-------
DISCLAIMER
The Information In this document has been funded wholly or In part by
the United States Environmental Protection Agency under Contract 68-02-4406
to PEI Associates. It has been subjected to the Agency's peer review and It
has been approved for publication as an EPA document. Any mention of trade
names or commercial products does not constitute an endorsement or recommen-
dation for use.
1i
-------
FOREWORD
Measurement and monitoring research efforts are designed to anticipate
environmental problems, to support regulatory actions by developing an In-
depth understanding of the nature and processes that affect health and the
ecology, to provide Innovative means of monitoring co-npllance with regula-
tions, and to evaluate the effectiveness of health and environmental protec-
tion efforts through the monitoring of long-term trends. The Environmental
Monitoring Systems Laboratory, Research Triangle Park, North Carolina, has
responsibility for assessment of environmental monitoring technology and
systems, Implementation of agency-wide quality assurance programs for air
pollution measurement systems, and supplying technical support to other
groups In the Agency, Including the Office of Air and Radiation, the Office
of Toxic Substances, and the Office of Solid Waste.
The determination of human exposure to toxic organic compounds 1s an
area of Increasing significance to EPA. The development of a new sampling
methodology for a heretofore poorly understood media will permit Important
new information to be gathered so this potential route of exposure to these
compounds can be properly evaluated.
Gary J- Foley
Director
Environmental Monitoring Systems Laboratory
Research Triangle Park, North Carolina 27711
ill
-------
ABSTRACT
House dust and the pollutants carried with house dust are potentially
Important contributors to exposure througS the pathways of Inhalation,
1ngest1on and skin penetration, especially for small children. Pesticides
may be one of the more Important contaminants of house dust.
A Mgh volume surface sampler (HVS2) for the collection of house dust
and the semi volatile organic; 1n house dust has been designed and tested.
The sampler consists of an Intake nozzle, cyclone, and filter. The position
of the nozzle Is regulated hy the static pressure In the nozzle. The HVS2
operates at approximately 9.5 L/s (20 cfm) and can collect more than 2 g of
floor dust from a rug 1n an average clean residence 1n less than 4 mln.
Over 95X of the sample 1s retained In the cyclone and would, thus, be usable
as a bulk sample for bloassays.
The KVS2 collects approximately 30X of the dust less than ISO urn from
level loop and plush carpets. It collects 93.4X of the total dust from a
smooth bare floor. The variation In the collection efficiency of fine dust
1n repeat trials on carpets was 6X for level loop carpets and 12X for plush
carpets. The variation as the surface loading was changed was of the same
order.
Previous studies of ambient < amp 1 ing for pesticides suggested that
semlvolatlle organ1cs In house dust would not be retained on the filter and
a polyurethane foam (PUP) absorbent filter would be necessary to collect
tnem. Both house dust and a test dust were spiked with 10 or 20 ppm chlor-
pyrifos and dleidrln and 50 or 100 ppm dlazlnon. Virtually all the pesti-
cide was retained In the cyclone or on the filter. Although a PDF filter
does not appear to *»e necessary, 1t can be used with the NVS2.
Several alternative sampling methods were also studied. The collection
efficiency for fine dust of conventional upright and canister-type vacuum
cleaners, as well as small hand-held vacuum cleaners, was not sufficient and
use as required here would have been difficult or Impossible. Gloves, semi-
sticky paper, and carpet squares were also found not to be satisfactory
substitutes.
A recommended sampling procedure for the use of the HVS2, based on the
test experiences reported here, Is provided.
This report Is submitted in fulfillment of Agreement for Services 7-
7009 CKX between Environmental Monitoring and Services, Inc. and Engineering
Plus, as a part of Subcontract No. 725-86 with PEI Associates under Contract
No. 68-02-4406 with the U.S. Environmental Protection Agency. This report
covers the period from 3/1/87 through 8/31/87, and work was completed as of
8/31/87.
iv
-------
CONTENTS
Page
Foreword 111
Abstract 1v
Figures and Tables vl
Acknowledgments v11
1. Introduction 1
2. Conclusions and Recommendations 2
3. Literature Review 4
Public Health Implications of Surface Dust .... 4
Characteristics of House Dust 5
Sampling Techniques 7
4. Sampler Design 9
Performance Goals 9
Initial Configuration 9
Modifications 11
5. Sampler Testing 14
Testing Procedures 14
Test Dusts 15
Analysis Procedures 16
Sampler Performance 17
Dust Collection Efficiency Tests 20
Pesticide Collection Efficiency Tests 21
Field Cleanup Procedure 24
6. Alternative Sampling Methods 27
Conventional Vacuum Cleaners 27
Hand-held Vacuum Cleaners 28
Dust Traps 28
7. Recommended Test Procedures 30
8. References 38
-------
FIGURES
Number Page
1 A1r flow schemata of HVS2 10
2 The HVS2 (front view) 12
3 The HVS2 (s1d'.« view) 12
4 Example chnMiatograms 18
5 Velociy In carpet for different nozzle static pressures. . 20
6 Ideal arrangement of rug test segments 36
7 Recommerded data sheet 37
TABLES
1 Size distribution of test sand 15
2 Instrumentation and operating conditions 17
3 Physical properties of tested pesticides 22
4 Mass balance of pesticides In sample and collected dust . . 23
5 Pesticide concentration In sample and collected dusts ... 24
6 Instrumentation and operating conditions 25
7 Amount of pesticide In clean dust sample 26
v1
-------
ACKNOWLEDGMENTS
The authors are grateful to Richard W. Boubel and Jack Hlrsch of CS3
for their work in the design and construction of the HVS2, to the Hoover
Company for their continuing assistance to this project, to Gerald Pade,
William Budd, and Wen Wei for their work In testing the surface sampler, and
to David A. Kalman, David L. Eaton, Carl Calleman, George G. McCaslIn,
Timothy V. Larson, and Lee E. Hontleth of the Un1ver1sty of Washington for
their insights, advice, and assistance during the course of this work. The
comments of our Project Officer, Nancy K. Wilson, Russsll W. Weiner, and
Merrill D. Jackson on the first draft added greatly to the final report.
Robert G. Lewis, Chief, Methods Development Branch, and James E. Howes, our
Project Manager for EMSI, gave Invaluable advice on the design and execution
of the testing program, particularly the pesticide evaluation tests.
-------
SECTION 1
INTRODUCTION
Recent studies of human exposure to air pollutants have Increasingly
recognized the Importance of pathways other than Inhalation (Ott et al.,
1986). Ingestlon of air pollutants deposited 1n water or on soil 1s poten-
tially the source of a significant portion of an Individual's total pollu-
tant burden, especially for very young children, with their lower body
weight and frequent hand-to-mouth activity. Direct penetration through the
skin may be Important for some organic materials, particularly pesticides.
Failure to consider dust as a pathway for air pollution may result 1n a
significant underestimation of health risks.
Dust can be both a medium for the transfer of pollutants from sources
to people and a medium for the accumulation of pollutants. House dust may
contain bacteria, viruses, allergens, smoke residues, pesticides, asbestos
particles, paint fragments, solvents, flame retardants, cleaners, fragments
and residues from synthetic fibers, building products, and a multitude of
other materials and pollutants created by activities 1n the home or tracked
1n from road dust, soil, or work sites.
At present there 1s no validated method for sampling the dust present
on surfaces, although a wide variety of techniques has been used by various
researchers. Unfortunately the variations In the methods used permit little
comparison between one study and another.
A useful sampling technique should have a known and reproducible re-
moval rate for dust on various surfaces and be able to achieve a relatively
constant efficiency at different loadings of surface dust accumulation. In
addition, the procedure should be able to collect and measure the low and
medium volatility organics which are expected to be found on dust particles.
Finally, the data collected should be capable of being compared to the
actual dust uptake by a test subject. This report describes the development
of a sampler designed to meet these objectives.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
The high volume surface sampler (HVS2) constructed by Cascade Stack
Sampling Systems (CS3) 1s an effective and efficient way to collect samples
of fine surface dust. A bulk sample of more than 2 g can be collected In
about 4 siin In an average clean residence.
The static pressure In the nozzle was found to be the best measure of
the appropriate height for the nozzle on carpets. When operated at the
defined optimal settings, the fine materials (less than 150 urn) collected
from carpets by the HVS2 are approximately 6% of the total load of a stan-
dard test dust and approximately 30% cf the fine materials 1n the test dust.
The test-to-test variation Is about 8% on level loop carptets and 12% on
plush carpets. Better than 93% of the test dust 1s collected from a bare,
hard surface.
Oust wall losses In the system were approximately 40 mg regardless of
the number of tests or the total loading on the system between cleanings.
This loss 1s Insignificant In comparison to the greater than 95% of the
total mass collected In the cyclone.
Semlvolatlle organic materials on the test dusts were retained on the
collected dust. Experiments with a test dust which contained organic mater-
ial, elemental carbon, sand, and talc found that a polyurethane foam (PUF)
absorbent filter was not necessary for collection of the three pesticides
tested. When both house dust and the test dust were spiked with 10 or 20
ppm of chlorpyrifos and dieldrin and 50 or 100 ppm of dlazlnon, less than
0.1% of the pesticide was found on the PUF filter. A mass balance on the
pesticide demonstrated mass closure to within the accuracy of the measure-
ment procedures.
One test done without the cyclone suggests the short sampling time Is
the primary reason the semi volatile compounds are not stripped from the
collected partlculate matter by the air stream.
The HVS2 can be used to measure complex mixtures of metals, solids, and
organics on a variety of surfaces. Perhaps one of the most obvious uses is
In support of studies of the health effects of indoor air pollutants and
studies of the relative Importance of pollutant pathways. Outdoors, the
sampler could be used to measure pollutant accumulations in potentially air-
mobilized soil surfaces. This might be useful In Investigations of the
potential risks associated with fugitive dust from hazardous waste land dis-
posal sites, for example.
A field test of the HVS2 would be an important next step for evaluating
this Instrument and the recommended procedures provided In this report.
-------
Because surface dust Is an Integrated record of the pollutants Introduced
Into the air above the surface, 1t 1s Important that such a field test
Include air sampling of both dust and gases during a period before and
between the collection of surface samples.
There are significant differences between suspended and surface dust.
The means of introduction of the dust, the size distribution, the half-life
of retention In a space, and the modes of exposure to the dust are quite
different. Each of these factors must be considered in designing the samp-
ling program for a field test which compares surface dust and suspended
dust.
The size distribution of house dust, the size distribution of dust on
the hands of small children, and the size of particles which pass the
cyclone and are found on the HVS2 filter should all be measured in order to
more properly characterize the results obtained from the HVS2.
Although the tests reported here support a conclusion that no PUF
absorbent filter Is required for some semivolatile organics, this should be
confirmed for more volatile compounds. It is suspected that the same
conclusion will be reached, as the more volatile compounds will also be less
likely to be found in the dust.
While the tests of alternative procedures did not find a simple proce-
dure which can meet the performance goals for the KVS2, an exploration of
such methods should continue, perhaps with the goal of finding a screening
method which would not yield data that meet rigorous standards but could be
used inexpensively for large, preliminary samples.
-------
SECTION 3
LITERATURE REVIEW
Public Health Implications of House Dust
Paniculate matter, especially fine, carbonaceous participate matter,
has been shown to be an efficient scavenger, by adsorption, of semivolatile
Materials tnd other particles (Natusch und Wallace, 1974). House dust and
orpet fibers, In particular, have been f;und to be coated with a wide
variety of other, smaller particles and droplets of organic materials
(Haeske et al., 1957).
The metals In house dusts, particularly lead and cadmium, have been
extensively studied (e.g., Vostal et al., 1974; Solomon and Hartford, 1976;
Harrison, 1979; Diemel et al., 1981; Selfert et al., 1984; Rablnowltz et
al., 1985). Organic materials 1n suspended house dusts have also been
analyzed (e.g., Haeske et al., 1957; Vocom et al., 1971; Starr et al., 1974;
Davies et al., 1975; Imhoff et al., 1982; Lloy et al., 1985).
