PB89-221246
Asbestos Fiber Reentra inment during Vacuuming and
Wet-Cleaning of Carpet at a Captive Research Site
•(U.S.) Environmental Protection Agency, Cincinnati, OH
31 Mar 89
J
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
ii
-------
EPA/.60Q/Dt89/055
ASBESTOS FIBER REENTRAINMENT DURING VACUUMING AND WET-CLEANING
OF CARPET AT A CAPTIVE RESEARCH SITE
by
John R. Kominsky
Ronald W. Freyberg
PEI Associates, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
William C. Cain
Roger C. Wilmoth
Thomas J. Powers
David L. Smith .
U. S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
2 6 West Martin Luther King Drive
Cincinnati, Ohio 45268
Paper Presented at the
National Asbestos Council Conference
March 29-31, 1989
Anaheim, California
-------
TECHNICAL REPORT DATA # ,
(fietat rtmd Inttrvctiofti on She nvtnt be fort eomptetrnf)
V REPORT NO. 2.
EPA/600/D-89/055
J. RECIPIENT'S ACCESSION NO,
PB89—22121*6
4. TITLE AND SUBTITLE
Asbestos Fiber Reentrainment During Vacuuming
and Wet-Cleaning of Carpet at a Captive Research Sit
ft. REPORT DATE
1. PERFORMING ORGANIZATION CODE
s
7. AUTHOR(S) 2
John Kominsky and Ronald Freyburg * ~
W. Cain,R. Wilmoth, T. Powers and D. Smith
1. PERFORMING ORGANIZATION REPORT NO.
8. PERFORMING organization name and address
1. PEI Associates, Inc. Cincinnati, OH 45246-0100
2. EPA, RREL, WHWTRD.TCB, Cincinnati, OH 45268
10, PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Risk Reduction Engineering Laboratory - Cincinnati, OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH
13. TVjPE OF REPORT AND PERIOD COVERED
Coraplete""'
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUWUMENTARY NOTES , .
NAC Conference - Sixth Annual Asbestos Abatement Conference and Exposition
Anaheim, California - March 29-31, 1989
1«. abstract
A statistical evaluation was made of airborne asbestos concentrations measured
before, during, and after removal of asbestos-containing fireproofing at three
university buildings. Of the three sites studied, all passed the AHERA Z-test when
the work area asbestos I eve! s wer.e compared to perimeter levels (outside the
abatement area but inside the building). Two sites also passed the Z-test when
work area and outdoor air levels were compared. At one site, contamination of the
perimeter area occurred at some point during the abatement project. Had this area
been used in the Z-test clearance comparison, a contaminated site would have been
falsely released.
17. KEY WORDS AND DOCUMENT ANALYSIS
1. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
C. COS ATI Fktd/Gcoup
IS. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
IS. SECURITY CLASS (Thii Rrpor,t
UNCLASSIFIED
2*. NO. Of "AGES
2.2.
30. SECURITY CLASS (Ttujfmff!
UNCLASSIFIED
_ _i
32. PRICE
&3
CPA t~m 1226-t
-------
Asbestos Fiber Reentrainment During Vacuuming and
Wet-Cleaning of Carpet at a Captive Research Site
ABSTRACT
A study was conducted to compare the effectiveness of
alternative carpet cleaning techniques and to evaluate the
potential for asbestos fiber reentrainment during cleaning of
carpet contaminated with asbestos. The equipment was evaluated
at two carpet contamination levels. Airborne asbestos
concentrations were determined before and during carpet cleaning.
Overall, airborne asbestos concentrations were two to four times
greater during the carpet cleaning activity. The level of
asbestos contamination and the type of cleaning method had no
statistically significant effect on the relative increase of
airborne asbestos concentrations during carpet cleaning.
INTRODUCTION
Buildings that contain friable asbestos-containing materials
(ACM) may present unique exposure problems for custodial workers.
A major concern is the extent which carpet and furnishings may be
reservoirs of asbestos fibers, and their subsequent behavior when
normal custodial cleaning operations are performed.
