United States
Environmental Protection
Agency
Atmospheric Research and
Exposure Assessment Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/S3-90/030 June 1990
SEPA Project Summary
Field Evaluation of a High
Volume Surface Sampler for
Pesticides in Floor Dust
W. T. Budd, J. W. Roberts, and M. G. Ruby
House dust and the pollutants
carried with it are potentially
important contributors to total
exposure through the pathways of
inhalation, ingestion and skin pen-
etration, especially for small children.
Pesticides may be one of the more
important contaminants of house
dust. The full report describes a pilot
study conducted as a part of the Non-
Occupational Pesticide Exposure
Study (NOPES), which provides pre-
liminary information on the pesticide
content of floor dust.
A high volume surface sampler
(HVS2) for the collection of house
dust and the semivolatile organics in
house dust has been developed and
tested in the laboratory. This study
also served as a field test and initial
validation study of the HVS2. The
HVS2 is designed to collect more
than 2 g of floor dust from a rug in an
average clean residence in a few
minutes. Such a large sample could
be used in bioassays or analyzed for
a variety of contaminants.
This study was conducted in nine
houses in the Jacksonville, Florida
Phase III segment of the NOPES
study. Both the NOPES questionnaire
and a supplemental questionnaire
were administered in each household
to develop information on pesticide
usage and other variables that might
be related to the floor dust samples.
All samples were collected from
carpeted surfaces.
The samples were collected and
processed using previously pub-
lished procedures. Both the HVS2
and the procedures were found to be
generally satisfactory. An average of
3.2 m2 (34 ft2) was necessary to
collect a 2 g sample. The total time
for collecting a single sample,
including sample processing, clean-
up of the HVS2, and travel time, was
approximately 4 hours.
The samples were analyzed for 33
pesticides by GC/ECD andGC/MS
following the NOPES standard
procedures. High concentrations of
interfering compounds in some
samples required substantial dilution
before they could be analyzed. As a
result, other analytes were diluted
below their detection limit.
On average, 7.5 target pesticides
were observed in the indoor air
samples and 11.8 in the floor dust.
The number observed in the floor
dust ranged from 2 to 23. Thirteen of
the pesticides were observed only in
the floor dust. The most consistently
observed pesticides were chlOrpy-
rifos, with a median concentration of
about 5 ppm in the dust, and
chlordane, with a median concentra-
tion of approximately 6 ppm. The
-------
median surface loading of
chtorpyrifos was approximately 6
yg/m2 and of chlordane, about 13
ug/m2. No significant correlation
between surface loading and dust
concentration was seen for these or
most other pesticides. Several
pesticides were observed in floor
dust which have not been in
widespread use for many years.
Exploratory statistical analyses
suggest that a relationship may exist
between the measured concentra-
tions of pesticides in the dust and in
the air for some pesticides. A
relationship was observed between
the number of pesticides detected
and the age of the house.
The HVS2's PUF plug adsorber was
necessary for the accurate
measurement of five pesticides. The
source of the pesticides on the PUF
plug (room air or blow off from the
glass fiber filter) was not determined
nor Is it clear why these five
pesticides are more likely to be found
in the PUF plug. A supplemental
experiment indicates a relationship
between blowoff and vapor pressure.
This Project Summary was
developed by EPA's Atmospheric
Research and Exposure Assessment
Laboratory, Research Triangle Park,
NC, to announce key findings of the
research project that is full-y
documented In a separate report of
the same title (see Project Report
ordering Information at back).
Introduction
The Non-Occupational Pesticide
Exposure Study (NOPES) was designed
to gather information on total human
exposure to selected pesticides among
persons not occupationally exposed,
using the Total Exposure Assessment
Methodology. During Phase I and Phase
II of NOPES, measurements were taken
to permit estimates of human exposure
through air, drinking water, and dermal
contact. The full report describes a pilot
study during the NOPES Phase III
activities which provides preliminary
information on possible human exposure
to pesticides through the additional
medium of household dust.
The importance of routes of exposure
other than inhalation has been
increasingly recognized in recent studies.