Activity In the house will resuspend deposited house dust (Anis and
An1s, 1972 and Hunt, 1972), which can then be Inhaled. Possibly more Impor-
tant, deposited house dust also can be Ingested directly, especially by
children. Many pesticides and some chlorinated and other organic compounds
present 1n residences can be absorbed directly through the skin, presenting
yet another pathway for these hou*edust-assodated pollutants (Wester and
Maibach, 1987).
Most children go through a phase of putting things 1n their mouths.
For children In the crawling stage, this means putting Into their mouths
fingers that have been in contact with the floor or ground. Studies have
estimated that between 6 and 12% of small children also compulsively eat
non-food material, such as dirt, soap, paper, or paint chips, a practice
termed pica (Nahaffey, et al., 1985). The toddler has only about one-fifth
the body weight of an adult and Ingests an estimated 2.5 times, or more, as
•uch dust, which Increases the potential health risk to the child of con-
taminants 1n the floor dust at least 12 times (Hawley, 1985). The risk to
such children will be further Increased by their higher ratio of surface
area to body weight and the stage of development of their organs, nervous
system, and immune system (IPCS, 1986).
An estimate of average daily soil dust Ingestlon by small children
ranges from 0.10 to 0.50 g, with a potential for higher rates for children
with pica who live in a dusty home or who have access to exposed soil
(LaGoy, 1987; cf. Binder, et al., 1986). Although the long residence times
in the lung of Inhaled dust would suggest that a high percentage of Inhaled
toxic materials will enter the blood stream, studies of children in areas of
heavy contamination have found a better correlation between the body burden
-------
of pollutants and the amount of the pollutant on their hands than with the
concentrations In the air (Roels et al., 1980; but cf., Oavles et al.,
1987). Davidson and Ellas (1986) estimate that a two year o'd child could
take In forty times as much lead from Ingesting dust contaminated by air
pollution as from directly Inhaling lead partlculate matter. Murphy and
Yocom (1986) provide a worst case estimate of almost 1000 times as much
long-lived soil toxics taken In by dust Ingestlon as by Inhalation for a
young child (2.5 years).
The retention on the hands of different size dust particles has been
measured by several researchers. Que Hee et al. (1985) found an approxi-
mately constant 30 mg of dust retained when hands Mere presented with dust
segregated Into different sizes. Ougan and Insklp (1985) and Dugan et al.
(1985) measured the particle size of dust removed from hands and found a
higher percentage of particles In the sizes less than 10 micrometers. How-
ever a re-analysis of their data looking at surface area, suggests a poten-
tially greater Importance of the particles between 10 and 100 micrometers, a
size range which they did not characterize.
Childhood Intake may represent a major source of the total adult body
burden of some pollutants. Data developed by Henke et al. (cited by Fox,
1979) suggests that up to one-third of the adult body burden of cadmium may
be accumulated by age three. Plnkerton (1973) reports that one-half of the
adult body burdens of cadmium and lead accumulate by late adolescence.
The Integrated mutagenlc activity of house dust has been found by
several research teams to be significant, although the Implications of the
measurements are not clear (e.g., van Houdt and Bolelj, 1984; Lloy et al.,
1985; Roberts et al., 1987). The measurements of floor dust mutagenlc
activity by Roberts et al. (1987) are higher but comparable to the results
from selected soils (Brown et al.,19P5).
Characteristics of House Dust
The sources of house dust are highly variable and studies have not
provided any detailed generalization of Its composition. Residues from food
and food preparation, skin scales and other organic residue from the human
and animal occupants of the house, fibers from clothing and household fur-
nishings, fibers and minerals from building materials, organic matter from
cleaning compounds, waxes, and other consumer products, fragments of vegeta-
tion and humus, and mineral particles, such as clay and sand, have all been
identified In house dust.
A study in six U.S. cities found between 37 and 86X of house dust was
Inorganic (of which about two-thirds was silica and clay), less than IX to
5% was grease, and the remainder other organlcs (J.N. Balough, personal com-
munication. 1986). A German study (Haeske et al., 1957) found significant
differences in the composition of airborne dust, dust recovered from a hori-
zontal shelf, and carpet dust. More dry organic matter was present In the
airborne and horizontal-shelf dust than the carpet dust, which had more
silicaceous material. A greater percentage of the airborne dust was soot
and other small particles.
-------
A limited survey of 29 homes 1n two areas 1n Seattle, HA (Roberts et
al.. 1985) found an average (removable) accumulation of floor dust of 19
g/m* and a (removable) accumulation rate of 1.6 g/^/day. In three homes
the dust recovered from rugs was found to be from 15 to 140 times greater
than that collected from bare floors In the same home.
Schaefer et al. (1972) found a mean Indoor dustfall rate of 0.02
g/m?/day In 100 homes In five cities. The amount of accumulated dust on
floors was measured by Solomon and Hartford (1976) In 12 residences and
found to be approximately 1.2 g/m2. The striking difference between the
amount of dust observed depositing from the air on the floor and the amount
of dust that accumulates on the floor suggests track-In as a major source.
One study of lead In house dust, which carefully compared the deposited (9.3
ug Pb/m'-day) and accumulated (166 ug Pb/m2-*--) dust levels concluded that
most of the lead 1s carried Into the house wUn "dust particles, adhering to
shoes, etc." (Dlemel et al., 1981).
The accumulation of dust 1n rugs can be expected to vary with the type
of carpet, the frequency and manner of cleaning, the sources of dust Inside
and outside the home, and the behavior of the occupants. Studies of vacuum
cleaner efficiency have found that the addition of a power-driven agitator
Increases the amount of dust and rug fiber pickup by a factor of 2 to 6. An
upright vacuum cleaner, with a power-driven agitator, can be expected to
remove approximately 35 to 55X of the accumulated dust from a plush rug and
70 to 80% of the dust from a flat rug. The canister or tank cleaner,
without a power-driven agitator, can be expected to remove about 10 X
of the dust from a plush rug and 40 to SOX of the dust In a flat rug (J.N.
Balough, personal communication, 1986).
It.Is estimated that about 8X of the homes have no vacuum cleaner and
1n an additional 17X the vacuum cleaner Is broken, or It 1s operating far
below design efficiency because of a loose belt, a full dust bag, or some
other maintenance problem (J.N. Balough, personal communication, 1987). A
removable (total) dust accumulation of 250 g/m2 was measured by Roberts et
al. (1987) 1n one home with no vacuum cleaner and an accumulation of 1000
g/m2 was measured In a high traffic area of a retail store (CAMRASO, 1983).
Visible dust can be raised by the stomp of a foot at these accumulation
levels.
The size distribution of dust removed from carpets 1n three studies 1n
1928, 1961, and 1975 varied little with time. Between 14 and 20X were
retained on a 50 mesh (300 urn) screen and between 51 and 54X passed a 200
mesh (75 urn) screen (J.N. Balough, personal communication, 1986 and Barager,
1961). Roberts et al. (1985) found approximately 40X (by weight) of the
house dust collected from 29 homes to be less than 150 urn.
Pesticides and other organic* with lower vapor pressure or higher
polarity are expected to partition more to dust than to air, making house
dust a potential major reservoir of persistent pesticides, such as chlor-
dane, and PAH compounds, such as benzo[a]pyrene (BaP). However, the propor-
tion of semi volatile organic compounds partitioned to the dust Is not clear.
Heasurements by Bamesberger and Adams (1966) found between 16 and 98X of the
2,4-0 herbicide esters 1n the partIculate phase.
-------
Measurements of chlordane and heptachlor In homes treated for termites
found significant levels of the pesticides In carpet swatches placed on the
floor In the homes (Wright and Letdy, 1982). Studies of house dust (Davles
et al., 1975; Starr et al., 1974) have found significant levels of pesti-
cides in the dust. Both BaP and pesticides have been measured at greater
concentrations 1n house dust than In the soil near the houses (Davles et
al., 1975 and Bailey, 1981).
Pentachloraphenol (PCP) and llndane have been found in housedust 1n
homes where wood preservatives had been used (Krause et al., 1987). The
dust concentrations of PCP were observed to be correlated to air concentra-
tions 1n homes where the application had occured within the past two years,
but the correlation was absent If the application was older than two years.
The concentrations of PCP In the dust tended to remain relatively high,
while the concentrations 1n the air dropped by about a factor of four.
Sampling Techniques
A wide variety of techniques have been used for collecting dust. Use
of moistened glass fiber filters wiped across surfaces that may collect dust
Is a standard technique 1n Industrial hygiene (OSHA, 1977). Moistened
towels have been used 1n a similar manner (Vostal et al., 1974). This
procedure has also been used for collecting water-soluble Ions from surfaces
(Sinclair, 1982). Adhesive tape (Royster and F1sh, 1965) has been used to
lift dust from surfaces for visual examination. Oust fall Jars (Selfert et
al., 1982} and other test surfaces (Schaefer et al., 1972 and Anzal and
Klkuchl, 1978) have been deployed to collect dust. Accumulated dust has
been collected by brushing from seldom-cleaned surfaces (Weschler, 1978).
Small test areas have been hand vacuumed using a cut section of tubing as a
nozzle and collected on a small, high-efficiency filter (Solomon and
Hertford, 1976 and Que-Hee et al., 1985). Dust samples have been collected
from family-used vacuum cleaner bags (Angle and Mclntlre, 1979). Specially
adapted (Oiemel et al., 1981 and Milar and Mushak, 1982 ) and standard
vacuum cleaners have also been used to collect samples directly, with some
analysts sieving the collected dust to remove larger material (Harrison,
1979). With the exception of the work by Que-Hee et al. (1985), there has
been very little effort made to validate the collection techniques used.
Vacuuming of dust from surfaces requires the dust at the surface to
be mobilized by the air flow Into the collector. Gillette et al. (1980)
conducted extensive tests of disturbed and undisturbed desert soil surfaces
to determine the threshold velocities at the surface required to mobilize
particles from the surface. For disturbed soils, which are expected to most
resemble the dust layers on floors and in carpets, the threshold velocities
for aeolian erosion were between 0.2 and 0.6 m/s (40 and 120 ft/mln).
In contrast to surface sampling of dusts, ambient sampling for semi-
volatile organic compounds has evolved to a proven sampling train (Lewis et
al., 1977; Lewis and Jackson, 1982). Early measurements of ambient pesti-
cides often utilized samplers similar to the high volume sampling system for
ambient total suspended participate (TSP). Study of these samplers led to
the conclusion that substantial portions of the organic compounds were being
volatilized and blown off the filter over the 24-hr sampling period (Lewis,
1976). In the current ambient pesticide sampler, the partlculate filter Is
-------
backed up by a trap specific for capture of the semivolatile organlcs.
Open-cell PUF has been found to be an effective collector for some of these
compounds, with an overall recovery of 70 to 100 X for many pesti-
cides. Soxhlet extraction of the filter and the 3-1n long PUF plug permit
analysis by gas chronatography.
-------
SECTION 4
SAMPLER DESIGN
Performance Goals
A useful sampling technique for house dust should have a known and
reproducible removal rate of dust on various surfaces and be able to achieve
a relatively constant efficiency at different loadings of surface dust
accumulation. Essentially all the paniculate natter In the house dust
(Including the very fine materials) should be collected at a high efficien-
cy. The size distribution of the material retained for analysis should be
similar to that which would stick to a child's hand or skin. In addition,
the procedure should be able to collect and extra:! the low and medium vola-
tility organlcs that are expected to be found on i
-------
The blower permits a maximum flow rate of 14.2 L/s at standard tempera-
ture and pressure. The flow can be reduced to about 1.9 L/s If desirable.
The flOK Is measured by a calibrated (NBS traceable) laminar flow element
(LFE).