The U. S. Environmental Protection Agency (EPA) performed a
series of controlled experiments at a captive research site to 1)
evaluate two cleaning methods for removal of asbestos fibers from
carpet, and 2) evaluate the potential for asbestos-fiber
reentrainment during these cleaning activities. Analysis of the
carpet samples to evaluate the removal effectiveness of the
cleaning methods is ongoing and will be reported in the future.
This paper presents the airborne asbestos concentrations
resulting from the dry-vacuuming and wet-cleaning of asbestos-
contaminated carpet.
EXPERIMENTAL DESIGN
This study was conducted at a captive research site
consisting of an unoccupied building scheduled for demolition.
Two rooms, each with approximately 600 square feet of floor
space, were constructed in a larger bay of the building.
A layout drawing of the test facility is shown in Figure 1.
The rooms were constructed of 2 x 4-inch lumber with studs spaced
on 24-inch centers, and 3/4-inch plywood floors. The ceiling,
floor, and walls were double-covered with 6-mil polyethylene
sheeting. (The interior layer of polyethylene sheeting was
encapsulated and replaced after each experiment). Separate
decontamination facilities for workers and waste-materials were
2
-------
HEPA-FILTERED
NEGATIVE AIR
UNIT
Figure t. Layout of test facility, bldg. 1422-A.
-------
connected to the experimental areas. Room size was determined
based on the minimum amount of time required to vacuum or wet-
clean the room and attain an adequate sample air volume to
achieve a specified analytical sensitivity. A 52-inch ceiling-
mounted, axial flow, propeller fan was installed in each room to
facilitate air movement and minimize temperature stratification.
Two carpet cleaning techniques (HEPA-filtered dry vacuuming
and HEPA-filtered hot water extraction) were used on carpet
artificially contaminated with 100 million and 1 billion asbestos
structures per square foot (a.s./ft2). Each treatment
combination was replicated four times to yield a total of 16
experiments. Each type of cleaning equipment was tested in each
room the same number of times for each contamination level. A
single experiment included contaminating a new piece of carpet in
a previously cleaned room, collecting work area air samples,
vacuuming or wet-cleaning the carpet for a specified period of
time while simultaneously collecting work area air samples,
removing the carpet, and cleaning the room.
Three work area air samples were collected before cleaning
and three air samples were collected during carpet cleaning for
each experiment. The air samplers were positioned in a
triangular pattern (Figure 1) at a height of approximately five :
feet above the floor.
Carpet Selection
A survey of fourteen General Service Administration (GSA)
field offices in eleven different states distributed across the
United States was conducted to determine the specific type and
manufacturer of carpet to be used in the study. Eight offices
indicated a preference for the same manufacturer and variety of
carpet. The carpet selected was first-grade, 100-percent Nylon,
with 0.25-inch cut pile, 28 ounces of yarn per square foot, and
dual reenforced vinyl backing. The carpet was manufactured in
roll sizes of 54 by 90 feet.
Cleaning Equipment
The same GSA field office survey identified a commonly used
HEPA-filtered vacuum cleaner. Four different HEPA-filtered
vacuum cleaners of the same model were used in this study. These
units were equipped with a motor driven carpet nozzle with a
rotating brush. The hot water extraction unit used was selected
based on the results of a survey of six trade associations for
commercial cleaners. Four different cleaners of the same model
were used. These units were equipped with an extractor tool
which uses a motor-driven cylindrical brush to agitate and scrub
the carpet during the extraction process.
4
-------
Sampling Methodology
Air samples were collected on open-face 25-run diameter,
0.4 5-um pore-size mixed cellulose ester membrane filters with a
5-um 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 4 5-degree angle
toward the floor. The filter assembly was attached to an
electric-powered vacuum pump operating at a flow rate of 10
liters per minute. 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.
