EPA/60G/Ar93/199
COMPARISON OF TRANSFER OF SURFACE CHLORPYRIFOS
RESroUES FROM CARPET BY THREE DISLODGEABLE
RESIDUE METHODS
David E. Camann, H. Jac Harding, Susan R. AgrawaJ
Southwest Research Institute
P.O. Drawer 28510
San Antonio, Texas 78228-0510
Robert G. Lewis
U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711
ABSTRACT
After chlorpyrifos which was broadcast-sprayed on carpet had dried, transfers by the Dow drag sled, the
California cloth roller, and the Southwest Research Institute polyurethane foam (PUF) roller were compared. On
plush nylon carpet, mean chlorpyrifos transfers were 4.5% by the cloth roller, 1.1% by the drag sled, and 0.65%
by the PUF roller. On level-loop polypropylene carpet, mean transfers were 2.5% by the cloth roller, 1.4% by
the drag sled, and 1.2% by the PUF roller. The cloth roller was found to be less suitable than the other methods
because its transfers exhibited greater variability and were altered by orientation of the roll relative to the lay of
the carpet fibers. Moistening the sampling media increased the transfer by the drag sled and the PUF roller, but
substantially increased the measurement variability of both methods.
INTRODUCTION
Dermal contact with residues of pesticides applied to carpets and subsequent skin absorption or ingestion
through hand-to-mouth activity are routes of human exposure which need better evaluation, especially for young
childrea The Dow drag sled1, the California cloth roller2, and the Southwest Research Institute (SwRI)
polyurethane foam roller3 (SwRI invention disclosure #2061, patent pending) are dislodgeable residue sampling
methods which have recently been developed to estimate the transfer of a chemical from a contaminated surface
to the skin. This paper reports three experiments conducted to determine which of the three methods deployed
as currently used by the developer provides the more reproducible and facile transfer of chlorpyrifos residues from
carpet:
Exp. 1. Transfer comparison of the three methods using dry sampling media on new plush cut-pile
nylon carpet.
Exp. 2. Transfer comparison of the three methods using dry sampling media on new level-loop
polypropylene carpet.
Exp. 3. Transfer comparison of the better two methods using both dry and moist sampling media on
new plush cut-pile nylon carpet.
METHODS
Prior to each experiment, the specified virgin carpet and padding were installed in 1 }/2 rooms (1 room
for Experiment 3) of an existing 42 ft x 10 ft 3-room trailer.
-------
DisJodgeable Residue Methods
Relevant characteristics of the dislodgeable residue methods are summarized and contrasted in Table I.
The California cloth roller was constructed and the method performed as described by Ross et al. (1991). A soap-
washed and precleaned dry 17 in. x 17 in, cloth of percale bedsheet was placed on the carpet and covered with
a sheet of plastic. A 2 ft long by 4 in. diameter sewer pipe, filled with 25 pounds of lead shot ballast and wrapped
in a sheet of high density PUF, was rolled forward and backward over the plastic/cloth/earpet sandwich ten times.
After the 20 passes, the percale cloth was picked up and analyzed.
The drag sled method was performed using the initial configuration described by Vaccaro and Cranston
(1990). Briefly, a precleaned dry 4 in, x 4 in. denim weave cloth supplied by B. Shurdut, Dow Chemical
Company, was attached beneath foil under a 3 in. x 3 in. plywood block and an 8-lb weight mounted (Figure 1).
The sled was dragged once over a 3 in. x 4 ft carpet strip at 6-8 cm/s. After the single pass, the denim cloth was
removed for analysis.
The original PUF roller sampler (Hsu et al., 1990) was used for Experiments 1 and 2. A precleaned dry
PUF ring (3 in. length, 3.5 in. OD, 1.62 in. ID) was secured on the 8 in. length x 2 in. OD cylindrical 7.2 Ib
stainless steel roller. The new (October 1992) model of the PUF roller sampler was used for Experiment 3. A
precleaned dry PUF ring was secured on the 3 in. length x 1.75 in. OD cylindrical 0.37 kg aluminum roller
(Figure 2). The PUF roller was rolled once over a 3 in. x 1.0 m carpet strip at 10 cm/s once in both directions.
After the two passes, the PUF ring was slit and removed from the roller for analysis.
In Experiment 3, PUF rings and denim cloth were used which had been moistened with deionized water.
