NEIC
AUG 211981
EPA 330/1-90-001
SAMPLING OF COMMON PESTICIDES AND PCBS
FROM INERT SURFACES
October 1989
National Enforcement Investigations Center, Denver
G.S. Environmental Protection Agency
Office of Enforcement
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF ENFORCEMENT AND COMPLIANCE MONITORING
EPA33011-90-001 (\l)6 2 1 l98l,
SAMPLING OF COMMON PESTICIDES AND PCBS
FROM INERT SURFACES
October 1989
Bryan L. Can-
Dean F. Hill
NATIONAL ENFORCEMENT INVESTIGATIONS CENTER
Denver, Colorado
-------
ABSTRACT
There is a little published data on the sampling and analysis of
common pesticides and polychlorinated biphenyls (PCBs) on inert surfaces,
particularly with respect to quantitative recovery measurements.
Consequently, various sampling parameters and the nature of different
inert surfaces were investigated for technical chlordane, chlorpyrifos,
malathion, diazinon, propoxur, bendiocarb, and a PCB mixture (Aroclor
1260). The goal of the study was to correlate various solvents, sorbent
materials, and sampling techniques with respect to recovery and
variability, as well as to determine the potential for analytical interferences
deriving from various types of inert surfaces. The importance of gained
experience in the sampling process was also evaluated. Optimum
sampling parameters were determined for the described pesticides and
PCBs. Achieved recoveries ranged from 4 to 93% within relative standard
deviations varying between 1 and 67%. The quantitative limitations of
surface sampling techniques were thus demonstrated by these extremely
wide ranges. As a result of this investigation, a technique is recommended
that optimizes to the degree possible the various sampling parameters. The
value of practice and/or gained experience was also demonstrated.
-------
CONTENTS
INTRODUCTION 1
METHODS AND MATERIALS 2
RESULTS 6
SUMMARY 15
REFERENCES
FIGURE
1 Gas Chromatograms of Samples Taken from an Untreated Vinyl
Tile Surface with an Electron Capture Detector 7
TABLES
1 Summary of Analyte Analysis 4
2 Recovery Data Using Various Common Solvents 5
3 Sampling and Extraction Efficiencies for Various Sorbent
Materials 9
4 Percent Recoveries for Various Surface Types 10
5 Recoveries of Wiping and Blotting Techniques 11
6 Comparison of Sampling Efficiencies Between Individual
Samplers of Varying Experience 11
7 Wiping Efficiency vs. Area Coverage 12
8 Volume of Applied Solvent Correlated with Analyte Recovery 13
9 Recovery Data of Incremental Sampling of Chlorpyrifos from
Vinyl Tile 13
10 Recovery of Analyte vs. Amount Analyte Present 14
-------
INTRODUCTION
Based on personal experience and the available limited literature,
analytical characterization of inert surfaces for chemical contaminants
has proven reliable from a qualitative standpoint, but reproducible and
accurate results have been difficult to achieve. Surface measurements for
common structural-use insecticides and PCBs are valuable for both
enforcement and general monitoring purposes. Surface concentration
measurements can yield information related to the degree of original
contamination in a given situation, as well as the effectiveness of various
cleanup measures. For example, the U.S. Environmental Protection
Agency (EPA) recently promulgated a rule requiring reduction of PCB
contaminated surface and soil concentrations to certain prescribed levels.7
The available literature related to inert surface sampling for
pesticides and/or PCBs is limited. Few standard operating procedures exist
with respect to surface sampling, though Occupational Safety and Health
Administration (OSHA) has a wipe sampling procedure to establish
violations with respect to pertinent workplace regulations.2 Chavalitnitikul
and Levin3 conducted a surface sampling study with respect to lead oxide;
recoveries were demonstrated to be as high as 90% with good repeatability
from nonporous surfaces.
The imprecision and generally poor recoveries can be associated
largely with the sampling techniques and inherent limitations of the
process since analytical methodology for most insecticides and PCBs has
been standardized for some time. This study was designed to assess the
different variables associated with the surface sampling process, as well as
to determine the degree of potential interferences to be expected from
different types of structural materials. The variables of absorbing solvent,
sorbent type, various mechanical sampling techniques, and experience
were evaluated for both recovery and precision for the various pesticides
and PCBs. The various inert surface types studied were glass, aluminum
foil, waxed and unwaxed vinyl tile, painted (latex) wood, and polyurethane
treated wood veneer. From the developed data, sampling procedures are
-------
recommended for both specific pesticides (or PCBs) and unknown home-
treating insecticides in general.