The Health and Safety Plans for the Hyde
Park disposal site in Niagara Falls, New
York developed under a Superfund
consent decree require the monitoring of
homes for contamination of the house
dust. However, i prior to this study, no
validated procedure existed to make such
measurements. I
The ingestion of pollutants deposited
on soil and dust (particularly house dust)
is expected to b|e especially important for
very young children, because of their
lower body weight and frequent hand-to-
mouth activity.! The toddler has only
about one-fifth [the body weight of an
adult and ingests an estimated 2.5 times
as much dust> which increases the
potential health frisk to the child by at
least 12 times. ^Further, the risk to such
children will be' increased by the early
stage of development of their organs,
nervous system, and immune system.
Estimates of '• the risks posed by
pesticides and other semivolatile
chemicals should therefore include an
evaluation of the dust-mediated pathway.
Validation of'a method for sampling
surface dust ife necessary if reliable
measurements • are to be made. The
study reported here also served as a field
test and initial validation study of a new
surface dust sahpling methodology, the
High Volume Surface Sampler (HVS2),
constructed by [Cascade Stack Sampling
Systems (CS3). The HVS2 was designed
and developed (Roberts and Ruby, 1988)
under EPA spbnsorship as a way to
collect rembvablesurface dust
accumulations on indoor surfaces (e.g.,
rugs and floors); and outdoor fugitive dust
surfaces (e.gj., bare ground at a
contaminated earth site).
The HVS2 has been extensively tested
in the laboratory. It is designed to collect
several grams of dust in a few minutes
from a rug or bare floor in an average
house. Such a large sample could be
used in bioassays or analyzed for a
variety of contaminants by separate
techniques. Because many of the
environmental pollutants of most concern
in the dust matrix are semivolatile
organics, the HVS2 has been, and is
here, carefully studied for its ability to
capture these organic compounds.
This study involved nine houses in the
Phase III Jacksonville NOPES set. The
HVS2 was useci to collect a sample of
dust in each house. Analysis included the
dust loading on the surface and the
concentration o'f pesticides in the dust.
The primary purposes of this study were
to field test th<3 HVS2 to validate the
methodology, to obtain preliminary data
on the amount and characteristics of dust
in residences, and to obtain preliminary
data on the species and concentration of
pesticides in house dust.
The full repoijt provides information on
the practical aspects of field operations
with the HVS2, the ability of the HVS2 to
retain the collected pesticide material, the
variation in dust loading observed in the
nine houses, the loading and
concentrations of pesticides in house
dust in the nine houses, and a
comparison of the concentrations in the
air and in the dust.
Data Collection Methods
The sampling activities for this study
were integrated with the NOPES Phase III
sampling activities. The houses for floor
dust sampling were obtained
opportunistically from volunteers during
one week of routine sampling activity.
The standard NOPES samples were
collected in each household, including
indoor and outdoor air, in addition to the
floor dust sample.
The selection of the floor area within
the house to be sampled was determined
primarily by convenience for sampling
and the probability that a sufficiently
large dust sample could be collected in a
short period of time. No attempt was
made to presurvey the entire house in
order to select a representative area or
an area where children would be likely to
play. All the samples in this study are
from carpeted surfaces'.
The NOPES Study Questionnaire was
administered in a personal interview. The
Study Questionnaire requested
demographic data for the individual
respondent, any occupational exposure to
pesticides, the potential for pesticide use
in the home, and an inventory of all
pesticides currently in the house. In the
floor dust sample houses an additional
questionnaire was administered. This
questionnaire included queries on the
type and frequency of floor cleaning, the
number of cigarettes smoked in the
household, and the frequency of fireplace
or woodstove use, if one were present.
Floor dust sampling was carried out
using the HVS2 and the sampling
procedures described by Roberts and
Ruby (1988). In brief, the HVS2 is a high-
powered vacuum cleaner equipped with a
nozzle that can be adjusted to a specific
static pressure within the nozzle, a
cyclone to separate the larger particles
from the air stream immediately after
their removal from the surface, a high
efficiency quartz fiber filter for particles,
and (optionally) a polyurethane foam
(PUF) plug adsorber for semivolatile
organic compounds.