A precleaned PUF ring was uniformly moistened with 5.0 ± 0.1 g of water in the laboratory by spraying the ring
surface with an atomizer, compressing with a squeeze tool to obtain uniform water distribution, weighing and
sealing in a steel canister until use. The sampling surface of the denim cloth was moistened with 0.5 ± 0.1 g of
water from the atomizer and weighed just prior to mounting under the drag sled. When moistened at these levels,
the PUF ring and denim cloth were observed to produce equivalent moisture trails at method pressure on a glass
surface.
Broadcast Application of Chlorpyrifos and Ventilation While Drying
Broadcast application of Chlorpyrifos to test carpeting was conducted by a licensed pest control applicator
according to label instructions for flea control treatment. The formulated product, Dursban® L.O. (E.P.A.
Registration No. 464-571), which contains 41.5% Chlorpyrifos (0,O-diethyl O-(3,5,6-trichloro-2-pyridyl)phosphoro-
thioate), was applied approximately 40 cm above the carpet as a 0.50% aqueous spray (40 mL/3.785 L water) with
a hand-held fan broadcast nozzle attached to an air pressurized tank. Application was accomplished in 2 to 3 min.
The trailer was ventilated for 2 h immediately after application. All windows were opened and window
air conditioning units were operated in fresh return air mode. During the first 30 min and last 15 min of the
ventilation period, both doors were opened and a box fan was operated outside the test room doorway to allow
maximum cross ventilation. Air conditioner units were returned to the usual reeirculated air mode just prior to
sampling and remained on throughout the sampling period of Experiments 1 and 2.
Experimental Design
Adjacent samples using each compared dislodgeable residue method (with dry and moist media in
Experiment 3) and a deposition coupon (2 coupons in Experiment 3) were collected sequentially within a
rectangular block of treated carpet. Six replicate blocks were sampled in Experiments 1 and 2; 4 l/2 replicate
blocks comprised Experiment 3.
-------
Table I. Characteristics of dislodgeablc residue methods.
Property
California cloth roller
Dow drag sled
SwRI PUF roller
Sampling medium (material) Percale bcdshcet (50% cotton,
50% polyester)
Surface of sampling medium Square (42.9 cm)2
Contact motion
Face (iastantaneous contact
area pressed through sampling
medium)
Roll
440 cm2 = 42.9 cm x 10.2 cm
Mass exerting pressure through 14.4 kg
sampling medium
2,300 Pa = (14.4 kg)(9.8 m/s2)/
[(0.61 m)(0.10m)l
Pressure exerted through
sampling medium
Sampled carpel area
0.184 m2 = (0.429m)2
Number of passes over sampled 20
carpet area
Sampling speed over carpel 0.23 m/s
Denim weave cloth
(predominantly cotton)
Square (10.2 cm)2
Drag
58 cm2 = (7.6 cm)2
3.53 kg
Polyurethanc foam ring (polycther,
0.029 g/cm3)
Curved exterior of ring, (OD = 8,9 cm,
length = 7.6 cm)
Roll
38.6 cm = 7.6 cm x 5.1 cm
3.25 kg;" 3.10 kgb
6,000 Pa = (3.53 kg)(9.8 m/s2)/ 8,300 Pa;a 8,000 Pab = (3.10 kg)
(0.076 m)2 (9.8 m/s2)/f(0.076 m)(0.05 m)J
0.093 m2 = 0.076 m x 1.22 m 0.076 m2 = 0.076 m x 1.0 m
1
0.07 m/s
0.10 m/s
a Original PUF roller sampler
b 1992 model of PUF roller sampler
-------
Weight
ooden Block
Cotton Denim
Collection Media
Figure 1. Dow drag sled.
Stalnlwn Stwil
Wtlghi
I
Snap-On
PUFRing
I
Figure 2. PUF roller sampling instrument
-------
Deposition coupons, consisting of absorbent alpha-cellulose pads (4 in. x 4 in.) backed with aluminum
foil were placed on the carpet prior to the chlorpyrifos application and picked up before the adjacent dislodgeable
residue samples from the block were collected. Residues measured on the coupon (pair in Experiment 3) gave
an estimate of the surface loading of residue remaining on adjacent carpeting during sampling in the block.
Field blanks of each method were obtained by sampling on the virgin carpet prior to the chlorpyrifos
application to assess contamination potential during sampling and handling. Deposition coupon(s) were placed
at the designated locations in each sampling block shortly before the application commenced. Field samples were
collected in each blocks upon label allowed re-entry, i.e., when the carpet was dry (which was operationally
defined as 2 hours after application, but checked by hand contact). The dislodgeable residue samples of a block
were collected from specified randomized locations in the block after the deposition coupon(s) were picked up.