Analytical Methodology
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 Emergency Response Act
(AHERA) final rule (52 CFR 41821). Because there are no OSHA
Permissible Exposure Limits or NIOSH Recommended Exposure Limits
for airborne asbestos measured by TIM, a subset of filters were
selected for additional analysis by phase contrast microscopy
(PCM) according to NIOSH Method 7400.
Carpet Contamination
Carpet contamination levels for this study were selected based on
field data reported by Wilmoth et. al.1 Wilmoth reported that
carpet samples from an occupied building were collected using two
different sampling techniques. Microvacuuming of the
contaminated carpet revealed asbestos levels ranging from
approximately 8000 to 2 billion a.s./ft2. Bulk sample
sonification of the samples showed asbestos levels ranging from
30 million to 4 billion a.s./ft2. These results indicate that
the experimental contamination levels of 100 million and 1
billion a.s./ft represent realistic carpet contamination levels.
The carpet was contaminated with a spray-applied dispersion
of asbestos in distilled water. Sealed ampules of asbestos fiber
dispersions were prepared such that the contents of one ampule,
dispersed in approximately 6 liters of freshly-distilled water,
would provide the required concentration of suspension to
artificially contaminate one 600 ft2 sample of carpet.
The original suspension was prepared by dispersing a known
weight of chrysotile in freshiy-distilled water. A weight of
4 09.5 mg of purified Calidria chrysotile was placed in an agate
mortar and, using a pestle, was lightly ground with a small
volume of water, gradually adding more freshly-distilled water
5
-------
until a creamy liquid was obtained. This liquid was made up to
400 mL in a polypropylene disposable beaker then placed in an
ultrasonic bath for approximately 30 minutes. The suspension was
then made up to 1500 mL with distilled water in a one-gallon
polyethylene bottle. The bottle was then placed in an ultrasonic
bath for 3 0 minutes, during which time the bottle was removed and
shaken vigorously. For the lower concentration, a volume of 150
mL of this suspension was made up to 1500 mL with freshly-
distilled water in another one-gallon polyethylene bottle. The
two suspensions had concentrations of 273 mg/liter and 27.3
mg/liter, respectively. A volume of 50 mL of suspension was used
to prepare each ampule.
The asbestos dispersion was applied to the carpet using a
meticulously cleaned hand-pumped garden sprayer. A fixed number
of pumps were used for each batch to provide consistent spray
pressure. The desired controlled spray was experimentally
determined by trial and error prior to beginning the tests with
asbestos. 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 following a grid work of
string placed over the carpet before starting each experiment.
The tank was periodically shaken and agitated to aid in keeping
the asbestos fibers suspended in the tank. 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 surfaces to simulate the effects of normal foot traffic in
working the asbestos into the carpet.
Prior to the start of the field experiments, the spray-
applied dispersion was evaluated in the laboratory to examine the
effectiveness of the technique to apply the asbestos dispersion
and define the degree of fiber loss, if any. An asbestos
dispersion was prepared in a hand-pumped sprayer identical to
that used in the field experiments. Three one-liter samples were
then collected by spraying the asbestos dispersion into separate
glass containers. These samples were then analyzed to determine
the asbestos concentration in terms of fibers per liter of water.
These results are presented in Table 1. The original dispersion
concentration was approximately (2 to 4) x 1013 asbestos fibers
per liter of water. These results indicate no significant loss
of fibers during the transfer of the liquid-dispersion through
the sprayer's hose and nozzle.
6
-------
TABLE 1. RESULTS FROM PRELIMINARY STUDY OF SPRAY-APPLIED
ASBESTOS DISPERSION
Volume in Sprayer at Sample
Time of Sample Collection Volume
(Liters) (Liters)
Mean Asbestos
Concentration
(Fibers/Liter)
6
4
2
1
1
1
2.38 X 10
13
2.22 X 10
2.20 X 10
13
13
Cleaning Technique
The carpet was vacuumed or wet-cleaned for a period of
approximately 65 minutes to allow the collection of air samples
of sufficient volume to obtain an analytical sensitivity of 0.005
asbestos structures per cubic centimeter. The carpet was cleaned
in two directions, the second at a 90-degree angle to the first.