As shown In the schematic in Figure 1, the dust Is entrained In the air
flow to the slot nozzle, which can be positioned at the desired distance
above the surface. A simple slot nozzle was chosen when preliminary tests
with conventional vacuum cleaners showed a higher collection efficiency for
fine participate matter with this type of nozzle. The agitating "power
head* nozzle used In many conventional vacuum cleaners does collect more
mass, but it was primarily larger particles and "carpet fluff". Although
agitation may better mimic the activity of a small child, the smaller par-
ticle size collected by the slot nozzle Is expected to better resemble the
particles which are collected on the child's skin.
The sample passes from the nozzle to a 7.6 cm body diameter cyclone
which has a calculated partlculate cut diameter of 5 micrometers at the
maximum flow rate. The larger particles are removed by the cyclone and
collected in the jar at its base.
From the cyclone, the sample passes into the filter section, which was
a 142 mm diameter quartz fiber filter in the prototype. A PUF plug was
located behind the filter to collect the semivolatlle organIcs. The PUF
•plug is cut 5 cm in diameter and 7.6 cm long.
Blower
Laminar
Flow
Element
>5 jim
particles
Figure 1. Air flow schematic of the prototype high volume surface sampler.
10
-------
The blower Is connected to the sampler by a flexible hose which allows
the blower to be located out of the room being sampled. This prevents
disturbance of the air In the area being sampled and also reduces the blower
noise.
Modifications
Testing of the HVS2 Identified several design problems with the proto-
type unit. Some of these were resolved by field modifications to. the proto-
type unit and others were resolved 1n the redesign and construction by CS3
of the final, delivered HVS2 unit. Photographs of the final design HVS2 are
provided In Figures 2 and 3.
It was found that the unit oust be quite rigid and balanced fore to aft
1f the nozzle Is to be held at a uniform height, Independent of the movement
of the wheels over a surface. There was a tendency for the front end of the
unit to "nose-dive" at the start of a forward movement, which was especially
noticeable on plush carpeting. When moving over an uneven surface, such as
multl-level carpeting, the unit would bob up and down. In both cases the
nozzle could not be maintained at a constant height above the surface.
Several modifications were made in the design to correct these prob-
lems. Field modifications Included additional braces and clamps at stra-
tegic locations on the unit to make it more rigid. A new design of the
carriage and the handle solved the problem In final unit. Two additional
changes in the design of the final unit also make it easier to use on a soft
or uneven surface. The casters on the re*r of the carriage have been
replaced by larger, non-swivel ing wheels. The front wheels are closer to
the nozzle, so the nozzle will follow the surface. When the surface 1s so
uneven that 1t would be better for the nozzle to remain at a constant
height, the wheels can be replaced by wheels which fit a rigid frame, which
is then laid over the surface to be sampled.
The nozzle can be fixed in its relation to the surface more easily in
the final design. The nozzle is held In place by two guides which can be
tightened down. A threaded rod is mounted to the nozzle, so turning a
knob on the n>d provides a fine adjustment to the nozzle height.
Initially, the flow velocity at the nozzle was not sufficient. To
correct this problem, the nozzle area was reduced from 24 cm2 to 12 cm2. In
addition, an approximately 13 mm wide flange was added around the nozzle
face. To reduce the pressure drop through the system a series of experi-
ments (described In Section 5) were carried out which established that the
PUF plug could be eliminated. Consequently, the final design has no holder
for a PUF plug. However, this requires minimizing the face velocity across
the particle filter. This was accomplished by replacing the 142 mm diameter
filter with an approximately 20 by 25 cm filter, Increasing the face area
(and reducing the face velocity) by a factor of 3.3.
The most useful measure of the height of the nozzle above carpeted
surfaces was the static pressure in the sampling nozzle. As the distance
from the top of the surface decreases, or as the nozzle is pushed deeper
into less dense materials, such as a plush carpet, the static pressure
11
-------
Reproduced from
D..I ...ii.fti. copy
Figure 2» The h'.gh volume surface sampler (front view)
Figure 3. The high volume surface sampler (side view).
12
-------
Increases. This proved to be a sensitive measure which could be used rela-
tive to many surfaces. A pressure tap was added to the nozzle and a magne-
helk gauge was added to the control panel.
Other minor changes were made 1n the design to make It. more leak-free
and to make It easier to assemble, disassemble, carry, and use. Once the
correct flow rate had been determined, the LFE could be replaced by a less
expensive orifice plate. The motor was placed 1n a carrying case, to reduce
the noise and make It easier to move around.
13
-------
SECTION 5
SAMPLER TESTING
Testing Procedures
The testing procedures to establish the operating characteristics and
sampling reproducibility of the HVS2 relied heavily on the established
procedures for testing household vacuum cleaners (ASTM, 1987b). This method
measures the relative cleaning efficiency of vacuum cleaners on a standard
•carpet. -A standard plush (typical of residential Installations) and level
loop (typical of commercial installations) carpet were used.
In ASTM Method F608-79, a test dust of 90X sand and 10% talc Is spread
on and embedded into a test carpet by dragging a large, smooth weight across
the surface. The vacuum cleaner is run forward and back over a defined test
area, measuring 46 cm by 137 cm, in a specific movement pattern. The
efficiency of the cleaning is measured by the fraction of the applied dust
which is collected by the cleaner in 16 passes of the length of the test
-area. The speed of the cleaner movement is specified as 2.5 seconds per
'pass.
The uniformity of application of the test dust was tested by applying
semi-sticky squares of paper to the test area, uniformly pressing them onto
the carpet with a weight, and measuring their weight gain. Obvious carpet
filaments which adhered to the squares were removed prior to re-weighing.
No loss of weight on application to a smooth, clean surface could be mea-
sured. The coefficient of variation over 32 sample areas was 15X for the
plush carpet and 42% for the level loop. The higher variation for the level
loop 1s clearly a result of the denser surface where the embedment tool has
less tendency to even out the distribution.
Th» ASTM method specifies a procedure for cleaning the test carpet
between tests with a conventional upright vacuum cleaner with a power-driven
agitator. By using the cleaner first on the top and then on the reverse
side of the carpet, the applied test dust can be effectively removed In
three pairs of front and back cleaning operations. Tests during preliminary
phase of these experiments demonstrated that more than 96% of the mass of
the applied dust could be accounted for during an efficiency test-clean
cycle.
The ASTN test procedures were modified in several minor ways. First,
the carpets were placed directly onto a flat surface, rather than over an
open grill. The pads that the ASTM methods suggests using under the carpets
were not used. Experience has shown that neither of these changes has any
significant effect on the efficiency measurements (J.N. Balough, personal
communication, 1986). Because the nozzle was narrower than on a conven-
tional vacuum cleaner, it was necessary to make twice as many passes in
twice the time to cover the area. However, the speed of the cleaner was the
14
-------
same. Only two of the standard test carpets, * plush and a single-level
loop were used. In addition, tests were made on a smooth, hard surface.
Tne standard test method provides for only one level of test dust
application to the surface, approximately 160 g/mz. To provide test levels
which were more consistent with the observed loadings 1n residences, most of
the tests were conducted with loadings of approximately 80 g/mz and some
tests were made at about 32 g/m2.
Test Dusts
The ASTM testing method defines a test dust which 1s 90% sand and 10%
talc (both with specific size characteristics). The sand used In the tests
did not-conform to the size distribution given 1n the ASTN method. As shown
In TABLE 1, It contained more fines and less coarse material. However, the
available sand was used as supplied rather than being selved and recombined
as It was felt the actual distribution provided material 1n a size range
that was not present 1n the talc and thus would otherwise not be tested. It
appears from the literature that this size range 1s Important In adherence
to hands.
Four tests of small quantities of the sand, handled In the same manner
.as when It was mixed with talc for application on a test rug, found a
significant sample-to-sample variation 1n the size distribution. The re-
sults of these tests are also shown 1n TABLE 1. Some of the sample-to-
sample variation is the result of varying humidity, which could not be
controlled 1n the room where the sizing was done.
The size distribution of the talc was found to be more than 96% less
than 75 urn. It was noted that when the talc was mixed with the sand there
was actually less material less than 7Sum than with the talc alone. It was
hypothesized that the talc was lost by adhesion to the larger sand parti-
cles.
System tests with pesticide-Inoculated dusts used a different dust
composition. The critical system element in the pesticide collection effi-
ciency tests was assumed to be the filter. Past experience with ambient
pesticide collection suggested the pesticides would be stripped off the dust
on the filter by the moving air. Thus the amount of fines was Increased to
put a greater portion of the dust onto the filter. Additionally, an organic
TABLE 1. SIZE DISTRIBUTION OF TEST SAND
Size range >300um 300-150 urn 150-106 urn 106-75 urn <75 urn
ASTM standard 36 X 61 X 3 X trace 0 X
Permitted range 26 - 38 X 48 - 64 X 2 - 6 X < 1 X OX
Actual test sand 20 X 70 X 2 X 7 X IX
Small sample #1 14 X 69 X 3 X 13 X IX
Small sample #2 19 X 69 X 3 X 9 X IX
Small sample 13 33 X 65 X OX 2 X OX
Small sample #4 19 X 72 X 2 X 7 X OX
15
-------
material and a substitute for soot were added to the dust to better match
the reported composition of house dusts. The mixture used with the pesti-
cide tests was 45% sand, 45% talc, 9.5X food-grade cornstarch, and 0.5%
technical-grade graphite. The cornstarch and graphite were found to be more
than 99% less than 75 um.
In the pesticide tests, the pesticides dieldrin, diazlnon, and chlor-
pyrifos were mixed in measured amounts with approximately 30 ml of hexane
and the mixture blended with the test dust. The concentration of pesticide
In the dust was approximately 10 or 20 pom (by weight) dieldrin and chlor-
pyrifos and 50 or 100 ppm (by weight) dUzinon (because of an expected lower
sensitivity to diazinon in the analytical procedure). The slurry was dried
while rotating in a glass flask. Some vnall, Irregularly-shaped Teflon
articles were placed in the flask to help break up any agglomerations that
might form. Approximately 1 hr at room temperature was required for com-
plete drying. The pesticide-lnnoculated dusts were used Immediately after
drying.
Analysis Procedures
A size-specific efficiency was determined for the dust collection
efficiency tests. The actual loading applied to the test carpet was deter-
mined from the weight of the application container, before and after appli-
cation. Similarly, the weight of the collected dust was determined by
differencing the weight of the cyclone catch cup and the filter before and
after use. The Pall flex 2500 QAO-UP quartz fiber air sampling filters were
desiccated to constant weight and weighed before and after sampling. Dust
from the cyclone catch was weighed and seived according to ASTM Method D422-
63 (ASTM, 1987a) to determine the size-specific weight of the catch.
The size range that was found to be the most useful for analysis was
for the material passing the 100 mesh (<150 um) screen. Thus, in each of
the tests of dust collection efficiency the results are stated as a collec-
tion efficiency of fine materials, which is the percent of the total dust
applied to the carpet that was recovered and passed the 100 mesh screen.
This is only a fraction of the total efficiency for recovery of all material
from the surface. For example, in one test with a fines efficiency of 6.1%
the total collection efficiency for all sizes is 26.5%.
The fine materials (<150 um) applied to the carpet would be only about
20% of the total load, if the size distribution of the individual test dust
sample were the same as the bulk measurement of the dusts (and U too often
will not be, as discussed above). That means a reported 5% fines efficiency
1s approximately equivalent to a 28% recovery of the applied fine materials.
The talc (< 44 um) is a constant 10% of the total load and approximately 56%
of the total material <150 um.
Sampling train preparation, sample handling, and analysis procedures
are similar to those described by Sherma (1981), EPA (1984), and ASTM
(1986). The PUF plugs were precleaned by extraction for at least 16 hr with
a 5% ethyl ether in hexane solution. After extraction, the plugs were com-
pressed and vacuum desiccated to remove excess solvent. The Pallaflex 2500
QAO-UP quartz fiber filters were baked at 600 C for two hours. After
16
-------
precleanlng the PUP plugs and filters were wrapped and stored in hexane
rinsed- aluminum foil.
