
119
85
82
40
20
16
3
0
0
0
0
0
0
0
0
Cumulative
fiber count
0
0
101
236
355
440
522
562
582
598
601
601
601
601
601
601
601
601
601
Percent
of total
0.00
0.00
16.81
22.46
19.80
14.14
13.64
6.66
3.33
2.66
0.50
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Cumulative
percent
• 0.00
i 0.00
: 16.81
' 39.27
' 59.07
: 73.21
; 86.86
93.51
196.84
199.50
100.00
100.00
100.00
100.00
100.00
ioo.oo
ioo.oo
ioo.oo
100.00
47
-------
APPENDIX C
CARPET ASBESTOS CONCENTRATIONS
BEFORE AND AFTER CARPET CLEANING
NOTE: Sample numbers ending with "B" indicate that the
sample was taken before carpet cleaning; those
ending with an "A" indicate that the sample was
taken after carpet cleaning.
48
-------
Sample Number
Number Asbestos
EXPERIMENT 1
01B-002B 43
01B-003B 36
01B-005B 46
01B-008A 3
01B-009A 6
01B-011A 22
EXPERIMENT 2
02B-014B 28
02B-016B 52
02B-017B 20
02B-020A 23
02B-021A 32
02B-023A 22
EXPERIMENT 3
03B-026B 14
03B-028B 33
03B-029B 25
03B-032A 40
03B-034A 26
03B-035A 6
EXPERIMENT 4
04B-038B 26
04B-039B 46
04B-041B 46
04B-044A 2
04B-045A 4
04B-047A 3
of Concentration,
Str. s/ft2
- WET CLEAN
53,830,766
46,213,405
48,257,407
41,311,983
7,835,031
28,482,906
- DRY VACUUM
39,205,882
70,911,509
27,512,900
31,094,320
51,735,597
29,742,394
- DRY VACUUM
20,210,093
49,823,310
35,641,711
58,254,124
41,541,412
9,058,784
- WET CLEAN
38,897,868
72,789,649
66,992,243
29,921,437
65,827,160
49,370,370
49
-------
Asbestos
r Sample Number
Number Asbestos
t"
EXPERIMENT 5
(05B-050B 35
05B-051B 50
05B-053B 23
i
1 05B-056A 1
05B-057A 0
j D5B-059A 0
i
EXPERIMENT 6
06B-061B 34
I 06B-064B 17
06B-065B 28
[ 06B-067A 107
06B-070A 7
06B-D71A 32
EXPERIMENT 7
07B-073B 17
07B-075B 21
D7B-07BB 20
07B-079A 30
07B-081A 36
Q7B-084A 17
EXPERIMENT 8
08B-086B 29
08B-088B 2
08B-090B 12
08B-092A 1
08B-094A 7
08B-096A 1
of Concentration,
Str. s/ft2
- WET CLEAN
53,830,622
70,328,163
34,725,337
15,147,727
<15,147,727
<14,960,718
- DRY VACUUM
54,323,385
24,005,297
44,014,151
215,793,694
11,435,049
47,607,143
- DRY VACUUM
30,134,309
36,073,454
64,536,432
46,284,722
52,618,421
53,801,045
- WET CLEAN
90,046,587
6,329,535
34,284,979
16,456,790
25,042,941
3,226,822
50
-------
Asbestos
Sample Number
Number Asbestos
EXPERIMENT 9
09B-097B 140
09B-099B 129
09B-102B 104
09B-103A 5
09B-105A 3
09B-108A 3
EXPERIMENT 10
10B-110B 116
10B-111B 108
10B-113B 122
10B-116A 150
10B-117A 109
10B-119A 129
EXPERIMENT 11
11B-122B 118
11B-124B 127
11B-125B 147
11B-128A 39
11B-130A 22
11B-131A 1
EXPERIMENT 12
12B-134B 103
12B-135B 125
12B-137B 145
12B-140A 41
12B-141A 106
12B-143A 120
of Concentration,
Str. s/ft2
- WET CLEAN
577,413,366
497,560,764
393,215,339
75,738,636
45,443,182
51,269,231
- DRY VACUUM
303,907,233
439,450,549
416,989,744
640,865,385
251,048,794
484,113,176
- DRY VACUUM
232,134,002
384,752,273
594,511,529
464,169,643
33,630,734
16,021,635
- WET CLEAN
416,562,500
425,063,776
732,140,152
152,491,629
379,833,333
393,602,362
51
-------
Asbestos
Sample Number of
Number Asbestos Str.