The HVS2 has been shown to collect a
consistent amount of the material on
either a plush or level loop carpet, the
two most commonly found residential
carpets. With a standard test dust it will
-------
collect approximately 30% of the material
in the test dust, which was less than 150
jim.
The recommended test procedures for
the HVS2 (Roberts and Ruby, 1988)
include calibration of the required
measuring instruments, pretest
preparation of the HVS2 and supplies, the
procedures for sampling, and a
procedure for clean up in the field
between samples. The sampling
procedure calls for laying out eight 46 by
137 cm (18 by 54 in.) rectangles, setting
the HVS2 flowrate and nozzle pressure
drop, sampling four of these rectangles,
and then continuing the sampling of the
additional rectangles until 2 g of material
has been collected.
Field Operations
In addition to obtaining preliminary data
on the amount and characteristics of dust
in homes and preliminary data on the
species and concentration of pesticides
in house dust, this field study provided
an opportunity to validate the High
Volume Surface Sampler and to field test
the sampling methodology. The latter
included an evaluation of the practicality
of the sampling, cleanup, and analysis
procedures, the approximate times
required to perform each of the tasks, a
determination of the amount of sample
that could be realistically collected in a
normal range of houses, and an
evaluation of the acceptability of the
HVS2 by the residents of the sampled
house.
Although the HVS2 weighs about 25 kg
(55 pounds), it could be carried into the
house by one person. After it was carried
into the house, the floors were examined
to locate areas where the 46 by 137 cm
(18 by 54 in.) sample sections could be
laid out. A folding frame was used as a
template for marking the selected
sampling areas with masking tape.
It was necessary to use from 1 to 13
(average of 5) sampling areas to collect
the required 2 g of sample. The sample
areas contain 0.63 m2 (6.8 ft2). Because
the amount of dust on the floor varied
widely, it was difficult to estimate prior to
sampling how many sample areas would
be required. The sampling continued
over additional sample sections until
about 2 g of material were collected. On
average, this required less than 7
minutes.
The HVS2 was cleaned between runs
using brushes and a mixed solvent (50%
acetone/50% hexane) rinse on the entire
sampling train between the nozzle and
the PDF plug holder.
The cyclone catch was weighed and
then seived to less than 150 pm, using a
shaker and sieves following ASTM
standard method 422-63.
Results
Seven of the nine sample sites were
single-family detached houses and the
remaining two were mobile homes. In all
but one house, most of the floors (other
than kitchen and bathroom) were
carpeted.
Dust Loading
The floor dust loadings at each site
were determined by dividing the total
sample collected in the cyclone and the
fraction of that less than 150 pm by the
total area sampled. The results are
presented in Table 1. The two houses
which have the greatest loading are also
the two houses which were not using a
vacuum. The ratio of the fine fraction to
the total dust load is about 0.5, if the two
high loading houses are not included.
Table 1. Total and Fine Fraction Dust
Loading by Household
Household
1 227-21 6
0753-039
1064-014
0490-026A
0490-026B
1440-01 6A
1 440-0 16B
1647-001
1064-011
0966-021
0782-038
Total Dust
load (g/m2)
10.8
4.2
0.3
2.2
0.8
1.4
4.3
0.8
6.6
33.7
812.7
Dust <150
Itm load
(g/m2)
6.6
3.0
0.1
1.2
0.3
1.0
1.1
0.3
4.7
23.3
168.9
Pesticide Loadings and
Concentrations in Air and Dust
The loading of the NOPES target
pesticides was calculated by dividing the
total mass of the pesticide by the area
sampled. The concentration of the
pesticide in the collected dust was
calculated by dividing the total mass of
pesticide by the mass of the fine fraction
of the cyclone catch plus the mass on the
quartz fiber filter. These results are
presented in Table 2 for each household.
The data are reported as both a
surface loading (ng/m2) and as a
concentration in the dust (ppb) as data
have been reported by others using both
measures. Previous work suggests
surface loading may be better related to
possible health effects. Davies et al.