After each sampling run the sample train was cleaned with hexane,
washed with detergent, rinsed with water, cleaned with Chemsolv or nitric
acid, rinsed with distilled water, and rinsed with methylene chloride.
During the pesticide tests all materials in contact with the test dusts were
either glass, aluminum, stainless steel, or Teflon.
Collected pesticide samples were Immediately wrapped 1n hexane rinsed-
aluminum foil and cooled to 0 C. Filters were not desiccated. Refrigerated
samples were shipped by overnight courier to the analytical laboratory.
Dust and vapors collected during the pesticide collection efficiency
tests were analyzed for pesticide content by Environmental Monitoring and
.•Services-, Inc. Pesticide levels were determined by solvent extraction of
the pesticides from the dust, filter, or PUF plug, reduction of the extract,
and measurement by gas chromatography. Samples were extracted for 16 hr
with 300 to 500 ml of 15% ether in hexane. The extract was dried in a
sodium sulfate column and the volume reduced to 4 ml in a Kaderna-Oanlsh
apparatus. Further reduction to 2 ml was accomplished in a stream of dry
nitrogen.
The pesticide levels in the samples was determined by gas chromatog-
-raphy with electron capture detection. (The Initial determination of dia-
'zinon was by flame photometry, however an equivalent response factor was
established and further analysis used electron capture for all compounds.)
Two columns were used; one for quant 1 tat ion and one for confirmation of the
Identity of the peaks. Details of the instrumentation and operating condi-
tions are given in TABLE 2. An example of the chromatogram obtained is
shown in Figure 4.
Prior to extraction, each sample was spiked with 200 ng of octachloro-
napthalene, which provided a known for determination of sample recovery. For
24 samples, the recovery on the known was 84 +/-
Both field blanks and niethod blanks were analyzed. In one experiment,
TABLE 2. INSTRUMENTATION AND OPERATING CONDITIONS FOR PESTICIDE ANALYSIS
Quant it at ion Confirmation
Instrument Varian 3700 Varian 3700
Column DB-5, 30 m x 0.53 mm DB-608, 30 m x 0.53 mm
Injector temperature 220 C 220 C
Detector temperature 300 C 300 C
Column program 180 C for 1 min 164 C for 1 mln
180 to 300 C 9 15 C/min 164 to 260 C 9 12 C/min
Hold 300 C for 1 min Hold 260 C for 1 min
Carrier flow 10 cc He/min 10 cc He/min
Makeup gas flow 20 cr Nj/min 20 cc Nj/min
Detector ECD ECD
Sensitivity 1x1 1x1
17
-------
Hit HuUtS|:HOWCSB2t)
939BB *
•s.
M
o
o
O
u
scmr. i
MHCC trtm. i 2.en TO 10.00
kil
a
Samplr: 6-30-1 F
Sample Type: Filter
Column: DB-5
6
ni MUTES
IB
FILE: HOWESl:HOUESaIS
•S.M8
o
32.88B.
MNGC (HIN.): 2.999 TO 18 B88
Sample: 6-30-1 P
Sample Type: PUF
Column: DB-5
MINUTES
Figure 4. Example chromatograms.
IB
IB
-------
the entire procedure was carried out without the addition of pesticides to
the test dust. The reported pesticide levels on the field blank were near
or below the detection limit of the analytical method and, 1n every case, at
least a factor of 100 below the pesticide levels measured in the tests dusts
and the filter. Method blanks were obtained by performing extractions with-
out samples and then proceeding with sample analysis. In every case the
method blank results were below the detection limit.
Sampler Performance
The collection efficiency of the HVS2 was measured using the procedures
described abo»-e, with variations 1n two important variables in order to
better define the optimal settings. The static pressure in the nozzle and
.the system volumetric flow rate were each changed while the other was held
.constant.
The dramatic difference In the dust removal behavior between the level
loop and plush carpets is well Illustrated by the system response to a
change in nozzle volumetric flow, which is essentially a change in the
velocity of the flow across and through the carpet. With a nozzle static
pressure of -41.7 cm (-16.41n.) HgO, the collection efficiency for fine
material (passed 100 mesh) on the level loop test carpet increased by ap-
proximately 17% as the volumetric flow Increased 67%. With the nozzle
static pressure at -28.5 cm (-11.2 In.) HjO on the plush test carpet the
fines efficiency increased only 4% for approximately the same change in
volumetric flow.
Similarly, a change in the nozzle static pressure when the volumetric
flow Is.field constant shows a much greater Increase in collection efficiency
with the level loop carpet. This is accomplished by setting the nozzle
closer to or deeper into the carpet surface. This effectively directs the
air flow more and/or deeper into the carpet. With a flowrate of 9.5 L/s (20
cfm), a 45% change In the pressure increased the fines collection efficiency
by 23% on the level loop carpet. At the same flowrate on the plush test
carpet a 12% increase in the nozzle static pressure (the maximum Increase
possible) made almost no difference In the fines collection efficiency.
Thus an Increase in the operating conditions beyond -28.5 cm (-11.2 in.)
HjO nozzle static pressure and -9.5 L/s (-20 cfm) on the plush carpet would
not significantly Increase the collection efficiency of fine materials
(passing 100 mesh) while it would significantly Increase the difficulty of
operation. On the other hand, reducing the nozzle static pressure on level
loop carpets to -28.5 cm (-11.2 in.) H?0 would substantially reduce +he fines
collection efficiency. At -41.7 cm (-16.4 in.) H?0 nozzle static pressure on
the level loop carpet and -28.5 cm (-11.2 in.) HjO on the plush carpet, the
fines efficiency is approximately the same. For this reason, It was deter-
mined that these two values should be used as the nominal operational values
In the remainder of the study.
Measurements were made of the actual horizontal velocities within the
plush test carpet by inserting the probe of a hot wire anemometer Into the
carpet, replacing a small plug of fiber that was removed. Values were taken
at approximately 3 mm (1/8 In.) and 10 mm (3/8 in.} from the lip of the nozzle
19
-------
flange, near the center of the nozzle. The static pressure in the nozzle
was changed by adjusting the height of the nozzle, with the system volu-
metric flow re-adjusted to 9.5 L/s (20 cfm) at each setting. The results
are shown In Figure S. The velocities obtained within the carpet are In
the same range and generally exceed the velocities determined by Gillette
(1980) to be required for mobilization of surface dust.
The loss of material to the walls of the sampling system prior to the
filter was measured en several occasions. It was observed that approximate-
ly 40 mg of material could be retrieved from the walls Irrespective of the
number of test runs or quantity of material collected since the last time
the walls were cleaned. Wnile this amount Is significant by comparison to
the amount that is collected on the filter, under average conditions, 1t is
not significant by comparison to the amount collected in the cyclone. Fur-
ther, since the loss appears to be a dynamic equilibrium 1t Is only Impor-
tant if.there are large changes 1n the amount of material that potentially
can be recovered or sensitive analysis techniques are being used and cross-
contamination of samples may be possible.
Dust Collection Efficiency Tests
The collection efficiency of the HVS2 was measured on three different
surfaces at different surface loadings of test dust to determine the pre-
cision of the measurement and the dependence on loading.
300-
200-
^ ioo H
>s
O
>
50 H
30-
10
I
15
20
I
25
30
I
35
40
Nozzle Pressure (cm H20)
Figure 5. Velocity within the plush test carpet for different nozzle
static pressures at 9.4 L/s (20 cfm) flow rate.
20
-------
The collection efficiency from a smooth painted surface was measured
with the nozzle approximately 16 mm (1/16 in.) above the surface. The
static pressure In the nozzle was approximately 2.5 cm (1.0 1n.) H?0. The
total collection efficiency at a loading of 7.7 g/m2 was 93.4 +/• *-3X.
This Is a coefficient of variation of 4.6X. At the lighter loading of 2.4
g/m2 the total efficiency was 97.2%. Thus, the HVS2 reprodudbly picks up
essentially everything from a smooth, hard surface.
The collection efficiency from carpeted surfaces was tested on a plush
and on a level loop carpet. The three other types of carpet, flat (or
Persian}, multilevel, and shag, represent a much smaller fraction of the
market and were not tested. Both the total collection efficiency and the
fines collection efficiency (the percent of the total surface load collected
which passes a 100 mesh, or approximately 150 urn, selve) were measured.
The total collection efficiency on the level loop carpet at the nominal
settings (nozzle static pressure of -41.7 cm (-16.4 In.) h^O and 19.4 L/s
(20 cfm) volumetric flow at a surface loading of 80 g/mz) was 24.0 +/- 4.0%.
The fines collection efficiency under the same conditions was 5.8 +/- 0.5%.
This Is a coefficient of variation of 8.OX.
The total collection efficiency on the plush carpet at the nominal
settings (nozzle static pressure of -28.6 cm (11.2 In.) H20 and 19.4 L/s (20
cfm) volumetric flow at a surface loading of 80 g/mz) was 10.2 +/- 0.5%.
The fines collection efficiency under the same conditions was 5.6 +/- 0.7%.
This is a coefficient of variation of 12.2%.
The greater consistency between the collection efficiency on the two
different carpets for fines, as contrasted to total, may be due to the
greater dependence of these particles on air movement for mobilization and
collection and a lesser Importance of physical movement of the carpet fibers
to liberate the particles from the carpet. The latter is significantly
different between the two carpets.
The change In collection efficiency with changes 1n loading are not
substantial on either the level loop or plush carpets. Measurements at a
loading of 32 g/m2 on plush were lower, with 4.9% fines collection effi-
ciency. At a loading of 158 g/m2 the fines collection efficiency was
higher, at 6.1%. However, the four measurements at these loadings were all
within one standard deviation of the average from the more extended test
runs at 80 g/m2. On the level loop carpets the fines efficiency was 6.2% at
119 g/m2 and 6.2% at 32 g/m2. Again, all three measurements at these
loadings were within one standard deviation of the nominal loading effi-
ciency values.
Pesticide Collection Efficiency Tests
Previous studies of the retention of organic materials on particles
collected on an air sampling filter suggested that It would be necessary to
back up the filter with a PUF filter to ensure capture of any semi volatile
organics 1n the surface dust. A series of tests were conducted using the
pesticides dieldrln, chlorpyrifos, ana diazinon as representative compounds.
The physical properties of these compounds are given In TABLE 3.
21
-------
TABLE 3. PHYSICAL PROPERTIES OF TESTED PESTICIDES
Boiling Vapor Pressure
ChlorpyHfos 42 C 1.9 x 10'f torr
D1az1non 84 1.4 x 10'4
Dleldrln 176 1.8 x 10''
A series of six tests were run to evaluate the distribution of pesti-
cides from the test dust In the HVS2. Samples of the dust collected In the
cyclone catch cup and on the filter were analyzed separately. These tests
were run with the PUF plug In place and It was also analyzed for pesticides.
Two tests were run at a nominal 10 ppm of pesticide in the dust, two tests
at double that amount, one test with ordinary housedust, and one test with
the cyclone removed from the system.
S1h*ce there was no question about the recovery of the pest1c1de-1nnoc-
ulated dust from the carpets, only a concern that the dust once collected
might be stripped of Its semlvolatlle compounds by the air movement through
the sampler during testing, there was no need to sample the dust on the
carpet. Instead a small tray was constructed of heavy aluminum sheet, where
the pestldde-innoculated dust could be evenly distributed and conveniently
vacuumed Into the HVS2 in the same manner as from any smooth, hard surface.
Alternate spoonfylls of Innoculated dust were placed 1n a sample bottle
for analysis and spread onto the tray. The HVS2 was moved slowly over the
tray, collecting most of the dust for 4 minutes and back over the tray,
possibly collecting more dust, but primarily just letting air flow over the
collected dust, for another 4 minutes. This was designed to simulate an
extreme condition of surface dust collection.