EXPERIMENT 13 - WET
13B-146B 231
13B-147B 215
13B-150B 107
13B-152A 22
13B-153A 10
13B-156A 14
EXPERIMENT 14 - DRY
14B-157B 116
14B-159B 107
14B-161B 25
14B-163A 113
14B-165A 104
14B-167A 99
EXPERIMENT 15 - DRY
15B-171B 113
15B-172B 107
15B-173B 101
15B-177A 112
15B-178A 114
15B-179A 108
EXPERIMENT 16 - WET
16B-182B 115
16B-183B 43
16B-185B 107
16B-188A 13
16B-189A 47
16B-191A 8
Concentration,
s/ft2
CLEAN
859,160,156
923,308,634
342,862,981
366,575,000
166,625,000
212,068,182
VACUUM
315,825,163
441,308,787
416,562,500
362,088,942
451,276,042
196,379,464
VACUUM
412,908,443
530,621,280
707,291,831
347,766,131
538,414,116
453,401,361
CLEAN
675,903,880
555,416,667
543,480,415
194,489,338
662,272,727
131,654,321
52
-------
Asbestos
1
V
1
1
1
1
IHf
1
1
1
.
Sample
Number
17B-1B
17B-3B
17B-4B
17B-1A
17B-3A
17B-4A
18B-1B
18B-3B
18B-4B
18B-1A
18B-3A
18B-4A
19B-1B
19B-2B
19B-4B '
19B-1A
19B-2A
19B-4A
20B-1B
20B-3B
20B-4B
20B-1A
20B-3A
20B-4A
Number of
Asbestos Str.
EXPERIMENT 17 - WET
2
19
25
0
1
6
EXPERIMENT 18 - DRY
19
15
28
28
18
27
EXPERIMENT 19 - WET
18
83
41
1
1
6
EXPERIMENT 20 - DRY
15
8
34
26
11
33
Concentration,
s/ft5
CLEAN
31,738,095
56,889,039
79,345,238
<14,426,407
14,426,407
19,431,487
VACUUM
51,103,713
49,590,774
85,448,718
76,609,195
53,894,879
77,902,597
CLEAN
56,008,403
292,695,767
114,145,781
2,601,483
17,632,275
79,345,238
VACUUM
44,912,399
22,272,348
110,111,759
76,406,526
33,569,139
96,977,513
53
-------
Asbestos
Sample
Number
21B-1B
21B-2B
21B-4B
21B-1A
21B-2A
21B-4A
22B-1B
22B-2B
22B-3B
22B-1A
22B-2A
22B-3A
23B-1B
23B-2B
23B-4B
23B-1A
23B-2A
23B-4A
24B-1B
24B-2B
24B-3B
24B-1A
24B-2A
24B-3A
Number
Asbestos
EXPERIMENT 21
102
223
128
23
9
51
EXPERIMENT 22
104
124
129
115
53
107
EXPERIMENT 23
65
100
104
9
11
6
EXPERIMENT 24
137
86
202
75
110
69
of Concentration,
Str. s/ft2
- WET CLEAN
894,513,158
1,499,490,517
860,694,108
403,407,895
157,855,263
674,434,524
- DRY VACUUM
467,624,637
864,678,803
718,283,208
630,843,621
792,918,070
808,932,623
- WET CLEAN
218,304,359
435,364,818
419,486,807
28,482,906
164,567,901
89,764,310
- DRY VACUUM
1,142,809,762
251,509,434
1,649,914,216
1,306,862,745
321,698,113
1,082,082,353
54
-------
APPENDIX D
. AVERAGE ASBESTOS CONCENTRATIONS BEFORE
AND AFTER CARPET CLEANING FOR EACH EXPERIMENT
55
-------
Asbestos
Concentration, s/ft2
Experiment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Cleaning
Method
WET
DRY
DRY
WET
WET
DRY
DRY
WET
WET
DRY
DRY
WET
WET
DRY
DRY
WET
WET
DRY
WET
DRY
WET
DRY
WET
DRY
Contaminat i on
Level
LOW
LOW
LOW
LOW
LOW
LOW
LOW
LOW
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
LOW
LOW
LOW
LOW
HIGH 1 ,
HIGH
HIGH
HIGH 1 ,
Before
49,433,859
45,876,764
35,225,038
59,559,920
52,961,374
40,780,944
43,581,398