METHODS AND MATERIALS
Sampling: In order to evaluate a specific surface sampling
variable, all other sampling parameters were held constant to the degree
possible. A measured solution of the analyte of interest was applied to a
fixed area (10 x 10 cm) with a microsyringe, and the solvent allowed to
evaporate. Thin-layer chromatography (TLC) saturation pads were the
sorbent material used for all the experiments except for the evaluation of
the optimum sorbent material itself. The saturation pad (11 x 11 cm) is
folded twice such that two ca. 2.5 cm square surfaces are exposed and two
inner pockets are formed. The solvent was then applied to both sides of the
pad and the inert surface of interest was sampled.
Based on previous experience, a sampling procedure using a blotting
technique demonstrated slightly better recoveries than a true wiping
motion; therefore all aspects of this study were conducted using a blotting
technique, except for the experiment comparing the two techniques. After
sampling of the designated surface area, the saturation pad was then rolled
such that the exposed sampling surfaces were inverted and subsequently
the pad was placed in a 40-mL volatile organics analysis (VOA) vial with a
Teflon®-lined cap. The other sampling materials used in the sorbent
experiment were also placed in a VOA vial after sampling for subsequent
analyte determination.
Solvent Evaluation: The following insecticides were evaluated for
absolute recovery and associated variability using four different common
solvents selected according to their availability and respective polarity index
(P1):4 (1) Technical chlordane, (2) chlorpyrifos, (3) malathion, (4) diazinon,
(5) propoxur, and (6) bendiocarb, as well as Aroclor 1260. The solvents used
were, acetone (P'=5.1), isopropanol (P'=3.9), dichloromethane (P'=3.1), and
isooctane (P'=0.1). 10 x 10 cm areas were treated with a known amount of
® Teflon is a registered trademark and appears hereafter without ®.
-------
analyte via microsyringe and sampled (after solvent evaporation) by a
blotting technique maintained as consistent as possible throughout the
entire evaluation. Each analyte was applied and sampled in triplicate or
quadruplicate for each solvent.
Sorbenfc Five different sorbents were evaluated to determine
maximum sampling efficiency. The sorbents evaluated were: (1) TLC
saturation pads, (2) commercially available pharmaceutical gauze pads,
(3) polyurethane foam plugs, (4) cotton balls, and (5) dental wicks. Three or
four replicate samples were taken from aluminum foil surfaces and
analyzed for each type of sorbent material. Technical chlordane,
chlorpyrifos, malathion, propoxur and Aroclor 1260 were evaluated using
the blotting technique.
Sampling Technique: Various aspects of technique were
evaluated with respect to recovery and precision in order to establish
optimum technical conditions of the sampling process. These included:
(1) a wiping versus a blotting motion, (2) solvent volume with respect to the
solvent applied to the sorbent material, (3) the effectiveness of repeating the
sampling process a second time on the same (10 x 10 cm) area. These
considerations were evaluated by sampling selected analytes from
aluminum foil using the appropriate solvent.
The relatively indeterminate variable of analyte concentration with
respect to recovery was also evaluated by sampling spiked levels of
chlorpyrifos equivalent to 0.12 ug/100 cm2, 1.2 ug/100 cm2, 12 ug/100 cm2,
and 120 ug/100 cm2. In addition, the effect of gained experience or practice
was also evaluated by having two relatively inexperienced laboratory
technicians take samples from prepared aluminum foil surfaces for
chlorpyrifos, in addition to the more experienced author.
Analytical Considerations: Upon completion of each sampling
exercise, the sorbent material was placed into a 40-mL VOA vial and
extracted for approximately 30 minutes in an automatic wrist shaker.
Table 1 summarizes all pertinent parameters with respect to the analyte
-------
analyses. Recoveries of pretreated sorbent materials were also determined
[Table 2] to validate the extraction efficiency.
Table 1
SUMMARY OF ANALYTE ANALYSIS
Compound
Chlordane
Chlorpyrifos
Malathion
Diazinon
Propoxur
Bendiocarb
Aroclor 1260
Extraction
Solvent
Hexane
Hexane
Hexane
Hexane
Methanol
Methanol
Hexane
Analysis
GC
GC
GC
GC
HPLC
HPLC
GC
Detector
ECD
ECD
FID
FID
UV
UV
ECD
Operating
Conditions
A
A
B
B
C
C
A
Operating Conditions
A = DB-5 wide bore capillary column, 0.32 id. Samples were analyzed with a temperature
gradient from 60 to 260 at 7 °C/minute.
B = XL Me Silicone wide bore capillary column, 0.32 id. Samples were analyzed with a
temperature gradient from 100 to 250 "Cat 7° I minute.