(1987) found a better correlation between
lead surface loading and hand lead than
between lead concentration in dust and
hand lead.
The correlation between the
concentration (ppb) and pesticide loading
(ng/m2) measures was significant (at the
99% level) and robust for only two
pesticides, chlorothalonil and g-BHC. For
the remainder that could be statistically
tested, the coefficients were also
generally near zero.
The total pesticide loading on the floor
is understated by these data. First, the
large particle size fraction (>150 iim
diameter) was separated and was not
submitted for analysis. Some pesticide
materials may have been present in this
segment of the floor dust. Second, the
HVS2 does not collect 100% of the dust
in a carpet. It is not known what fraction
of the dust in the carpet is "available" for
human uptake or how this relates to the
fraction of dust which is collected by the
HVS2. The HVS2 does collect a fraction
which is approximately constant with
loading and for different carpet materials.
An average of 7.5 target pesticides
were identified in the indoor air at the
nine sites, while an average of 11.8
pesticides were identified in the dust. The
number observed in the floor dust ranged
from 2 to 23. Thirteen pesticides were
found only in the dust in these nine
houses. Four of these (PCP, ODD,
atrazine, and carbaryl) were not observed
in indoor air in any of the samples taken
during the Phase III Jacksonville segment
of NOPES. In addition, six others
(heptachlor epoxide, captan, methoxy-
chlor, cis- and trans-permethrin, and
DDT) were observed in indoor air of no
more than four sites in the entire
Jacksonville Phase III study. Conversely,
of the six pesticides which were observed
in samples from at least eight sites
during this pilot study (heptachlor,
chlorpyrifos, aldrin, dieldrin, chlordane,
and ortho-phenylphenol), all but one were
also observed in the air at most of those
locations. The exception is aldrin, which
was observed in the air at only two sites.
A summary of the median value (of the
sites where that pesticide was observed)
and maximum value for floor dust
concentration in ppm and loadings in
ng/m2 of the most commonly observed
pesticides is provided in Table 3.
The relatively fewer number but higher
loadings of pesticides observed in four of
the households (1064-011, 0966-021,
0782-038, and to a lesser extent 1647-
-------
Table 2.
Concentration of Pesticides in Surface Dust
1227-216 0753-039
Household
DichlorvQs
o-Phenylphenol
Propoxur
Bondiocarb
a-BHC
HCB
Atrazine
PCP
g-BHC
Diazinon
Chlorothalonil
Carbaryl
Heptachlor
Malathton
Chlorpyrifos
Aldrin
Dacthal
Heptachlor epoxlde
Oxychlordane
Cap/an
Chtofdane
Folpet
2.4.D
DDE
Die/drift
DDT
ODD
Methoxychlor
c-Permothrin
t-permethrin
nglm2 ppb nglm2
399" 60"
8730 1315 5778*
13228 1992 2565"
21
2646" 398" 22
17
254 38 249
76190 11474 17851
3810 574 10
1812 273 19550
600
187 28 1440
4772-
roos-
ppb
1896"
842"
7
7
5
82
5857
3
6415
197
472
1566"
330"
1064-014
nglm2
22"
464
1"
r
635
9
21"
13"
427
55
3*
71
225
19
32
4"
I PPb
• 360*
7600
13"
[
| 10"
^0400
\ 152
340"
216"
7000
I 908
50"
; 7770
' 3680
306
524
60'
0490-026A
nglm2
85*
203"
276*
40*
4*
1111"
255
' 215"
213"
1984
1894
5758
12"
43
651**
18206
520
414
330"
73"
ppb
69*
165"
224*
32*
3"
903"
207
175*
173'
1613
1540
4681
10"
35
529"
14800
423
336
268"*
59"
0490-026B 1440-016A
nglm2
433
160*
<40
124
5*
167"
3017
594
136*
1095
457*
3150
6947
226
14
187
31218
70
371
5765
1263
372
654
ppb nglm2
1368 171*
505* 167"
<100
391
14*
526"
9526
1876 12*
428* <200
3458 12*
1443*
9947 112
21937 260
713 13*
43
589
98584 3655
220
1173 234
18205 176
3988 27
1174
2066
ppb
179*
175"
13*
<100
12"
118
273
14*
3838
246
785
28
Notes: " Value less than defined quantitation limit. Value is mojre uncertain than others.