All samples were collected in one block before proceeding to the next block. Spikes of the precleaned
dislodgeable residue media (both dry and moist in Experiment 3) and of a deposition coupon were made both
before and after the replicate block sample sets were collected; these field spikes were used to assess and adjust
for losses during transport, storage, and extraction.
Sample Analysis
All samples were Soxhlet-extracted with 6% ethyl ether/94% hexane; extraction commenced within 24
hours after sampling. In Experiment 3, the pair of deposition coupons from a block were extracted together.
Extracts were analyzed for chlorpyrifos by GC/ECD on two dissimilar columns and quantitated from the DB-5
column results.
Data Adjustment
Crude results (mg/sample) from each field sample were adjusted for contamination and extraction
inefficiency by subtracting the field blank result and dividing the difference by the mean recovery proportion of
the two field spikes for that method. The adjusted result was divided by the sampled carpet area (see Table I)
to determine the measured transfer rate (mg/m2 of carpet contacted) for dislodgeable residue samples and the
measured surface loading (mg/m2) for coupon samples.
RESULTS
Recovery of chlorpyrifos in field spikes of all sampling media is shown in Table n. Amounts in field
blanks were negligible in Experiments 1 and 3, but within two-to-three orders of magnitude of field samples in
Experiment 2, which was performed six days after Experiment 1.
Table II. Percent recovery of chlorpyrifos in field spikes of sampling media.
Sampling medium Use
Field spikes, Standard
Moisture n Mean deviation
a-Cellulose pad
Denim cloth
Percale sheet
PUFring
Deposition coupon Dry
Drag sled
Cloth roller
PUF roller
Dry
Moist
Dry
Dry
Moist
5
2
4
2
94.2
102.8
96.6
96.2
110.2
107.7
13.9
11.0
1.6
14.0
4.6
2.1
-------
The surface loading of chlorpyrifos on treated carpets determined at the time of block sampling via
deposition coupons is presented in Table ED. The chlorpyrifos loading was slightly lower and much more uniform
on the level-loop polypropylene carpet in Experiment 2 than on the plush cut-pile nylon carpets in Experiments 1
and 3.
Table HI. Chlorpyrifos surface loading4 on treated carpets during dislodgeable residue sampling5.
Experi-
ment:
Date
1:8/4
2:8/10
3:11/24
Carpet type
Plush nylon
Level-loop
polypropylene
Plush nylon
Deposition
coupons
per
replicate
1
1
2
No. of
replicates
n
6
6
5
Chlorpyrifos loading8 (mg/m2)
Mean
154.1
120.9
183.3
Standard
deviation
94.3
14.0
30.6
Range
74-329
105-135
146-216
Loading
coefficient
of variation
0.611
0.116
0.167
a Loading, mg/m2 = [(deposition amount, mg - blank amount, mg)/(mean spike recovery)]/(coupon area,
m2)
The transfer rates of chlorpyrifos from treated carpet by the dislodgeable residue methods are summarized
in Table IV. Transfers with dry sampling media were highest for the cloth roller, intermediate for the drag sled,
and lowest for the PUF roller, both from plush nylon carpet and from level-loop polypropylene carpet As shown
by the coefficient of variation, all three dislodgeable residue methods gave more repeatable performance on the
plush nylon carpet than on the level-loop polypropylene carpet. The cloth roller displayed more variation in
chlorpyrifos residue transfer using dry sampling media than did the drag sled and PUF roller, particularly on the
plush nylon carpet. Transfers obtained with the cloth roller differed substantially [8.1 ± 1.5 (n=4) vs. 4.5 ± 0.1
(n=2) for Experiment 1 mean ± std. dev.] for rolls oriented with/against vs. across the lay of the carpet fibers.
Transfers with the drag sled and PUF roller did not vary with the orientation of the drag/roll relative to the lay
of the carpet fibers. The additional transfer variability observed with the cloth roller is largely attributable to the
directional sampling effect.
Transfers using moist media were larger than transfers using dry media for both the drag sled and the PUF
roller. However the measurement variability of both methods increased substantially when moist media were used.
The percentage of the chlorpyrifos loading that was transferred by each method in each experiment is
presented in Table V. The mean percent transfer of chlorpyrifos residue with the cloth roller on level-loop
polypropylene carpet in Experiment 2 (2.5%) was only 55% of its transfer (4.5%) on plush cut-pile nylon carpet
in Experiment 1. In contrast, percent transfers with the drag sled and PUF roller were slightly larger on the level-
loop polypropylene carpet.