Quality Assurance
Specific quality assurance procedures used to ensure the
accuracy and precision of the collection and tem analysis of air
samples included the use of filter lot blanks, laboratory blanks,
field blanks, replicate and duplicate sample analyses.
Filter lot blanks are unused filters selected at random and
analyzed to determine the background contamination level of
asbestos. One hundred (100) lot blanks were submitted for TEM
analysis. No asbestos structures were detected in 1000 grid
openings analyzed. The lot of filters was subsequently
considered acceptable for use.
During the setup of the air sampling pumps, pre-loaded
filter cassettes were selected as field blanks. These filters
were labeled and handled in a manner similar to that for the
actual sample filters, but they were never attached to the pump.
One field blank was collected for each of the 16 experiments.
Two of the 16 filters each contained one asbestos structure.
Additionally, prior to each of the sixteen experiments, one
sample cassette was.selected from the filter inventory to be used
as a laboratory blank. These samples were sealed and submitted
for use by the analytical .laboratory to ensure that there was no
blank interferences during the analytical procedures. Two of the
7
-------
16 sealed blanks each contained two asbestos structures.
Analysis of the field and laboratory blanks demonstrated filter
contamination was significantly below the guideline of 3 asbestos
structures average per 10 grid openings.
Duplicate sample analysis provides a means of quantifying
intralaboratory precision and refers to the analysis of the same
grid preparation by a second microscopist. Five samples were
randomly selected for duplicate analysis. There was no evidence
of inconsistency among the two sets of analyses. A paired t-
test2 did not detect any statistically significant tendency for
one analyst to give higher or lower structure counts (p=0.6195).
Replicate sample analysis provides a means of quantifying
analytical variability introduced by the filter preparation
procedure and refers to the analysis of a second grid preparation
from the original filter, but not necessarily by the same
analyst. Five samples were randomly selected for replicate
analysis. While a paired t-test did detect a statistically
significant tendency for the replicate analysis to yield lower
asbestos concentrations (p=0.0259), the effect due to the filter
preparation procedure is confounded by the effect of a second
analyst. Hence, an overall statement regarding analytical
reproducibility is not appropriate.
Statistical Analysis
Methods—
Airborne asbestos concentrations were determined before and
during carpet cleaning to study the effect of 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 concentration before and during cleaning
was then determined by averaging the three respective work area
samples. As a measure of relative change in airborne asbestos
concentration, the ratio of the concentration during cleaning to
the concentration prior to cleaning was computed. The natural
log of this ratio was then analyzed using a two factor analysis
of variance (ANOVA) with cleaning method and contamination level
as the main effects. The two-factor interaction term was also
included in the model.
Summary statistics (arithmetic mean and standard deviation)
were calculated according to cleaning method and contamination
level.
8
-------
RESULTS AND DISCUSSION
Figure 2 presents the average airborne asbestos
concentrations measured before and during cleaning for each
cleaning method and carpet contamination loading. Table 2
presents the summary statistics.
Air sampling results from two of the sixteen experiments
showed that for both wet cleaning and HEPA vacuuming, the average
airborne asbestos concentrations decreased during carpet cleaning
The explanation for this anomaly is that the HEPA filtration
system used to ventilate the test rooms was operating during the
carpet cleaning phase of these two experiments. Therefore, these
results were omitted from the statistical analysis of the data.
Results from the two factor ANOVA indicated no statistically
significant difference between cleaning methods with respect to
fiber reentrainment during carpet cleaning (p=0.5847). That is,
the mean relative increase in airborne asbestos concentrations
during carpet cleaning with a HEPA vacuum was not significantly
different from that found during wet cleaning. When averaged
across both contamination loadings, airborne asbestos
concentrations increased approximately 2.6 times during HEPA
vacuuming and approximately 3.2 times during wet cleaning.