The reference dust sample, the cyclone catch, the filter, and the PUF
plug were placed In containers and Immediately refrigerated. The HVS2 was
washed down with hexane and the washings poured Into storage containers for
evaporation. These were retained for possible future analysis. The other
samples were shipped In refrigerated containers by overnight courier to
Environmental Monitoring and Services, Inc. for analysis.
The analytical results were reported 1n either total mass or mass
concentration terms, depending on the size of the sample. These have been
converted to total mass terms for all samples and are presented In TABLE 4.
Also included In TABLE 4 1s an estimate of the pesticide in the unaccounted
for dust (mostly wall loss}, the difference between the measured sample
weight and the weight of the dust collected by the cyclone and filter. The
concentration of pesticides 1n the sample were used to estimate the amount
lost; the concentration 1n the fines on the filter may have been a more
representative of the fine materials which are deposited on the walls. The
results show a closure on total mass within the accuracy of the reported
sample recovery for the analytical procedure for all samples. If the higher
concentrations of the fines had been used in the calculation, the results
would have been closer to 100%. The first nominal 10 ppm sample overstates
the total amount of pesticide In the system. The unusually low entry for
lost material for this sample suggests the recorded weight of either the
sample or the cyclone catch may have been 1n error.
22
-------
In all of the samples shown in TABLE 4 the amount found In the PDF plug
Is either less than the analytical detection limit or slightly above It. In
no case is the amount even within a factor of 100 of the amount on the
filter.
Also to be noted on TABLE 4 are the results of the test with ordinary
house dust. This dust was collected by a canister vacuum cleaner from four
residences and then sieved through a SO mesh (passes 300 urn) screen. The
remainder of the test was carried out the same as the other tests of pestl-
cide-innoculated dust. The lower amount of pesticide recovered on the
filters reflects the lower percentage of the dust which was collected on the
filter. This suggests the test dust used in the pesticide collection effi-
ciency tests overstated the proportion of fines and thus the proportion of
pesticide exposed to air stripping on the filter.
One test was made with the cone of the cyclone blanked off. This
eliminated the size discrimination of the cyclone although It did Increase
the amount of wall loss. The result was to put the entire load onto the
filter. A much smaller amount of dust was used to avoid overloading the
filter. The high retention of pesticide and continuing low concentrations
of pesticide in the PUP plug found in this test suggests the primary reason
for the contrast between these tests and the experience with ambient samp-
lers is the dramatic difference in sampling times (8 minutes against 24
TABLE 4. MASS BALANCE OF PESTICIDES IN SAMPLE AND COLLECTED DUST (ug)
Diazlnon Sample Lost Cyclone Filter PUT Total X
Nominal 10-1 "55273 ~T7 49J"l ~4TT~ O 54175 107.6
Nominal 10-2 533.0 10.7 424.8 35. nd 470.7 88.3
Nominal 20-1 1140.0 37.0 1038.7 63. nd 1138.9 99.9
Nominal 20-2 1232.8 22.9 1053.3 23. nd 1099.4 89.2
House dust 326.0 2.8 287.9 1.2 -- 291.9 89.5
No cyclone 31.0 6.6 -- 22.5 nd 29.2 94.4
Chlorpvrifos
Nominal 10-1 80.1 0.3 80.4 7.8 0.1 88.6 110.6
Nominal 10-2 84.0 1.7 73.4 6.5 nd 81.6 97.1
Nominal 20-1 188.0 6.1 165.8 11.0 0.1 183.0 97.3
Nominal 20-2 184.4 3.4 160.8 7.7 nd 172.0 93.3
House dust 249.9 2.1 247.6 0.6 -- 250.3 100.2
No cyclone 9.3 2.0 -- 5.5 nd 7.6 82.4
Dieldrin
Nominal 10-1 99.9 0.3 92.0 9.2 0.1 101.6 101.8
Nominal 10-2 114.0 2.3 98.5 9.1 0.0 109.9 96.4
Nominal 20-1 268.0 8.7 232.5 16.3 0.1 257.6 96.1
Nominal 20-2 263.4 4.9 229.5 14.8 0.1 249.3 94.7
House dust 105.0 0.9 96.0 0.2 -- 97.1 92.5
No cyclone 11.8 2.5 -- 11.8 0.1 14.4 121.9
nd: not detected
23
-------
TABLE 5. PESTICIDE CONCENTRATION IN SAMPLE AND COLLECTED DUST (ug/g)
Diazlnon Sample Cyclone Filter Enrichment
Nominal 10-1 50.8 51.1 265. ~^
Nominal 10-2 53.3 44.9 239. 4.5
Nominal 20-1 114.0 109.0 423. 3.7
Nominal 20-2 121.7 107.4 172. 1.4
House dust 32.6 29.0 217. 6.5
No cyclone 45.4 -- 42.
Chlorpyrifos
Nominal 10-1 8.1 8.3 47. 5.8
Nominal 10-2 8.4 7.6 44. 5.2
Nominal 20-1 18.8 17.4 74. 3.9
Nominal 20-2 18.2 16.4 57. 3.1
House dust 25.0 25.0 104. 4.2
No cyclone 13.6 -- 10.1
Dleldrin
Nominal 10-1 10.1 9.5 55. 5 4
Nominal 10-2 11.4 10.2 62. 5.4
Nominal 20-1 26.8 24.4 109. 4.1
Nominal 20-2 26.0 23.4 110. 4.2
House dust 10.5 9.7 37. 3.5
K- cyclone 17.2 — 21.9
Notes: nd: not detected
hours).,
TABLE 5 reports the same Information In terms of concentration. The
partitioning of the pesticide by particle size can be seen 1n the slight
reduction 1n concentration 1n the cyclone and the substantial Increase 1n
the concentration In the filter. The average enrichment factor 1s 4.4.
Field Cleanup Procedure
The cleanup procedure described above would be prohibitively cumbersome
and time consuming for use 1n a field study, where 1t can be expected that
three to four houses should be sampled In a single day. An alternate
cleanup procedure was designed to decrease the time required and the amount
of supplies needed.
There are essentially two methods for removing the particulate-adherred
pesticides from the HVS2. One may either remove the dust particles them-
selves by physically dislodging the particles or the organic compounds may
be extracted from the particles with a solvent. Both approaches were used.
The field cleanup procedure begins with a rlnsedown of the sampling
train with a 1:1 mixture of hexane and acetone. A bristle brush is soaked
in the solution and the portions of the HVS2 1r, front of the filter scrubbed
down. These parts are then rinsed a second time and allowed to air dry.
24
-------
The rlnsate was collected and retained but not analyzed.
The field cleanup procedure was tested by first collecting with the
HVS2 a test dust which had been Innoculated with a high concentration of
three pesticides. The system was then cleaned with the field cleanup pro-
cedure. Next, a clean dust was collected 1n tne system. The cyclor.e catch
and the filter were then analyzed for the pesitlcldes In the Innoculated
dust.
The worst case established for the test was a high percentage of fines
In the dust and a high concentration of pesticides. The samp organic-rich,
high fines dust used previously 1n the pesticide tests was >:st was collected by the HVS2
In the same manner as previously. Next the sampling train was cleaned using
the procedure described above. Finally, IP g of the same dust, but not
.Innoculated, was collected by the HVS2. The cyclone catch and the filter
were pooled and analysed for the three pesticides. The procedure was re-
peated four times.
The test dust, clean cyclone catch, and clean dust filters were ana-
lyzed bj( Analytical Resources, Inc. The samples were extracted by sonica-
tion In 50 ml of hexane twice and the solutions combined. After concentra-
tion, half of the sample was treated with diazomethane to methyl ate the PCP.
All analyses were performed on a GC with an electron caputre detector using
the conditions given in TABLF 6.
The concentrations of the pesticides in the subsequent sample of clean
dust was less than 0.3 uo for <*ach of the pesticides, as shown in TABLE 7.
TABLE 6. INSTRUMENTATION AND CONDITIONS FOR FIELD CLEANUP ANALYSIS
Instrument
Column
Injector temperature
Detector temperature
Colu'.m program
larrler flow
Makeup gas flow
Detector
ftctenuation
Quantitatlon
HP 5890
DB-5, 15 m x 0.53 mm
250 C
375 C
150 C for 1 min
150 to 240 C 9 3 C/mln
Hold 275 C for 5 min
5 cc He/m1n
30 cc ArCH4/min
ECD
2 degrees
Confirmation
HP 5890
08-608, 15 m x 0.53 mm
250 C
375 C
150 C for 1 min
150 to 240 C 9 3 C/min
Hold 275 C for 5 m1n
5 cc He/mln
30 cc ArCH4/m1n
ECD
2 degrees
25
-------
TABLE 7. AMOUNT OF PESTICIDE IN CLEAN OUST SAMPLE AFTER FIELD CLEANUP
Test 1 Test 2 Test 3 Test 4
PCP <0.2* ug <0.2* 0.003 0.012
ChlorpyHfos 0.3 0.08 0.06 0.4
Dieldrin 0.1 0.02 0.014 0.14
* detection limit for this run
The surrogate of bromodlchloropheol was found to have an extraction effi-
ciency of 94% for the dust sample and 47% for the quartz filters.
We can then estimate the possible amount of cross-contamination that
might occur from the data in TABLE 7. If we assume the extreme case is
*.represented by the residue at the mean plus three standard deviations (and
'correcting for extraction efficiency), the upper estimate o* the amount that
would be found in a clean dust would be 0.437 ug, 0.711 ug, ar.'l 0.254 ug,
for PCP, chlorpyrlfos, and dieldrin, respectively. If the collocted sample
were 2 g, this would correspond to 0.35 ppm, 0.57 ppm, and 0.20 pm, in the
same order.
Thus under this worst case, using the field cleanup procedure might cau:e a
pristine location to be categorized as a background situation, but would
have little if any effect on the classification of sites with any signifi-
cant amount of pesticides present.
26
-------
SECTION 6
ALTERNATIVE SAMPLING METHODS
Several alternative test methods were evaluated, both to determine how they
compare with the HVS2 method and to determine 1f any of them should be
explored as possibly better approaches to determining ths concentration of
dust on floors.
Conventional Vacuum Cleaners
Several studies of house dust have collected grab samples from the bags
of residents' vacuum cleaners. While this Is a quick way to collect a large
number of samples 1n a retrospective study, 1t provides neither a consistent
sampling efficiency nor any assurance that the samples have retained any
semi volatile materials.
Both an upright and a canister vacuum cleaner were tested for cleaning
efficiency using the ASTM procedures previously described. In addition,
both were evaluated with respect to their usefulness 1n a sampling program
similar to that Intended for the HVS2. A number of problems were identified
which limited the potential applicability of each.
A convertible upright vacuum cleaner with i power-driven agitator
(Hoover Model U4367) was used to to clean the rugs between test runs In this
experiment. Although it performed that task quite satisfactorily, Its fine
materials recovery (less than 150 urn) was less than 2.3% of the test dust.
It was also observed that about six grams of the collected sample could not
be removed from the bag after the test. The material left 1n the bag may
have been a significant fraction of the fines. In addition, the power-
driven agitator and nozzle of the upright could not be cleaned without
excessive effort.
A canister vacuum (Kenmore Model 87870), without a power-driven agita-
tor head, collected fine materials of 1.3X of the test dust. Problems
similar to those with the removal of fines from the bag 1n the upright were
noted with the canister.
The collection efficiency of thf paper bag material 1s not known,
however It is undoubtedly less than the air sampling filters used in the
HVS2. The face velocity of the air througn the bags 1n the vacuum cleaners
is about 30% less than the face velocity across the filter In the HVS2.
This is expected to reduce the evaporation of semivolattle materials cap-
tured In the bag walls. However, the canister bag design brings almost the
entire air flow through the previously collected material which may have the
opposite effect. This also results in much of the fine material being
captured on larger fibers and hairs, making it difficult to separate and
27
-------
recover the fines. Further, an examination of the bags found several with
questionable seals at corners and with other possibilities of by-pass.