C = Reverse Phase C-18 column, isocratic conditions; mobile phase composition at 50/50
acetonitrile I water.
-------
Table 2
RECOVERY DATA USING VARIOUS COMMON SOLVENTS
Sorbent: TLC saturation pads
Compound
Chlordane
Chlorpyrifos
Malathion
Diazinon
Aroclor 1260
Bendiocarb
Propoxur
Solvent
Isooctane
Acetone
Isopropanol
Dichloromethane
Isooctane
Acetone
Isopropanol
Dichloromethane
Isooctane
Acetone
Isopropanol
Dichloromethane
Isooctane
Acetone
Isopropanol
Dichloromethane
Isooctane
Acetone
Isopropanol
Dichloromethane
Isooctane
Acetone
Isopropanol
Isooctane
Acetone
Isopropanol
No. of
Replicates
4
4
4
3
4
4
4
3
4
4
4
4
4
4
4
4
4
4
4
4
3
4
4
4
3
3
Spike
(US)
1.3
1.3
1.3
1.3
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
375
375
375
375
50
50
50
50
515
515
1,650
2,560
2,560
2,560
Recovery
%
54
71
42
23
56
72
34
39
80
66
42
81
70
57
62
55
80
76
71
60
51
85
84
49
90
96
RSD
%
9
17
24
14
14
14
28
28
5
6
12
7
7
9
13
9
3
15
13
10
18
6
8
10
4
8
-------
RESULTS
The basic focus of this study was to evaluate the parameters of
surface sampling for PCBs and common home-treating pesticides in an
attempt to evaluate sampling variables and improve quantitative
characterization. The results not only statistically confirmed the
quantitative limitations of the process, but also lent some insight as to
possible improvements that can be used in sampling site characterization.
The old adage "practice makes perfect" was apparently appropriate when
attempting to obtain maximum recovery from a surface area.
The solvents studied were chosen over a wide polarity range to
establish correlation between analyte recovery and polarity. The data
demonstrated the inherent solvent qualities of acetone [Table 2], which
yielded the best overall recoveries. The disadvantage of acetone was its
propensity for also removing potential interfering compounds (this seems
most prevalent when using an electron capture detector) when sampling a
finished surface [Figure la]. The nonchlorinated organophosphorous
pesticides diazinon and malathion, demonstrated better overall recovery
when isooctane was used as the solvent. Isooctane did not remove as much
interference [Figure Ib], but did give poorer recoveries with respect to the
chlorinated pesticides. Isopropanol demonstrated significant recoveries
with respect to the carbamate pesticides, though less than expected
recoveries were obtained with respect to the other analytes. The significant
finding concerning the solvents studied was that pesticides may be more
solvent specific than earlier surface sampling studies suggested. This
finding is only significant when a specific pesticide of interest is known
before a site is sampled. The recovery ability is not the overriding factor in
choosing a solvent, since qualitative characterization may also be inhibited
by the interference. Isooctane represented the more desirable solvent with
respect to potential interference, and gave recoveries comparable to acetone.
Isopropanol (100%) demonstrated low recoveries except for the carbamates
bendiocarb and propoxur.
-------
Figure 1
GAS CHROMATOGRAMS OF SAMPLES TAKEN FROM AN UNTREATED
VINYL TILE SURFACE WITH AN ELECTRON CAPTURE DETECTOR
(a) Chromatogram of a Surface Sample Taken From Vinyl Tile Surface
Using Acetone as the Solvent.
*
R
(b) Chromatogram of Surface sample Taken From Vinyl Tile Surface
Using Isooctane as the Solvent
-------
8
There appears to be correlation between surface area of the sorbent
material and analyte recovery. The TLC saturation paper and the gauze
pads both possessed the greatest surface areas and correspondingly yielded
higher recoveries [Table 3], with the saturation paper demonstrating more
reproducible results. These conclusions with respect to surface area are in
disregard of any other intrinsic properties which may have influenced the
final results. The accessibility of the gauze pads and the TLC saturation
paper (essentially a heavy filter paper) make both the materials practical
choices as a sorbent. The rigidity of the TLC saturation paper allows for
easier handling throughout the sampling process.
Surface type was the most limiting factor with respect to obtaining
representative quantitative results. Recoveries were correlated to surface
porosity [Table 4], with the porous wood and veneer surfaces yielding
correspondingly lower recoveries, as compared to the nonporous
aluminum foil. Considering that glass is also a relatively nonporous
surface, the chlorinated pesticides and Aroclor 1260 yielded poorer than
expected recoveries. Strong adsorption of PCB to glass surfaces was
demonstrated by Hattula et al.5 Since the majority of field surface samples
are taken from non-ideal surface types the data presented may give some
insight as to recovery and reproducibility of results from specific surface
samples.