9
-------
Average
Airborne
Asbestos
Concentration
(s/cm 3
/
0.3
Low Contamination
Hot Water
Extraction
HE PA
High Contamination
/ 0.2248/
fptw
wM%- &
W'-iM.f;
Hot Water
Extraction
HE PA
Figure 2. Average airborne asbestos concentrations before and during
carpet cleaning.
-------
TABLE 2. SUMMARY STATISTICS FOR AIRBORNE ASBESTOS CONCENTRATIONS
BEFORE AND DURING CARPET CLEANING.
Approximate
Contamination
Loading
(a.s./ft*)
Cleaner
Number of
Samples
Airborne Asbestos
Concentration (s/cra3)
Average
Standard
Deviation
100 million
Before Cleaning
Extraction 9 0.0673 0.0947
HEPA-Vacuum 9 0.0571 0,0300
During Cleaning
Extraction 9 0.1639 0.1020
HEPA-Vacuum 9 0.2531 0.1729
1 Billion
Before Cleaning
Extraction 16 0.0761 0.0425
HEPA-Vacuum 16 0.1424 0.1340
During Cleaning
Extraction 16 0.1577 0.0602
HEPA-Vacuum 16 0.2248 0.1114
Note: The data points used in the calculation of each summary statistic
are the averages of the three work area samples before and during
cleaning.
11
-------
Similarly, no statistically significant difference between
carpet contamination loadings with respect to fiber reentrainment
was evident (p=0.0857). That is, the mean relative increase in
airborne asbestos concentrations during carpet cleaning when the
carpet contamination level was 100 million a.s./ft2 was not
significantly different from that found when the carpet
contamination loading was 1 billion a.s./ft2. When averaged
across both cleaning methods, airborne asbestos concentrations
increased approximately 2.5 times at the high contamination level
and approximately 5 times at the low contamination level.
Given that no differences due to cleaning method or
contamination level were detectable, the question of whether
there was an overall increase in mean airborne asbestos
concentration during carpet cleaning was tested. ANOVA results
indicate that, overall, the mean airborne asbestos concentration
during carpet cleaning was significantly higher during carpet
cleaning than that just prior to cleaning (p=0.oooi).
Specifically, the mean airborne asbestos concentration was
between 2 and 4 times greater during carpet cleaning.
Airborne Asbestos Fiber Distribution
TEM analysis of the 47 work area samples before and during
cleaning yielded a total" of 2,839 structures. Of these, 2757
(97.1%) were chrysotile, 8 (0.3%) were amphibole, and 74 (2.6%)
were ambiguous. The structure morphology distribution is
summarized in Table 3. *
TABLE 3. STRUCTURE MORPHOLOGY DISTRIBUTION OF AIR SAMPLES
COLLECTED BEFORE AND DURING CARPET CLEANING.
Number
Number
Number
Number
Structure
of
of
of
of
Type
Bundles
Clusters
Fibers
Matrices
Chrysotile
30
7
2661
59
Amphibole
0
2
5
1
Ambiguous
2
0
70
2
12
-------
These data indicate that the original chrysotile fibers used
to prepare the asbestos dispersion remained intact as fibers.
That is, there appeared to be no significant tendency for the
fibers to clump together as a result of the dispersion
preparation, carpet contamination, or cleaning technique.
The presence of amphibole asbestos fibers in the air is
probably due to existing conditions prior to experimentation.
Pre-study air monitoring identified three amphibole asbestos
fibers in seven air samples collected.
The structure length distribution of asbestos particles
found in the air before and during carpet cleaning is summarized
in Table 4 and illustrated in Figure 3. Eighty-four percent
(84%) of the chrysotile structures identified were one micron or
less in length. Only nine particles were identified with.lengths
greater than five microns. Compared to the fiber length
distribution of chrysotile used to contaminate the carpet (see
Figure 4), these data certainly suggest that the larger asbestos
particles either remained in the carpet or were prevented from
escaping into the air by the carpet cleaning activity.