A conventional canister vacuum cleaner may provide a possible alterna-
tive to the HVS2. It would require the design and construction of a substi-
tute bag from high efficiency filtering materials (e.g., Gore-Tex laminates)
and a test program to determine If semivolatile materials can be retained on
the sample. If It 1s necessary to recover the materials that are trapped on
or In the bag surface, 1t will also be necessary to Identify a bag material
that can withstand both a pre-sampling clean-up and a post-sampling
extraction.
Hand-held Vacuum Cleaners
A small hand-held vacuum cleaner (Black and Decker Oust Buster Plus
Model 933) gave a fine materials recovery of 4.7X of the test dust on a
level loop carpet but only 0.6% on a plush carpet. This unit may have been
even more efficient than the conventional vacuum cleaners in picking up the
fine material, but It did not retain It. The fines were observed to pass
directly through the unit and exit the exhaust as a white cloud, which was
directly Into the face of the user. It was not determined If this was due
to low efficiency on the filter or a by-pass of the filter.
The use of an Industrial hygiene partlculate sampling cassette with a
cut section of tubing as a nozzle (Solomon and Hartford, 1976 and Que Hee et
al., 1985) does appear to avoid most of the problems with the other proce-
dures described In this section. However the time required to collect a
sample of sufficient mass 1n even a marginally clean situation may be pro-
hibitive. The size of the sample would be limited by the volume available
In the cassette.
Dust Traps
Three methods of collecting dust directly were explored. White cotton
gloves were tried as a surrogate for direct skin contact with a surface,
semi-sticky paper was used to pick up surface dust, and two types of carpet
were used to trap the dust 1n a home.
The gloves were tested by embedding 100 g of test dust In the plush
test carpet. They were were first weighed and then placed over powderless
surgeons' gloves on each hand of the test technician. The gloves were then
pressed Into the test carpet from 9 to 100 times and weighed a second time.
Additional tests were done on hall carpets near the front door of a ten unit
apartment and on a sidewalk within ten feet of a busy Intersection. None of
the tests showed a significant weight gam, even though one showed visible
soiling.
Semi-sticky paper (3M Post-Its, Models 7651 and 7652) was used to
measure the variation in the application of the test dust, as described
earlier. Although It was felt to be useful for this very limited intercom-
parison of loading, the experience suggested that It would not be useful in
any broader analysis of surface dust. Covering the entire test surface area
28
-------
with the squares was time consuming and recovered less than 0.5X of the dust
applied to the surface. It would not have been feasible to make size
measurements or any chemical analysis of the recovered particles.
It was hypothesized that carpets In a home act as a sink for dust In
the home and will come to a dynamic equilibrium with the house dust 1n the
room. If this Is true, then a small carpet sample placed in the home should
serve as a representative sample collector that could be brought back to the
lab for analysis. Five 35 by 46 cm (14 by 18 in.) cleaned and weighed
plush rug samples were placed in three homes for three weeks. The residents
were Instructed not to vacuum the sample rugs. The rugs were brought back
into the lab and reconditioned to the Initial relative humidity and re-
weighed. Although two rug samples did hive a weight gain H was much less
than the equilibrium accumulation expected. The other samples did not have
a significant weight gain.
Entry rugs are constructed 1n a different manner, with the specific
Intention of trapping dirt. An 86 by 132 cm (34 by 52 in.) flat entry rug
was placed In two homes for three weeks each. In one home it gained 0.9 g
and in the other 0.19 g. The weight gain was determined by Inverting the
rug on a clean piece of paper and vibrating the back of the rug for four
minutes using an upright vacuum. However, none of the dust collected passed
through a 200 mesh screen (75 urn).
29
-------
SECTION 7
RECOMMENDED TEST PROCEDURES
The following test procedure is recommended for the use of the HVS2 in
sampling surface dust.
1. Scope
1.1 This method can be used to determine the quantity of surface dust
accumulated on a surface and the amount of pesticides present in the
surface dust. It may also be used with appropriate modifications to
measure metals, organics, or other chemicals present in surface dust.
Collection of semivolatile organics may require an absorbent following
the filter.
1.2 The results of this test method should be stated in grams per square
meter of recovered surface dust less than 150 urn and nanograms of a
specific pesticide per meter square of surface and/or parts per billion
(ppb) concentration in the collected dust, to aid comparison between
tests.
1.3 This method may be used in areas where hazardous materials are present.
However, the method does not address the safety problems associated
with such use. It is the responsibility of each user to determine
appropriate health and safety practices prior to use.
2. Applicable Documents
2.1 ASTM Standards
D 422-63 Particle-size analysis of soils
D 4536-86 High Volume Sampling for Solid Particulate Matter and Deter-
mining of Particulate Emissions
F 608-79 Carpet-embedded Dirt Removal Effectiveness of Household Vacuum
Cleaners
2.2 Other documents
Manual of Analytical Quality Control for Pesticides and Related Com-
pounds. U. S. Environ. Prot. Agency (EPA 600/2-81-059)
Compendium of Methods for the Determination of Toxic Organic Compounds
1n Ambient Air (EPA-600/4-84-041, 1984)
3. Significance
3.1 Dust on the floor or other horizontal surfaces can be suspended into
the air by normal activities, where It Is available for Inhalation. It
has been estimated that small children may Ingest 0.1 to 0.5 g and
adults about 0.1 g of dust per day. Some pesticides and other organics
in the dust may be absorbed through the skin. This method will provide
a standardized way to measure the concentration and quantity of dust on
floors or other horizontal surfaces and help assess exposure and risk
from the air, skin, and ingestion pathways of exposure. The particles
below 150 urn in diameter can be collected from surfaces in a more
30
-------
reproducible manner. The limited evidence that 1s available suggests a
higher percentage of such small particles will stick to the hands and
skin.
4. Apparatus
4.1 Weighing scale(s) for samples, 0.1 mg to 1000 g.
4.2 Stopwatch.
4.3 Masking tape for outlining sample sections.
4.4 High volume surface sampler (HVS2) train, including a 13 cm (5 in.)
aluminum nozzle, an aluminum or stainless steel (SS) cyclone, and an
aluminum 20 by 25 cm (8 by 10 in.) filter holder. The filter holder
should be followed by a calibrated orifice or SS laminar flow element,
flow control valve, flexible tubing, and blower capable of 14 L/s (30
ft3/"iin). One magnehelic gauge is used to measure the pressure drop
across the flow measuring device and a second to measure the static
pressure in the nozzle. The magnehelic gauge for the nozzle should
have a range of 0 to 6.23 kPa (25 in. H20).
5 Humidity gauge.
6 Humidity chamber.
7 Teflon jar (and lid) for cyclone. A polypropylene jar may be used
when only dust accumulation is measured.
8 Glass sample jars.
9 Thermometer.
10 Rubber stoppers for leak check.
4.11 Sample recovery equipment.
4.11.1 Brushes equal to size and length of nozzle section and connec-
ting tubing.
4.11.2 50% hexane-50% acetone solution for cleanup. Approximately 250
ml required per location.
4(. 11.3 Wash bottles and storage containers for liquid.
4.11.4 Standard 100 mesh (passes 150 urn) screen, pan, cover, and
shaker.
4.11.5 Aluminum foil and clean manila folders and envelopes for han-
dling and storing filters.
4.11.6 Ice chest.
5. Summary of method.
5.1 A samel ing nozzle is used to draw air over a surface to mobilize loose
material which is collected in a cyclone and on a filter. The cyclone
is used to collect material above 5 to 10 urn in diameter and to prevent
overloading the filter. The cyclone cup can collect 100 or more grams
of sample. A total of at least 2 grains of dust are needed in the
cyclone to provide sufficient material for chemical analysis and bio-
assays. The cyclone also removes 95% plus of the dust from the air
stream to reduce evaporation of the pesticides on the dust. With a
short sampling period of 8 min or less U is estimated that less than
0.5% of the semivolatile pesticide compounds will evaporate and pass
through the filter. The flow measuring device is used to measure and
maintain a constant a'.r flow sufficient to collect dust particles and
to provide a stable cyclone particle size separation. A magnehelic
gauge is used to measure and maintain a predetermined static pressure
In the nozzle to ensure a constant efficiency of pick-up of particles
of various sizes.
31
-------
6. Reagents *nd rliters
6.1 Purity of Reagents. Pesticide grade chemicals shall be used to collect
and analyze samples when pesticides are to be measured. Unless other-
wise Indicated, It 1s Intended that all reagents shall conform to the
specifications of the American Chemical Society.
6.2 Filters. Quartz fiber filters without organic binder and with a low
carbon background equivalent to Pal Iflex 2500 QAO must be used when
pesticides or other organlcs are to be measured. When only dust accum-
ulation Is measured, air sampling filters with 99.99 % efficiency,
using 0 3 urn particles, are sufficient.
6.3 Sample Recovery. Pesticide grade acetone, hexane, or methanol In glass
bottles shall be used 1n recovery of sample from the cyclone bottle and
HVS2 train. Use of materials other than Aluminum, SS, teflon, or glass
for handling chemicals or sampling, have been associated with poor
results in collecting pesticide samples.
7. Sampling Strategy
7.1 A representative sample of at least 2 g of surface dust shall be
collated from a rug by sampling four 46 by 137 cm (18 by 54 1n.)
sections. On very clean rugs additional sections may be sampled 1f the
cyclone catch Jar does not show sufficient sample. If more thin one rug
is available the rug should be sampled where small children are more
likely to play.
7.2 A representative sample shall be collected from a bare or hard-surface
covered floor by measuring and sampling all the available area or until
2 g of sample are collected.
8. Calibration and Pretest Preparation
8.1 Calibration- The magnehelic gauge, orifice or laminar flow element,
thermometer, humidity gauge, and scale should be calibrated against a
primary standard.
8.1.1 Nagnehellc gauge. The magnehelic gauge must be calibrated against
a manometer or other primary standard In the orientation In which
it will be used In sampling. This calibration should be done
once a month or after 25 tests whichever comes first. One means
to check the magnehelic gauge Is to set up a flow through the
train. If the manometer and magnehelic gauge do not agree wHhin
3% the magnehelic gauge is leaking or Is in need of repair or
calibration. A minimum of five flow rates should be used In the
calibration.
8.1.2 Orifice or laminar flow element. A full calibration of the ori-
fice should be performed before use and should be done a minimum
of once a year. The orifice or laminar flow element should be
designed for easy visual Inspection and should be cleaned regu-
larly. They should be calibrated by use of either a splrometer
with a minimum capacity of 0.3 m3 (10 ft3) or a roots meter.
8.1.3 Thermometer. Calibrate thermometers against mercury-in-glass
thermometers. Ice bath and boiling water (corrected for baro-
metric pressure) are acceptable reference points.
8.1.4 Relative humidity gauge. Calibrate the relative humidity gauge
against the wet and dry bulb hygrometer method on at least two
points once each six months.
6.1.5 Analytical scale. Calibrate according to manufacturers Instruc-
tions.
32
-------
8.2 Pretest preparation.
8.2.1 Contact the householder to explain the purpose of the test,
obtain Informed consent, and set a date for the test. Attempt to
schedule the test so there will be a decent Interval between the
test and the last vacuuming by the householder, or ask them to
refrain from vacuuming prior to the test.
8.2.2 Each filter should be exposed to a light source and Inspected for
plnholes, particles, or other Imperfections. Filters with visi-
ble Imperfections should be rejected. A small brush may be used
to remove particles loosely adhering to the filters. Latex
gloves should be used when handling the filters.
8.2.3 Filters should not be numbered or folded before collection of the
sampl* If pesticides, organlcs, or metals are to be determined.