DISCUSSION
This research was performed to allow intercomparison of dislodgeable residue transfers obtained by
recently-developed methods in different studies conducted to support registration of pesticides used in the home.
Consequently the cloth roller, drag sled, and PUF roller methods were evaluated as performed by their developers
(Table I), despite differences in properties which are likely to affect transfer including sampling pressure and
speed, sampled carpet area, and number of passes over the area. The methods were compared upon label-allowed
re-entry after application of the pesticide product at the maximum permitted rate, as in pesticide registration
studies.
Serious sampling difficulties were experienced in use of the cloth roller method, but not the other methods
(Table VI). In particular, the sampling cloth tended to bind and shift from its original position with successive
-------
passes of the roller, so that the actual carpet area contacted by the cloth differed in magnitude and location from
the nominal sampled area in an unknown and uncontrollable manner. The variation in transfer with orientation
of the passes relative to the lay of the carpet fibers, which was observed only in use of the cloth roller, may relate
to the propensity of the cloth to bind and shift during sampling. In situations where the carpet has a distinct lay
of pile, especially in areas of traffic, the direction of the drag sled may also affect dislodgeable transfer.
Performance of dislodgeable residue sampling using moistened media may provide a more realistic
simulation of transfers experienced by young children who exhibit frequent hand-to-mouth behavior. The goal
for moistness of the sampling media is problematic, since the probability distribution of moisture on children's
hands is broad and decreases with age. Moistening does increase dislodgeable residue transfer, as expected.
However, moistening so exacerbates transfer measurement variability that many more measurements are needed
to obtain the precision in mean transfer achieved with dry sampling media.
Mean percent chlorpyrifos transfer obtained from plush cut-pile nylon carpeting declined markedly from
Experiment 1 to Experiment 3, both with the drag sled (1.12% vs. 0.44%) and the PUF roller (0.71% vs. 0.26%).
Through a multifactional experiment, Dow Chemical Co. has indications that drag sled transfer increases with
room air temperature (B. Shurdut, personal communication, 1992). The temperature of both the ambient air
(Table V) and the room air was higher during sampling in Experiment 1 (performed in August) than in
Experiment 3 (in November).
Additional experiments are underway to determine the recovery of residues from the human hand with a
handwipe method, and to determine the effect of sampling variables including air temperature on transfers with
the drag sled and PUF roller. Future experiments will compare transfers by human hand presses to the
dislodgeable residue methods.
Table IV. Comparison of dislodgeable residue method transfer" of chlorpyrifos residue by carpet type and
media moisture.
Dislodge- Sampling
Experi- able residue media
mem Carpet type method moistness
1 Plush nylon Cloth roller
Drag sled
PUF roller
2 Level-loop Cloth roller
polypropylene Dfag s,ed
PUF roller
Dry
Dry
Dry
Dry
Dry
Dry
No. of
replicates
n
6
6
6
6
6
6
Chlorpyrifos transfer8
(mg/m2) f
Standard
Mean deviation
6.92
1.73
1.10
3.00
1.66
1.43
2.20
0.29
0.26
1.34
0.70
0.56
Range
4.4-10.0
1.4-2.2
0.8-1.5
1.6^.9
1.1-2.9
0.9-2.3
Transfer
coefficient
of
variation
0.318
0.168
0.238
0.446
0.420
0.393
3
a
Plush nylon
Transfer, mg/m2 =
Drag
PUF
sled
roller
[(dislodged amount,
m2)
Dry
Moist
Dry
Moist
mg - blank
4
5
4
5
amount,
0.81
1.36
0.48
3.91
mg)/(mean
0.34
0.98
0.07
2.02
0.5-1.3
0.6-2.9
0.4-0.6
0.6-5.4
0.414
0.718
0.141
0.517
spike recovery)]/(sampled area,
-------
Table V. Percent8 of ehlorpyrifos loading transferred by dislodgeable residue method and sampling
conditions.
Ambient air range
during sampling
Carpet type
Plush nylon
Plush nylon
Level-loop
polypropylene
Moistness
of samp-
Experiment ling media
3 Moist
Dry
1 Dry
2 Dry
Temp
CO
19-21
28-32
29-32
Relative
humidity
(%)
24-37
55-67
50-67
Percent transfer* (mean
Cloth
roller Drag sled
0.74±0.53
0.44±0.18
4.49±1.42 1.12±0.19
2.48±U1 1.37±Q.S8
± std dev)
PUF roUer
2.13±1.10
0.26±0.04
0.71±0.17
1.18±0.46
a Percent transfer = 100 x (dislodgeable residue method transfer, mg/nr)/(surfaee loading, mg/m2)
Table VI. Observations from field use of dislodgeable residue methods.