The wiping versus blotting aspect came of interest from preliminary
work that suggested a true wiping motion may have dispersed the analyte
along the surface rather than removing the analyte. Significant differences
between the two techniques were not apparent in the current study, though
the blotting technique demonstrated slight improvement in reproducibility
[Table 5].
-------
Table 3
SAMPLING AND EXTRACTION EFFICIENCIES
FOR VARIOUS SORBENT MATERIALS
Solvent: Optimum solvent from Table 2
Compound
Chlordane
Chlorpyrifos
Malathion
Propoxur
Aroclor 1260
Material
Polyurethane
Cotton balls
Gauze
Saturation paper
Dental wick
Polyurethane
Cotton balls
Gauze
Saturation paper
Dental widk
Polyurethane
Cotton balls
Gauze
Saturation paper
Dental wick
Polyurethane
Cotton balls
Gauze
Saturation paper
Dental wick
Polyurethane
Cotton balls
Gauze
Saturation paper
Dental wick
Sampling
Recovery
%
30
ND°
71
65
24
48
20
72
77
45
53
31
64
80
56
72
26
71
93
66
50
30
80
48
RSD
%
5
-
10
8
20
21
40
10
4
18
13
29
14
4
9
7
19
13
3
8
20
13
3
19
Extraction
Recovery
%
101
101
95
105
82
85
99
95
114
90
89
86
86
95
109
103
99
100
102
85
96
84
83
101
RSD
%
15
12
2
19
7
14
8
9
4
8
3
14
1
7
3
1
3
1
3
1
13
23
18
9
a = Not detected
-------
10
Table 4
PERCENT RECOVERIES FOR VARIOUS SURFACE TYPES
Sorbent: TLC saturation pads
Solvent: Optimum solvent from Table 2
Compound
Chlordane
Chlorpyrifos
Malathion
Propoxur
Aroclor 1260
Surface
Type
Veneer
Glass
Aluminum foil
Tile (wax)
Tile (no wax)
Wood (painted
Veneer
Glass
Aluminum foil
Tile (wax)
Tile (no wax)
Wood (painted)
Veneer
Glass
Aluminum foil
Tile (wax)
Tile (no wax)
Wood (painted)
Veneer
Glass
Aluminum foil
Tile (wax)
Tile (no wax)
Wood (painted)
Veneer
Glass
Aluminum foil
Tile (wax)
Tile (no wax)
Wood (painted)
No. of
Samples
3
4
4
3
3
3
3
4
4
3
3
3
3
4
4
3
3
3
3
4
4
3
3
3
3
4
4
3
3
3
Recovery
(%)
10
28
65
30
36
16
6
58
72
42
31
23
3
84
80
67
35
6
6
82
93
22
18
15
20
41
80
45
54
19
RSD
(%)
30
39
8
20
14
13
67
19
6
10
7
6
33
5
4
8
29
17
50
7
9
14
17
13
30
15
3
7
13
5
-------
11
Table 5
RECOVERIES OF WIPING AND BLOTTING TECHNIQUES
Sorbent: TLC saturation pads
Solvent: Optimum solvent from Table 2
Surface: Aluminum foil
Wiping Blotting
Recovery RSD Recovery RSD
Compound Replicates (%) (%) (%) (%)
Chlordane
Chlorpyrifos
Malathion
Propoxur
Arochlor 1260
4
4
4
4
4
60
65
87
58
78
12
6
5
3
12
56
62
90
49
79
7
5
2
2
8
Experience and consistency of technique were determined to play
significant roles concerning the overall accuracy and precision, in
obtaining surface samples. Three samplers of varying degrees of
experience collected surface samples using the described technique.
Higher recoveries and greater reproducibility were reflected by the more
experienced sampler [Table 6].
Table 6
COMPARISON OF SAMPLING EFFICIENCIES BETWEEN
INDIVIDUAL SAMPLERS OF VARYING EXPERIENCE
Sorbent: TLC saturation pads
Solvent: Acetone
Surface: Aluminum foil
Analyte: Chlorpyrifos
Sampler
a
b
c
Sample
No.
4
4
4
Recovery
(%)
44
53
62
RSD
(%)
16
17
6
Sampling Experience: c > b > a
-------
12
Area coverage with respect to this study was defined as the number
of times a sampled area (10 x 10 cm) was completely wiped both horizontally
and vertically. Preliminary work at this laboratory had established an
optimum area coverage of seven wipes per unit sample. The present study
demonstrated a somewhat cyclic correlation, with recovery and
reproducibility optimized after an area coverage of only one per unit sample
[Table 7].