Value is more uncertain than others due to continuing calibration drift.
< Pesticide confirmed as present in quartz fiber filter or PUF plug but not detected in cyclone due to
required for analysis. Value shown is approximately detection limit of cyclone catch sample.
Blanks and analytes not listed are not detected or. more properly, "less than detection limit."
(Continued)
high dilution
-------
I
Table 2.
Continued
1440-016B
Household
Dichlorvos
o-Phenylphenol
Propoxur
Bendiocarb
a-BHC
HCB
Atrazine
POP
g-BHC
Diazinon
Chlorothalonil
Carbaryl
Heptachlor
Malathion
Chlorpyrifos
Aldrin
Dacthal
Heptachlor epoxide
Oxychlordane
Captan
Chlordane
Folpet
2,4,0
DDE
Dieldrin
DDT
ODD
Methoxychlor
c-Permethrin
t-permethrin
nglm2
233*
698
9*
254*
86
357
25
74"
7076
589
254
405
6*
ppb
270*
629
8*
229*
77
327
22
67*
6374
530
229
364
6*
7647-007 7064-077 0966-027 0782-038
nglm2 ppb nglm2 ppb ng/m2
<200 <800 <7000 <1000 < 32000
<100 <500 47540*
<20 <90
<300 < 7000 57473*
<9000 <2000 < 39000
<20 <700 <700 <200 74729*
2073 8226 4437* 947* 59623
263* 7045 7390* 295* 7747*
36* 743
< 79000
7767* 4630 72965* 2753* 783894*
<3000
352* 7398 3672* 780* 7433*
8990*
5758*
27477*
26254*
ppb nglm2 ppb
<7000
7786*
2469*
<2000
633*
2564 793657 7747
333* 77270* 67*
<800
7908*
<700
320"
387*
248*
923*
7729*
Notes: * Value less than defined quantitation limit. Value is more uncertain than others.
** Value is more uncertain than others due to continuing calibration drift.
< Pesticide confirmed as present in quartz fiber filter or PUF plug but not detected in cyclone due to high dilution required
for analysis. Value shown is approximately detection limit of cyclone catch sample.
Blanks and analytes not listed are not detected or, more properly, "less than detection limit."
001) illustrates one of the more difficult
problems in the analysis of such multi-
component samples. In order to measure
.the maximum compound, it was
necessary to dilute the extract to the
point that the effective detection limit was
raised above the expected values of
many of the other compounds.
Some of the pesticides were observed
on the quartz fiber filter or the PUF plug
in samples from these four households
even though the compound was not
reported in the (diluted) cyclone catch
extract. The concentrations of these
pesticides have been reported in Table 2
as less than the detection limit for the
cyclone catch plus the observed value on
the filters.
Conclusions and
Recommendations
The use of the HVS2 in this nine-home
pilot study has shown it to be an effective
and efficient way to collect household
floor dust samples of sufficient size to
permit detailed chemical analysis.
Experience with the HVS2 suggested
some minor modifications to the device
to make it more maneuverable. Additional
study should be undertaken to determine
the variability of recovery efficiency of
the HVS2 with samples of real house
dust. The operating procedures and
sampling documentation were found to
be workable and complete, although
additional comments on the availability of
a portable hood for use during cleanup
should be added.