Strengths
California Cloth Roller
• Simple in design
• Inexpensive to build from available
materials
Dow Drag Sled
• Simple in design
• Inexpensive to build from available
materials
• Simple to use
SwRI PUF Roller
• Consistent use across operators due to
few variables
• Relatively simple to use
• Foam roller contact is more like skin
contact
Weaknesses
Sampling cloth tends to bind and shift from
original position
Plastic bag cover may adhere to PUF sleeve
on roller from static
Difficult to operate due to mass of roller
Operator must contact treated surface
Susceptible to added pressure from operator
Transfer affected by roll orientation relative
to lay of carpet fibers
Drag contact unlike most skin contact with
carpel
Drag contact is potentially directional relative
to lay of carpet fibers
Expensive to build or purchase
-------
CONCLUSIONS
1. The California cloth roller is less practical and more variable than the drag sled and PUF roller methods.
2. Transfers as currently performed by the developer were lowest for the PUF roller, intermediate for the
drag sled, and highest for the cloth roller when using dry sampling media on two types of carpet.
3. An interaction in transfer percentage exists between carpet type and dislodgeable residue method.
Transfers with the cloth roller are not predictive of transfers with the other methods across different types
of carpet.
4. Transfers with moist media are larger and much more variable than transfers with dry media for both the
PUF roller and the drag sled.
ACKNOWLEDGEMENT
This research was funded by the U.S. Environmental Protection Agency (Contract 68-DO-0007) under
subcontract from Battelle. This paper has been submitted to EPA's peer and administrative review, but no official
endorsement should be inferred.
REFERENCES
1. J. R. Vaccaro and R. J. Cranston. "Evaluation of dislodgeable residues and absorbed doses of chlorpyrifos
following indoor broadcast applications of chlorpyrifos-based emulsifiable concentrate", Internal Report,
Dow Chemical Co., Midland, Ml, 1990.
2. J. Ross, H. R. Fong, T. Thongsinthusak, et al. "Measured potential dermal transfer of surface pesticide
residue generated from indoor fogger use: using the CDFA roller method," Chemosphere 22: 297-84
(1991).
3. J. P. Hsu, D. E. Camann, H. J. Schattenberg, et al. "New dermal exposure sampling technique," in:
Proceedings of the 1990 U.S. EPA/A&WMA International Symposium on Measurement of Toxic and
Related Air Pollutants. VIP-17, Air and Waste Management Association. Pittsburgh, PA, 1990. pp 489-97.
-------
KEYWORDS
dislodgeable residue
surface sampling
carpet
chloipyrifos
dermal exposure
pesticides
10
-------
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-93/299
2.
4. TITLE AND SUBTITLE
COMPARISON OF TRANSFER OF SURFACE CHLORPYRIFOS
RESIDUES FROM CARPET BY THREE DISLODGEABLE RESIDUE
METHODS
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D.E.Camann, H.J.Harding, S.R.Agrawal, R.G.Lewis
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southwest Research Institute
San Antonio, Texas
10. PROGRAM ELEMENT NO.
EO 599
11. CONTRACT/GRANT NO.
68-DO-0007
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Research Triangle Park, NC
13.TYPE OF REPORT AND PERIOD COVERED
Symposium paper
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
After chlorpyrifos, which was broadcast-sprayed on carpet, had dried,
transfers by the Dow drag sled, the California cloth roller, and the Southwest
Research Institute polyurethane foam (PUF) roller were compared. On plush nylon
carpet, mean chlorpyrifos transfers were 4.5% by the cloth roller, 1.1% by the drag
sled, and 0.65% by the PUF roller. On level-loop polypropylene carpet, mean
transfers were 2.5% by the cloth roller, 1.4% by the drag sled, and 1.2% by the PUF
roller. The cloth roller was found to be less suitable than the other methods
because its transfers exhibited greater variability and were altered by orientation
of the roll relative to the lay of the carpet fibers. Moistening the sampling
media increased the transfer by the drag sled and the PUF roller, but substantially
incresed the measurement variability of both methods.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED TERMS
c.COSATI
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
21.NO. OF PAGES
20. SECURITY CLASS (This Panel
Unclassified
22. PRICE
------- |