Samples Analyzed by PCM
Twelve samples were selected to be analyzed by phase
contrast microscopy based on their respective high asbestos
concentrations determined by TEM. Results from both TEM and PCM
analyses are compared Table 5. Airborne fiber concentrations
determined by PCM were significantly lower than the corresponding
asbestos concentrations determined by TEM. This difference is
presumably due to the limitation of PCM to detect small fibers.
It should be noted that the majority of asbestos fibers applied
(Figure 4) did not meet the dimensional criteria (length >5 um)
of NIOSH Method 7400 and hence were not counted.
13
-------
TABLE 4. FIBER LENGTH DISTRIBUTION OF AIRBORNE ASBESTOS IN THE
WORK AREA BEFORE AND DURING CARPET CLEANING
Number of Structures{%)
During Cleaning
Structure
Length Before
(Micrometers) Cleaning Extraction HEPA-Vacuum
0.5 to
1.0
866(83.7)
710(82.0)
731(85.3)
1.0 to
2.0
138(13.3)
132(15,2)
110(12.8)
2.0 to
3.0
22
( 2.1)
11
( 1-3)
9
( 1.1)
3.0 to
4.0
5
( 0.5)
7
( 0.8)
3
( 0.4)
4.0 to
5.0
2
( 0.2)
1
( 0.1)
1
( 0.1)
5.0 to
6.0
1
( 0.1)
1
( 0.1)
2
( 0.2)
6,0 to
7.0
0
( 0.0)
2
( 0.2)
1
( 0.1)
7.0 to
8.0
0
( 0.0)
1
( o.i)
0
( 0.0)
8.0 to
9.0
0
( 0.0)
0
( 0.0)
0
( 0.0)
9.0 to
10.0
0
( 0.0)
0
( 0.0)
0
( 0.0)
> 10
1
( 0.1)
1
( 0.1)
0
( 0.0)
14
-------
Frequency of Fibers
1000
800
600 -
400 -
200
,5-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 >10
Fiber Length (micrometers)
Figure 3. Fiber length distribution for airborne asbestos before and
during carpet cleaning.
-------
1.0 10.0
Fiber Length,, microns
100.0
Figure 4. Distribution of chrysotile fiber lengths in the aqueous asbestos dispersion.
-------
TABLE 5. COMPARISON OF TEM AND PCM ANALYSES OF
SELECTED AIR SAMPLES
PCM Fiber TEM Asbestos
Sample Concentration Concentration
Number (f/cm3) (s/cro3)
03-A457D
0.0035
0.5507
03-A458D
0.0023
0.3658
03-A459D
0.0081
0.3464
10-A496B
0.0026
0.3656
10-A497B
0,0078
0.2909
10-A498B
0.0068
0.3375
10-A499D
0.0116
0.3871
10-A500D
0.0109
0.4891
10-A501D
0
0.0070
14-A523D
0.0061
0.3177
14-A524D
0.0138
0.3779
14-A525D
0.0138
0.3368
17
-------
CONCLUSIONS
Dry-vacuuming and wet-cleaning of carpet artificially
contaminated with asbestos fibers resulted in a statistically
significant increase in airborne asbestos concentrations. The
increase did not vary significantly with type of cleaning method
(wet or dry) or with the two levels of asbestos contamination
applied to the carpet.
Although this study suggests that performing normal
custodial cleaning of asbestos contaminated carpet may result in
elevated airborne concentrations, these data should not be
directly extrapolated to real-world situations. Further research
is required to determine the actual exposure risk to custodial
workers performing these activities in buildings containing
friable asbestos containing materials.
18
-------
REFERENCES
Wilmoth, R. C., T. J. Powers, and J. R. Millette,
"Observations on Studies Useful tc Asbestos O & M
Activities", Presented at the National Asbestos Council
Conference held in Atlanta, Georgia in February, 1988.
Neter, J., Wasserman, W., Kutner, M. H. Applied Linear
Statistical Models. Second Edition. 1985. Richard D.
Irwin, Inc.
19
------- |