43,553,700
489,396,490
386,782,509
403,799,268
524,588,809
708,443,924
391,232,150
550,273,851
591,600,321
55,990,791
62,047,735
154,283,317
59,098,835
084,899,261
683,528,883
357,718,661
014,744,471
' After
j
25,876,640
37,524,104
36,284,773
48,372,989
5,049,242
91,611,962
50,901,396
14,908,851
57,483,683
458,675,785
171,274,004
308,642,441
248,422,727
336,581,483
446,527,203
329,472,129
11,285,965
69,468,890
33,192,999
68,984,393
411,899,227
744,231,438
94,271,706
903,547,737
56
-------
APPENDIX E
FIBER LENGTH DISTRIBUTIONS OF ASBESTOS
IN CARPET SAMPLES COLLECTED BEFORE
AND AFTER CARPET CLEANING
57
-------
TABLE E-l. FIBER LENGTH DISTRIBUTION OBSERVED IN THE CARPET SAMPLES
COLLECTED BEFORE CARPET CLEANING
Particle Number of Cumulative
size range, ym fibers counted fiber count
0.23-0.34
0.34-0.54
0.50-0.73
0.73-1.08
1.08-1.58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
50.00-73.40
73.40-107.70
107.70-158.10
158.10-232.10
232.10-340.60
18
78
165
404
875
1150
1149
877
439
171
42
3
1
0
1
0
0
0
0
18
96
261
665
1540
2690
3839
4716
5155
5326
5368
5371
5372
5372
5373
5373
5373
5373
5373
Percent
of total
0.3
1.5
3.1
7.5
16.3
21.4
21.4
16.3
8.2
3.2
0.8
0.1
0
0
0
0
0
0
0
Cumulative
! percent
1 0.3
1.8
i 4.9
12.4
; 28.7
: 50.1
71.4
87.8
i 95.9
99.1
' 99.9
'• 100
; 100
: 100
100
! 100
: 100
100
'. 100
58
-------
TABLE E-2. FIBER LENGTH DISTRIBUTION OBSERVED IN CARPET SAMPLES
COLLECTED AFTER DRY VACUUMING OF CARPET CONTAMINATED WITH THE LOW
uunucm rw-n iun uiorc.r\oiuiN
Particle
size range, ym
0.23-0.34
0.34-0.54
0.50-0.73
0.73-1.08
1.08-1.58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
50.00-73.40
73.40-107.70
107.70-158.10
158.10-232.10
232.10-340.60
Number of
fibers counted
2
7
21
35
79
84
97
86
74
25
7
1
1
0
0
0
0
0
0
Cumulative
fiber count
2
9
30
65
144
228
325
411
485
510
517
518
519
519
519
519
519
519
519
Percent
of total
0.4
1.3
4.0
6.7
15.2
16.2
18.7
16.6
14.3
4.8
1.3
0.2
0.2
0
0
0
0
0
0
Cumulative
, percent
0.4
1.7
5.8
: 12.5
27.7
43.9
62.6
79.2
93.4
98.3
: 99.6
; 99.8
: 100
100
100
: 100
100
100
100
59
-------
TABLE E-3. FIBER LENGTH DISTRIBUTION OBSERVED IN CARPET SAMPLES
COLLECTED AFTER WET CLEANING OF CARPET CONTAMINATED WITH THE LOW
CONCENTRATION DISPERSION
Particle
size range, ym
0.23-0.34
0.34-0.54
0.50-0.73
0.73-1.08
1. 08-1. 58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
50.00-73.40
73.40-107.70
107.70-158.10
158.10-232.10
232.10-340.60
Number of
fibers counted
0
3
2
14
15
11
11
5
3
0
1
0
0
0
0
0
0
0
0
Cumulative
fiber count
0
3
5
19
34
45
56
61
64
64
65
0
0
0
0
0
0
0
0
Percent
of total
0
4.6
3.1
21.5
23.1
16.9
16.9
7.7
4.6
0
1.5
0
0
0
0
0
0
0
0
Cumulative
; percent
; 0
-. 4.6
7.7
; 29.2
52.3
69.2
: 86.2
93.8
• 98.5
! 98.5
: 100
100
100
: 100.