Table 7
WIPING EFFICIENCY VS. AREA COVERAGE
Sorbent: TLC saturation pads
Solvent: Acetone
Surface: Aluminum foil
Analyte: Chlorpyrifos
Recovery
Area Coverage Sample No.
1
3
5
7
9
11
13
3
3
3
3
2
2
2
71
53
62
65
50
44
34
Correlation between solvent volume used on the sorbent material
[Table 8], the highest overall recovery was demonstrated by the saturated
condition, but the relative imprecision and undefined volume of this
condition limit the usefulness of these data. Repeatability of results
justified the use of a total volume of 2 milliliters of solvent as applied to both
sides of the sorbent material.
-------
13
Table 8
VOLUME OF APPLIED SOLVENT CORRELATED WITH
ANALYTE RECOVERY
Sorbent: TLC saturation pads
Solvent: Optimum solvent from Table 2
Surface: Aluminum foil
Solvent Volume
0 mL 1 mL 2 mL 3 mL Saturated
Compound
Chlorpyrifos
Aroclor 1260
Chlordane
24
24
24°
43
58
28
59
62
59°
39
75
470
87
56
65
a = Percent recovery is an average of two samples
An evaluation of using two sorbents in sequence was carried out,
similarly to the technique by Leidy et al.6 This portion of the study was
conducted on vinyl tile considered a relatively porous surface. The results
using the TLC saturation pads to remove chlorpyrifos with acetone and
isopropanol are shown in Table 9. The significant increases in analyte
recovery using the sequential sampling procedure on a nonporous surface
are most likely due to solubilizing of the analyte by the first sorbent
treatment while the second treatment removed the analyte. The
experiment was duplicated on a nonporous surface; however, the
sequential procedure did not yield significant increases in recovery.
Table 9
RECOVERY DATA OF INCREMENTAL SAMPLING OF
CHLORPYRIFOS FROM VINYL TILE
Solvent
Acetone
Isopropanol
Spike
(Hg)
1.26
1.26
Recovery
Sorbent 1
(Hg)
0.66
0.91
Recovery
Sorbent 2
(Hg)
0.31
0.18
Total
Recovery
(%)
77
86
-------
14
The concentration of analyte present did not yield any significant
correlation with respect to recovery [Table 10]. This would appear to
suggest that the concentration of the analyte is not a factor with respect to
analyte recovery at or above the analytical detection level for quantitation.
Table 10
RECOVERY OF ANALYTE VS.
AMOUNT ANALYTE PRESENT
Sorbent: TLC saturation pads
Solvent: Acetone
Surface: Aluminum foil
Replicates: 2
Analyte Amount
(ug)/100cm2 % Recovery
0.1 62
1.2 58
12.0 59
120.0 61
-------
15
SUMMARY
Based on the results of this limited study, the following guidelines
are provided to personnel who plan to engage in a surface sampling project
for pesticides and/or PCBs:
• TLC saturation pads or 3-inch x 3-inch gauze pads as the
sorbent of choice
• Isooctane as the most effective solvent applied to the sorbent for
most pesticides of interest; if a carbamate or a known polar
pesticide are to be sampled, isopropanol is the more effective
solvent.
• A single sorbent with a single pass in both the horizontal and
vertical direction can be used for a nonporous surface such as
glass or metal; however, the process should be repeated with a
second solvent treated sorbent for a porous surface such as
wood and/or tile.
-------
REFERENCES
1. 40 Code of Federal Regulations: Part 761, Subpart G.
2. U.S. Department of Labor: Wipe Sampling Policies and Procedures
In., Industrial Hygiene Manual. Chapter VI, OSHA (1977).
3. Chavalitnitikul, C and Levin, L: A Laboratory Evaluation of Wipe
Testing Based on Lead Oxide Surface Contamination. Am. Ind. Hvg.
Assoc. si. 45(5):311-317 (1984).
4. Snyder, L. R,: Classification of Solvent Properties of Common
Liquids, si. Chromatogr. 92:223-230 (1974); J. Chromatogr. Sci.
16:223-234(1978).
5. Hattula, M.L. and O. Karlog: Adsorption of Polychlorinated
Biphenyls (PCS) on Glass Surfaces. Dansk. Tidsskr. Farm. 45:259-
261 (1971).
6. Leidy, R.B., C.G. Wright and H.E. Dupree, Jr.: A Sampling Method
to Determine Insecticide Residues on Surfaces and its Application to
Food-Handling Establishments. Environmental Monitoring and
Assessment. 9:47-55(1987).
------- |