Analytical procedures were generally
satisfactory. However, high concentra-
tions of pesticides in some samples;
required substantial dilution before they
-------
Table 3. Summary of Dust and Pesticide Data
Analyte
Dust <150 nm
o-Phenylphenol
Propoxur
Diazinon
Heptachlor
Chlorpyrifos
Aldrin
Chlordane
DDT
DDE
Dieldrin
wiec
1.2 g/m2
0.2 ng/m2
0.4
0.2
0.3
5.6
0.3
12.9
0.4
0.5
0.4
nan
; ,
; 1 .3 ppm
' 0.6
0.4
: 0.1
4.7
: 0.3
i 6.3
0.4
0.3
0.5
Maxu
169 g/m2
8.7 ng/m2
41.5
57.4
14.7
193.7
11.3
183.9
9.0
0.6
7.4
num
--
1.9 ppm
7.6
10.4
9.9
21.9
1.0
98.6
4.0
1.2
18.2
could be analyzed. As a result, other
analytes were diluted below their
detection limit. A cleanup procedure
should be developed for the cyclone
catch that would remove interfering
compounds, allowing accurate measure-
ment of all the target compounds.
Several pesticides were observed that
have not been in widespread use for
many years. It is not known if these were
long-lived residues from earlier indoor
applications, recent applications of old
pesticide material, or material tracked in
from outside areas which were treated in
years past.
A relationship was observed between
the number of pesticides in the samples
and the age of the house. No relationship
was found with any of the other physical
or socio-economic variables. A significant
correlation was observed between sur-
face loading (e.g., ng/m2) and dust
concentration (e.g., ppmb) for only two
pesticides. For other pesticides the cor-
relation coefficient was generally small as
well as not significant.
When a pesticide was present in both
the air and dust samples, simple
statistical tests suggest, with a high
degree of confidence, that a relationship
may exist between the air concentration
and the concentration in the dust for
some pesticides. The sample size was
not sufficient to quantify the relationship
or to identify which pesticides could be
so related.
A PDF plug filter was necessary for
accurate measurement of alpha-BHC,
HOB, gamma-BHC, heptachlor, and
aldrin. The source of pesticides found on
the PUF plug could be either the indoor
air drawn through the HVS2 or pesticide
material blown off the dust in the cyclone
or on the quartz fiber filter. It was not
possible to refute either alternative with
the data available. A supplemental study
did show that j blowoff could occur over
time and was a function of vapor
pressure, especially over the short term.
Further stud^ will be necessary to
determine if the PUF plug, or a lower
pressure drop Alternative, is necessary to
retain these pesticides and other semi-
volatile compounds.
This work I leaves a number of
questions about household exposure to
pesticides unanswered. Although this
study has demonstrated the presence of
a wide variety pf pesticides in household
dust in significant concentrations, it is not
clear that this rnaterial presents any risk
to residents. However, preliminary
studies do suggest that the dust-
mediated pathway may be a significant
route of exposure, especially for very
young children. A series of studies
should be conducted to determine the
sources of these pesticides and the
mechanism of [transfer, if any, from the
dust to residents. Specifically, 1)
longitudinal studies should be made of
pesticide application events, 2) studies
should be made of surrounding soils to
evaluate "track-in" as a source for the
pesticide materials, and 3) studies of
pesticide materials on residents, or dirt
and surrogates' on very young children,
should be conducted to determine the
amount and manner of transfer from
house dust to residents.
References
Davies, D.J.A., Watt, J.M., Thornton, I.
1987. Lead Levels in Birmingham Dusts
and Soils. Sci. Total Environ. 67:177-185.
Roberts, J.W. and Ruby, M.G. 1988.
Development of a High Volume Surface
Sampler for Pesticides in Floor Dust. U.S.
Environmental Protection Agency
(EPA/600/4-88/036, PB 89-124630).
-------
-------
W. 7. Budd, J. W. Roberts, and M. G. Ruby are with Environmetrics, Seattle, WA
89103.
Robert G. Lewis is the EPA Project Officer (see below).
The complete report, entitled "Field Evaluation of a High Volume Surface
Sampler for Pesticides in Floor Dust," (Order No. PB 90-192 006/AS;
Cost: $ 17.00, subject to change) will be available only from:
National Technical Information Service I
5285 Port Royal Road '
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Prefect Officer can be contacted at:
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC 27711
United States
Environmental 'Rrtitecfion
Agency
Center for Environmental Research
UnfomnatioiJ ;
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-90/030
------- |