100
"• 100
; 100
1 100
: 100
60
-------
TABLE E-4. FIBER LENGTH DISTRIBUTION OBSERVED IN CARPET SAMPLES
COLLECTED AFTER DRY VACUUMING OF CARPET CONTAMINATED WITH THE HIGH
CONCENTRATION DISPERSION ;
Particle
size range, ym
0.23-0.34
0.34-0.54
0.50-0.73
0.73-1.08
1.08-1.58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
50.00-73.40
73.40-107.70
107.70-158.10
158.10-232.10
232.10-340.60
Number of
fibers counted
4
23
60
123
262
389
346
266
108
36
7
0
0
0
0
0
0
0
0
Cumulative
fiber count
4
27
87
210
472
861
1207
1473
1581
1617
1624
1624
1624
1624
1624
1624
1624
1624
1624
Percent
of total
0.2
1.4
3.7
7.6
16.1
24.0
21.3
16.4
6.7
2.2
0.4
0
0
0
0
0
0
0
0
Cumulative
; percent
0.2
1.7
: 5.4
12.9
1 29.1
53.0
' 74.3
90.7
: 97.4
: 99.6
100
: 100
i 100
: 100
' 100
1 100
100
100
: 100
61
-------
TABLE E-5. FIBER LENGTH DISTRIBUTION OBSERVED IN CARPET SAMPLES
COLLECTED AFTER WET CLEANING OF CARPET CONTAMINATED WITH THE HIGH
CONCENTRATION DISPERSION
Particle
size range, ym
0.23-0.34
0.34-0.54
0.50-0.73
0.73-1.08
1.08-1.58
1.58-2.32
2.32-3.41
3.41-5.00
5.00-7.34
7.34-10.77
10.77-15.81
15.81-23.21
23.21-34.06
34.06-50.00
50.00-73.40
73.40-107.70
107.70-158.10
158.10-232.10
232.10-340.60
Number of
fibers counted
0
4
21
35
101
102
116
67
36
13
4
0
0
0
0
0
0
0
0
Cumulative
fiber count
0
4
25
60
161
263
379
446
482
495
499
499
499
499
499
499
499
499
499
Percent
of total
0
0.8
4.2
7.0
20.2
20.4
23.2
13.4
7.2
2.6
0.8
0
0
0
0
0
0
0
0
Cumulative
'. percent
0
0.8
: 5.0
; 12.0
32.3
52.7
I 76.0
' 89.4
; 96.6
i 99.2
: 100
1 100
1 100
. 100
: 100
: 100
100
100
; 100
62
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
3. RECIPIENT'S ACCESSION NO.
1. REPORT NO.
2.
4. TITLE AND SUBTITLE
Evaluation of Two Cleaning Methods for Removal of
Asbestos Fribers From Carpet
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
John R. Kominsky, Ronald W. Freyberg, Jean Chesson
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
PEI Associates, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-03-4006, WA 2-10
12.SPONSORING AGENCY NAME AND ADDRESS
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
1/88 - 9/89.
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer:
FTS: 684-7550
Thomas J. Powers Comm: 569-7550
16. ABSTRACT
This research study examined the effectiveness of dry vacuuming and wet
cleaning for the removal of asbestos fibers from carpet, and evaluated the
potential for fiber reentrainment during carpet cleaning activities. Routine
carpet cleaning operations using high-efficiency particulate absolute (HEPA)
filtered dry vacuum cleaners and HEPA-filtered hot-water extraction cleaners
were simulated on carpet artificially contaminated with asbestos fibers.
Overall, wet cleaning the carpet with a hot-water extraction cleaner reduced
the level of asbestos contamination by approximately 70 percent. ;There was
no significant evidence of either an increase or a decrease in asbestos
concentration after dry vacuuming. The level of asbestos contamination had
no significant effect on the difference between the asbestos concentrations
before and after cleaning. Airborne asbestos concentrations were'two to four
times greater during the carpet cleaning activities. The level of asbestos
contamination in the carpet and the type of cleaning method used had no
significant effect on the difference between the airborne asbestos concentra-
tion before and during cleaning. '
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
18. DISTRIBUTION STATEMENT
Release To Public
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
------- |