(The organic compounds and metals In the Ink could affect the
results.) This Includes filters required for blanks and trial
runs. If pesticides or organlcs are to be measured the filters
should be conditioned by being baked for two hours at 600 C to
remove organic compounds and placed In a humidity chamber free
from pesticide contamination for 24 hr where the humidity Is
maintained at SOX +/- 5%. Blanks need to be run to determine
background levels of contanlnation. The filters should be weighed
to the nearest 0.1 mg and placed in hexane-rinsed heavy aluminum
foil and double folded on each side. The Identification number
should be written on the outside of the foil and the foil placed
1n a numbered envelope.
8.2.4 Clean the aluminum nozzle, aluminum or SS cyclone, cyclone cup,
glass sample jars, SS connecting tubing, and aluminum filter
holder. A thorough cleaning with methanol, washing with deter-
gent, and rinsing with distilled water is sufficient when only
dust accumulation 1s measured. A thorough cleaning with hexane
may be sufficient for field clean-up between samples. When
pesticides or organics are to be measured the following cleaning
schedule should be followed prior to collecting a sample:
6.2.4.1 Glass should be rinsed in hexane and placed In a hot (50
C) detergent solution (Sparkleen or equivalent) for 10
m1n and scrubbed with a nylon brush. Clean with a Chem-
solv 2157 solution, rlns: with hot water, rinse with
distilled/ deionized water, and finally rinse with methy-
lene chloride. Store the glass In cleaned aluminum foil
and rinse with methanol 1f more than 24 hr passes between
cleaning and use. Methylene chloride, methanol, and hex-
ane should be applied with a Teflon squeeze bottles.
Teflon and rubber gaskets should be cleaned in the same
way that glass 1s cleaned.
8.2.4.2 SS should be cleaned the same as glass except a 15%
solution of nitric acid is substituted for Chemsolv.
8.2.4.3 Aluminum should be cleaned the same as glass except the
Chemsolv wash should be omitted.
8.2.5 Assemble the sampling train and place a filter 1n the filter
holder. Check for leaks by plugging the entrance to the cyclone,
start the blower, and place the control valve in the open posi-
tion. The system volumetric flow should be no greater than 0.5
L/s (1 cfm). If the flow Is greater than 0.5 L/s, the leaks must
be located, sealed and the sampling train rechecked.
33
-------
9. Sampling procedure
S.I Fill out test data form (see Figure 7) recording all data.
9.2 Select a rug or bare floor for sampling where small children are more
likely to play. Use approximately 2 cm (3/4 in.) masking tape and a
steel measuring tape or template to lay out eight 46 by 137 cm (18 by
54 in.) squares in an approximately 1.8 by 2.7 m (6 by 9 ft) sampling
area (see Figure 6) if there is sufficient space. The 46 by 137 cm
squares can be placed on a rug or floor In any convenient layout.
9.3 Connect the blower to 125 volt AC house electric circuit that has 10
amp of spare capacity. Make certain there are no other large loads on
the circuit such as an electrt: heater, iron, or air conditioner.
Assemble the HVS2 train. Adjust the flow rate to 9.5 L/s (20 cfm)
while adjusting the static pressure (delta P) in the nozzle to match
that recommended for the following carpets or bare floor:
Plush 28.4 cm (11.2 In.) H?0
Level loop 41.7 cm (16.4 in.) H?0
Flat 41.7 cm (16.4 in.) H?0
Bare Floor 2.5 cm (1.0 in.) H?0 (approximately 1/16 in. above
floor)
If the delta P on a plush carpet cannot be adjusted to 28.4 cm (11.2
1n.) H20, it should be adjusted to the maximum that still allows the
HVS2 to be pushed without excess effort. The nozzle delta P for the
multilevel, shag, and other types of.deep pile carpets should be ad-
justed to the maximum which still a?lows the HVS2 to be pushed at the
same speed as specified in section 9.4.
9.4 Sample each 46 by 137 cm (18 by 54 In.) segment for 80 s moving the
nozzle eight times over each section of rug and completing a forward
and return stroke of the HVS2 in 5 s. A stop watch should be used to
time the strokes. A 13 by 137 cm (5 by 54 in.) section of surface should
be sampled with eight strokes of 2 1/2 s each before moving to the next
section. When one 46 by 137 cm (18 by 54 in.) segment is sampled stop
the HVS2 and move to the next segment as shown in Figure 6, alternating
from one side of the 1.8 by 2.7 m rectangle to the other until four
segments have been completed. If It appears that less than 2 g of dust
have been collected in the cyclone catch bottle, sample additional
segments. Adjust the flow control valve to maintain a constant 9.4 L/s
(20 cfm) flow and nozzle height to maintain a constant pressure drop
across the nozzle. A portable electronic scale (0.1 to 300 g) may be
used to measure the approximate amount of dust in the cyclone jar. Run
the HVS2 blower for 20 s with the nozzle at least 3 in. above the floor
to remove loose material In the train at the end of the test. Do not
do a leak check at the end of sampling as the high vacuum may cause
volatilization of the pesticide. Record the nozzle delta P at the
beginning and end of sampling. A change of more than 10% in the delta P
Indicates a potential leak.
9.5 Sample recovery. The following procedures assume an Intent to analyse
the sample for pesticides or other trace materials. If only particle
mass accumulation is being measured, many of the Hems may be relaxed
and performed in a manner similar to the handling of ambient high
volume filter samples.
9.5.1 Open the filter holder and use a fine brush to remove any panic-
ulate matter deposited on the front side of the holder onto the
34
-------
filter. Any filter fibers adhering to the filter holder gasket
must be removed with a sharp edged blade and placed with the
filter. Fold the filter twice lengthwise with the particle side
1n. Replace the folded filter 1n the numbered aluminum foil
until 1t Is weighed. Then place the filter (In the numbered
alumlnun foil) In a numbered clean glass bottle so It Is sup-
ported and cannot move around. Record the bottle number on the
test data sheet. Place a layer of clean aluminum foil Inside the
lid so that 1t provides a seal when the lid 1s screwed on. Place
1n the Ice chest. Remove and cap the cyclone bottle using clean
aluminum foil Inside the I1d and label It with a number corres-
ponding to the number on the filter. Place 1t In the Ice chest.
9.5.2 Wash the nozzle, cyclone, filter holder, and connecting tubing
with a SOX hexane-SOX acetone rinse. Brush down the same parts of
the HVS2 with a stiff brush. Follow the brushing with another
rinse. Allow the system to air dry for 4 minutes before reassem-
bly. The small amount of sample lost to the walls (usually less
'iian 0.5X) makes analysis of the washings unnecessary.
9.S.3 Taking care to minimize the exposure of the samples to sunlight
and high ambient temperatures, selve the contents of the cyclone
catch bottle as described In ASTH method D422-63 (without
drying). Weigh the material that passes the final 150 mesh
screen and prepare 1t for chemical analysis. Place It 1n a
clean, folded aluminum package Inside a clean glass Jar and store
1t with the filter 1n a refrigerator at 0 C or 1n an 1ce chest
with dry Ice. Do not cool the filter or cyclone catch prior to
selvlng or weighing.
10. Analysis for pesticides and other organlcs
10.1 The filter, and dust from the cyclone should be extracted together.
The filter may be weighed before extraction but should not be placed
1n the humidity chamber or desiccated. The removal of water from the
filter may also remove pesticides. If the relative humidity of the
sampled air Is similar to that found In the humidity chamber (45 to
55X) there will be only a small change 1n weight In the humidity
chamber.
10.2 The results should be recorded In nanograms of each specific compound
measured along with sensitivity, precision, and accuracy of the
measurement.
11. Calculations
11.1 The concentration 1n ng/m? for each compound 1s found by dividing the
total ng of the compound In the sample by the area sampled.
11.2 The ppb for each compound Is found by dividing the ng of the compound
1n the sample by the number of grams of dust passing the final screen
and on the filter which were analyzed.
12. Precision
12.1 The percent coefficient of variation of the test method 1s about 12%
for plush carpets and 8X for level loop for particles under 150 urn.
These values for the rugs were determined by running a series of iden-
tical tests on standard test carpets using a standard test dust simi-
lar to that specified in ASTM Method F608-70.
35
-------
(9ft)
2.7m
(54")
-137cm'
(6ft)
1.8m
B1
A2
B3
A4
A1
B2
A3
B4
(18")
46cm
Figure 6. Ideal arrangement of test segments In 1.8 by 2.7 m (6 by 9
ft) area. Sample either A or B segments.
36
-------
HIGH VOLUME SURFACE SAMPLER DATA SHEE1
Operator
Date
Sample Ident. No.
Sampling site
.Rug
Hard floor Other
Type of surface
Type of rug Plush Level loop Flat Multilevel Shag
Last vacuumed Temp Humidity % Comments
Location of area sampled Area
Sketch of area sampled:
Total sample time
Filter final wt
Cyclone final wt
Sieve size
Wt sieving + filter
Filter sample no.
Lab sample no.
Name of compound
mln sec Flow rate
_ g Filter tare wt
_ g Cyclone tare wt
Wt passed sieve
wt
Nozzle dP
g Net wt g
g Net wt g
g % passed sieve %
Accumulation wt g/m2
Cyclone sample no.
cone (ng/m?)
cone (ppm)
Figure 7. Recommended data sheet.
37
-------
SECTION 8
REFERENCES
Amer. Soc. Testing and Materials (ASTM). 1986. Standard recommended prac-
tice for pesticides and polychlorinated blphenyls 1n Indoor air, draft
92.
Amer. Soc. Testing and Materials (ASTM). 1987a. Evaluation of carpet-
embedded dirt removal effectiveness of household vacuum cleaners (F608-
79). In: Annual Book rf ASTM Standards. Amer. Soc. Testing and Mater-
ials, Vol. 15.07: Philadelphia, 15 pp.
Amer. Soc. Testing and Materials (ASTM). 1987b. Particle size analysis of
soils (D422-63). in: Annual Book of ASTM Standards. Amer. Soc. Test-
Ing and Material?., Vol. 15.07: Philadelphia, 12 pp.
Amer. Soc. Testing *nd Materials (ASTM). 198/c. High volirne sampling for
solid participate matter and determination of particm ate emissions
(D4536-86). In: Annual Book of ASTM Standards, Amer. Soc. Testing and
Materials: Philadelphia, In press.
Angle, C.R. and M.S. Mclntire. 1979. Environmental lead and children: The
Omaha ftudy. J. Tox. Environ. Health 5:855-870
Annis, J.C. and P.J. Annis. 1972. Vacuum Cleaning: a Source of Residential
Airborne Dust. In: ASHRAE Symposium for leaner Indoor Air: Progress
and Problems. ASHRAE: Cincinnati, 1972
Anzai, I. and T. Kikuchi. 1978. A new monitoring technique of surface
contamination: The test surface methoM. Health Physics 34:271-273
Bailey, G. 1981. Air quality and respiratory health: A comparison study of
two Seattle communities. Submitt'.d as a Master's thesis at the Univer-
sity of Washington, Seattle, Was1..
Bamesberger, W.L. and D.F. Adams. 1966. An atmospheric survey for aerosol
and gaseous 2,4-D compounds. In: Organic Pesticides In the Environ-
ment, A.A. Rosen and H.F. Kraybill, eds. Washington, D.C.: Amer. Chem.
Soc., pp. 219
Baragar, A. 1961. A procedure for measuring the rug cleaning ability of
vacuum cleaners. Univ. of Nebraska Agricultural Experiment Station,
Res. Bull. 195: Llncrin, Neb., 13 pp.
38
-------
Binder, S., D. Sokal, D. Haughn. 1986. Estimating the amount of soil Inges-
ted by young children through tracer elements. In: East Helena, Mon-
tana Child Lead Study, Summer, 1983. Montana Dept. Health and Envir.
Sciences: Helena, Mont., App. 23
Boubel, R.W. 1971. A high volume stack sampler. J. Air Poll. Control
Assn. 21:783-787
Brown, K.W., K.C. Donnelly, J.C. Thomas, P. Davol, B.R. Scott. 1985. Muta-
genlclty of three agricultural soils. Scl. Total Environ. 41:173-186
CAMRASO. 1983. The Economic Cleaning of Soft Floorcoverings. CAMRASO:
Leeds, U.K., 59 pp.
Davidson, C.L. and R.W. Ellas. 1986. Environmental concentrations and
potential pathways to human exposure. In: A1r Quality Criteria for
Lead. U.S. Environ. Prot. Agency: Research Triangle Park, N.C., Chap-
ter 7
Davles, J.E., W.F. Edmundson, A. Raffonelll. 1975. Role of house dust 1n
hum&n ODT pollution. AIR. J. Public Health 65:53-57
Dlemel, J.A.L., B. Brunekreef, J.S.M. Bolelj, K. Biersteker, S.J. Veenstra.
1981. The Arhnem Lead Svudy; II. Indoor Pollution, and Indoor/Outdoor
Relationships. Environ. Res. 25:449-456
Dugan, M.J. and M.J. Insklp. 1985. Childhood exposure to lead In surface
dust and soil: A community health problem. Public Health Rev. 13:1-
54
Dugan, M.J., M.J. Insklp, S.A. Rundle, J.S. Moorcroft. 1985. Lead In Play-
ground Dust and on the Hands of Schoolchildren. Scl. Total Environ.
44:65-79
Fox, N.R.S. 1979. Nutritional Influences on metal toxlclty: Cadmium as a
model toxic element. Environ. Health Per. 29:95-104
Gillette, D.A., J. Adams, A. Endo, D. Smith, R. Kihl. 1980. Threshold
velocities for Input of soil particles Into the air by desert soils.
J. Geophys. Res. 85C:5621-5630
Haeske, H., H. Krupp, 6. Sandstede, G. Halter. 1957. Eighth progress
report on a study of the scientific principles of household washing and
cleaning to the Hoover Company. Battelle-Instltut E.V.: Frankfurt,
F.D.R., pp. 7-22
Harrison, R.M. 1979. Toxic metals in street and household dusts. Scl.
Total Environ. 11:89-97
Hawley, John K. 1985. Assessment of health risk from exposure to contami-
nated soil. Risk Analysis 5:289-301
39
-------
Hunt, C M. 1972. Simple observations of some common Indoor activities as
producers of airborne participates. In: ASHRAE Symposium for Cleaner
Indoor Air: Progress and Problems. ASHRAE: Cincinnati, 1972
Imhoff, R.E., J.A. Hanntng, w.M. Cooke, T.H. Hayes. 1982. Preliminary
report on a study of the ambient Impact of residential wood combustion
in Petersville, Alabama. In: Solid Fuels; Environmental Impacts and
Solutions. 0. Cooper and D. Halek, eds. Oregon Graduate Center:
Beaverton, Or., pp. 520-538
International Programme of Chemical Safety (IPCS). 1986. Principles for
Evaluating Health Risks from Chemicals during Infancy and Early Child-
hood: The Need for a Special Approach. Environmental Health Criteria
No. 59. World Health Organization: Geneva, pp. 40-47
*.Krause, I., H. Englert, P. Dube. 1987. Petachlorophenol containing wood
preservatives: Analyses and evaluation. In: Proceedings; Indoor Air
'87, Vol. I. B.Seifert, H.Esdorn, H. Fischer, H. Ruden, J. Wegner,
eds. Inst. for Hater, Soil and Air Hygiene: Berlin, pp. 220-224
LaGoy, P.K. 1987. Estimated Soil Ingestion rates for use in risk asses-
sment. Risk Analysis 7:335-359
Lewis, R.G. 1976. Sampling and analysis of airborne pesticides. In: Air
Pollution from Pesticides and Agricultural Processes, R.E. Lee, ed.
CRC Press: Boca Raton, Fl., pp. 51-94
Lewis, R.G. and H.O. Jackson. 1982. Modification and evaluation of a high
volume air sampler for pesticides and other semivolatile industrial
organic chemicals. Analytical Chem. 54:592-594
Lewis, R.G., A.R. Brown, M.O. Jackson. 1977. Evaluation of polyurethane
foam fo' sampling of pesticides, polychlorinated biphenyls, and poly-
chlorinaied naphthalenes in ambient air. Analytical Chem. 49:1668-1672
Lioy, P.J., N. Avdenko, R. Harkov, T. Atherholt, J.M. Daisey. 1985. A
pilot Indoor-outdoor study of organic particulate matter and particu-
late mutagenicity. J. Air Poll. Cont. Assn. 35: 653-657
Hahaffey, K.R. and J.L. Annest. 1985. Association of blood level, hand-to-
mouth activity, and pica among yound children. Federation Proceedings
44:752
Hilar, C.R. and P. Mushak. 1982. Lead contaminated housedust: Hazard,
measurement and decontamination. In: Lead Absorption in Children;
Hanagement, Clinical, and Environmental Aspects. J.J. Chisolm and D.H.
O'Hara, eds. Urban and Schwarzenberg: Baltimore, pp. 143-152
Murphy, B.L., and J.E. Yocom. 1986. Migration factors for particulates
entering the Indoor environment. Presented to the annual meeting of
the Air Pollution Control Assn., Minneapolis, Minn. (86-7.2)
Natusch, D.F.S., and J.R. Wallace. 1974. Urban aerosol toxicity: The in-
fluence of particle size. Science 186:695-699
40
-------
Occupational Safety and Health Admin. (OSHA). 1977. Wipe Sampling Policies
and Procedures. In: Industrial Hygiene Manual. U.S. Oept. of Labor.:
Washington, Chapter VI
Ott, V., L. Wallace, 0. Hage, G. Akland, R. Lewis, H. Sauls, C. Rodes, D.
Kleffman, 0. Kuroda, K. Norehouse. 1986. The Environmental Protection
Agency's research program on total human exposure. Environ. Intl.
12:475-494
Pinkerton, C., 0. Hammer, K.B. Ridbord, J.P. Creason, J.L. Kent, G.K.
Murphy. 1973. Human milk as a dietary source of cadmium and lead. In:
Trace Substances in Environmental Health. B.D. Hemphill, ed. Univ. of
Missouri: Columbia, Mo.
*.-Que-Hee, S.S., B. Peace, C.S. Clark, J.R. Boyle, R.L. Bornschein, P.B.
Hammond. 1985. Evolution of efficient methods to sample lead sources,
such as house dust and hand dust, In the homes of children. Environ.
Res. 38:75-95
Rabinowitz, M., A. Leviton, H. Needleman, D. Bellinger, C. Waternaux. 1985.
Environmental correlates of infant blood lead levels in Boston. Envi-
ron. Res. 38:96-107
Rappaport, S.M., M.C. McCartney, E.T. Wei. 1979. Volatization of mutagens
from beef during cooking. Cancer Ltrs. 8:139-145
Roberts, J.W., J. Broto, M.6. Ruby, G.R. Warren. 1985. An exploratory
study of toxics in house dust. Presented to the Annual Meeting of
PNWIS-CPANS APCA, Calgary, Alberta
Roberts, J.W., M.G. Ruby, 6.R. Warren. 1987. Mutagenic activity of house
dust. In: Short-term Bioassays in the Analysis of Complex Environmental
Mixtures, S.S. Sandhu, O.H. OeMarini, H.J. Mass, H.M. Moore, J.L.
Mumford, eds. Plenum Press, New York, pp. 355-367
Roels, H.A., J.P. Buchet, R. Lauwerys, P. Bruaux, F. Claeys-Thoreau, A.
Lafontafne, 6. Verduyn. 1980. Exposure to lead by oral and pulmonary
routes of children living in the vicinity of a primary lead smelter.
Environ. Research 22:81-94
Royster, G.W. and B.R. Fish. 1965. Techniques for assessing removable
furface contamination. Oak Ridge National Lab.: Oak Ridge, Tenn.
(ORNL-TM-1045)
Schaefer, V.J., V.A. Mohnen, V.R. Veirs. 1972. A1r quality of American
hones. Science 175:173-176
Seifert, 8., M. Drews, K. Aurand. 1984. Indoor heavy metal exposure of the
population around a secondary lead smelter. In: Indoor air, Vol 2. B.
Berglund, T. Lindvall, J. Sundell, eds. Proceedings of the 3rd Interna-
tional Conference on Indoor Air Quality and Climate, Stockholm, Sweden,
Aug. 20-24, 1984. Swedish Cncl. for BIdg. Res.: Stockholm, Sweden, pp.
177-181
41
-------
Sherma, J. 1981. Manual of Analytical Quality Control for Pesticides and
Related Compounds. U.S. Environ. Prot. Agency (EPA 600/2-81-059)
Sinclair, J.D. 1982. Paper extraction for sampling Inorganic salts on
surfaces. Analyt. Chem. 54:1529-1533.
Solomon, R.L. and J.W. Hartford. 1976. Lead and cadmium In dusts and soils
In a small urban community. Environ. Sc1. & Tech. 10:773-777
Starr, H.G., F.D. AldMch, W.D. McDougall, L.H. Hounce. 1974. Contribution
of household dust to the human exposure to pesticides. Pesticide
Monitoring J. 8:209-211
U.S. Environmental Protection Agency. 1984. Method for the Determination
of Organochlorine Pesticides and Polychlorinated Biphenyls In Ambient
Air. In: Compendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air. (EPA 600/4-84-041)
van Houdt, J.J. and J S.M. Bolelj. 1984. Mutagenlc activity of Indoor
airborne particles compared to outdoors. In: Indoor air, Vol 2. B.
Berglund, T. Undvall, J. Sundell, eds. Procedlngs of the 3rd Interna-
tional Conference on Indoor Air Quality and Climate, Stockholm, Sweden,
Aug. 20-24, 1984. Swedish Cncl. for Bldg. Res.: Stockholm, Sweden, pp.
169-176
Vostal, J.J., E. Taves, J.W. Sayre, E. Charney. 1974. Lead analysis of
house dust: A method for the detection of another source of lead expo-
sure in Inner city children. Environ. Health Perspec. 7:91-97
Weschler, C.J. 1978. Characterization techniques applied to Indoor dust.
Environ. Scl. 4 Tech. 12:923-926
Wester, R.C. and H.I Naibach. 1987. In vivo percutaneous absorption. In:
Dermatotoxlcology. F.N. Marzulll and H.I. Mai bach, eds. Hemisphere
Publishing: New York, pp. 135-52
Wright, C.G. and R.B. Leldy. 1982. Chloradane and heptachlor In the am-
bient air of houses treated for termites. Bull, of Environ. Contam.
and Toxicology 28:617-23
Yocom, J.E., W.L. Clink, W.A. Cote. 1971. Indoor/outdoor air quality
relationships. J. Air Poll. Control Assn. 21:251-259
42
-------
Reproduced by NTIS
National Technical Information Service
U.S. Department of Commerce
Springfield, VA 22161
This report was printed specifically for your
order from our collection of more than 2 million
technical reports.
•
For economy and efficiency, NTIS does not maintain stock of its vast
collection of technical reports. Rather, most documents are printed for
each order. Your copy is the best possible reproduction available from
our master archive. If you have any questions concerning this document
or any order you placed with NTIS, please call our Customer Services
Department at (703)487-4660.
Always think of NTIS when you want:
• Access to the technical, scientific, and engineering results generated
by the ongoing multibillion dollar R&D program of the U.S. Government.
• R&D results from Japan, West Germany, Great Britain, and some 20
other countries, most of it reported in English.
NTIS also operates two centers that can provide you with valuable
information:
• The Federal Computer Products Center - offers software and
datafiles produced by Federal agencies.
• The Center for the Utilization of Federal Technology - gives you
access to the best of Federal technologies and laboratory resources.
For more information about NTIS, send for our FREE NTIS Products
and Services Catalog which describes how you can access this U.S. and
foreign Government technology. Call (703)487-4650 or send this
sheet to NTIS, U.S. Department of Commerce, Springfield, VA 22161.
Ask for catalog, PR-827.
Name
Address.
Telephone
- Your Source to U.S. and Foreign Government
Research and Technology.
-------