PROCEEDINGS
THIRD ANNUAL SYMPOSIUM
United States
Environmental Protection Agency
Symposium
on
SOLID WASTE TESTING
and
QUALITY ASSURANCE
^^•^••i
Volume I
JULY 13-17, 1987
WASHINGTON, D.C
WEST1N HOTEL
Symposium Managed by American Public Works Association
-------
PROCEEDINGS
THIRD ANNUAL SYMPOSIUM
United States
Environmental Protection Agency
Symposium
on
SOLID WASTE TESTING
and
QUALITY ASSURANCE
Volume I
JULY 13-17, 1987
WASHINGTON, D.C.
WESTTN HOTEL
Symposium Managed by American Public Works Association
-------
PROCEEDINGS INTRODUCTION
One of the major environmental problems facing the United States, as
weJI as other nations, is the need for safe handling and disposal of
hazardous waste. A fundamental component of all programs relating to
waste management is the need to perform measurements. These
measurements include waste composition and properties; effectiveness
f management processes; engineering properties of materials used in
constructthg management units; and, last but not least, long term
p Lfc mance of such management units. Thus, the pivotal roles played
by the measurement methodology and, its attendent, quality assurance.
The analysis of complex waste matrices presents the environmental
ci amimity with demanding analytical problems for which solutions are
being developed at a rapid rate. This annual symposium series,
presented by the EPA’S Office of Solid Waste, is designed to focus on
recent developments in testing methods and quality assurance of
importance to both the RCRA and CERCLA programs.
The symposium highlights developing requirements for quality assurance
as well as new analytical procedures intended to be used in EPA ’s
natirnai RCRA and CERCLA hazardous waste management programs. Our
purpose in holding these symposia is several fold. First, as a means
of communicating what EPA is doing regarding the activities EPA has
already initiated to upgrade the state—of—the—art as reflected in the
regulations and in SW—846. Second, to describe the direction EPA’s
program is taking with respect to testing and quality assurance
issues. Third, as a forum for discussion between Agency personnel and
representatives from public and private laboratories involved in waste
sampling and evaluation.
DAVID FRIEDMAN
CHIEF, METHODS SECTION
OFFICE OF SOLID WASTE
-------
PROGRAM COMNIT EE
David Friedman
Chief Methods Section
Office of Solid Waste
U.S. EPA
Denise Zabinski
Chemist
Office of Solid Waste
U.S. EPA
David Bennett
Chief, Toxics Integration Branch
Hazardous Site Evaluation Division
(WH—548A)
U.S. EPA
Billy Fairless
Chief E1’ICM/ENSV
Region 7
U.S. EPA
Paul Friedman
Chemist
Office of Solid Waste
U.S., EPA
Duane Geuder
Chemist
Office of Emergency and
Remedial Response
U.S. EPA
Gary Ward
Chemist
Office of Remedial Response
U.S. EPA
Connie Glover
Manager
Lancy Environmental
Services Division
Liew Williams
Deputy Director
Quality Assurance and Methods
Research Division
U.S. EPA
Las Vegas, NV
Gail Hansen
Chemist
Office of Solid Waste
U.S. EPA
Kenneth Jennings
Environmental Scientist
Office of Waste Program
Enforcement
U.S. EPA
Tom Logan
Engineer
Environmental Monitoring
and Support Lab
U.S. EPA
Research Triangle Park,NC
William Loy
Chemist
Environmental Services
Division
Region 4
Athens, Gt
Theador Martin
Research Chemist
Environmental Monitoring
and Support Lab
U.S. EPA
Cincinnati, OH
Florence Richardson
Quality Assurance
Officer
Office of Solid Waste
U.S. EPA
Reva Rubenstein
Chief, Health Assessment
Section
Office of Solid Waste
U.S. EPA
Robert Stevens
Chief
California Department
of Health Services
-------
TABLE OF CONTENTS
Volume I
AIR AND GROUND WATER MONITORING
Comparison of TOX and GC,4 1S Data for RCRA Groundwater
Monitoring Well Samples
S. Pruskin et al. 1—1
Nonvolatile Organics as Probes for Contaminated Ground Water
Plumes from Hazardous Waste Management Facilities
R. Stephens et al. 1—15
Gas Chromotography Matrix Isolation Fourier Transform Infrared
(GC/MI—IR) Spectroscopy for Monitoring Air Pollutants
B. Fairless et al. 1—17
Evaluation of Flux Chamber Method for Measuring Air Emissions
of Volatile Organic Compounds from Surface Impoundments
A. Gholson et al. 1—59
Field Evaluation of Three Methods for Soil—Gas Measurement
for Delineation of Ground Water Contamination
H. Kerfoot 1—67
Procedures Used to Measure the Amount of Dioxin in the Ambient Air
Near a Superfund Site Clean—up Operation
B. Fairless et al. 1-81
VOC Emission Rates from Solid Waste Landfills
W. Vogt et al. 1—103
A Control Chart Strategy for Ground Water Monitoring
G. Flatman, T. Starks 1—115
The Use of Geostatistics for Contour Maps
E. Englund, G. Flatman 1-129
BIOLOGICAL TEST METHODS
Utility of Immunoassay for Trace Analysis of Environmental
Contaminants
J. Van Emon 2-1
Derivation and Use of Nonoclonal Antibodies for
Environmental Monitoring
A. Karu 2-7
-------
Imeunoassay for the Determination of Pentachiorophenol and Related
Compounds in Water Samples
T. chiang 2-9
Ultra Sensitive Bioassay for Dioxin
R. Schuman, K. Hunter 2—il
Strategies for Using Bioassay Methods for the Identification
of Hazardous Components and Comparative Risk Assessment
of Complex Mixtures
.1. Lewtas 2—19
Mutagenicity in Salmonella of Hazardous Wastes and Urine
from Rats Fed These Wastes
D. DeMarini et al. 2—45
Application of a Simple Short—Term Bioassay for the Identification
of Genotoxins From Hazardous Wastes
S. Sandhu 2—63
Application of Battery of Aquatic Toxicity Tests to Solid
Waste L 1 eachate characterization and Environmental Effects Prediction
D. Mount 2—79
Screening of Complex Solid Wastes for Chemicals Which Bioaccuxnulate
and Cause Environmental Hazards
G. Veith 2—81
Bioactivity Differences of Water and Sodium Acetate Eluate from
Municipal and Industrial Wastes
S. Peterson et al. 2—83
Use of Mosses as Indicators of Air Pollution
G. Sage 2—95
Use of Tradescantia for Toxicity Testing of Hazardous Waste
W. Lower 2—101
Statistical Approaches to Screening Hazardous Waste Sites
for Toxicity
3. Thomas 2—113
Application of a Biomarkers—Waste Characterization Approach
to the Prediction of Organism Responses following Exposure to
Cont niinated Marine Sediments
G. Pesch 2—135
Environmental Monitoring
0. Meyn 2—137
Assessment of the Microscreen Phage—Induction Assay for
Screening Hazardous Wastes
V. Houk, D. DeMarini 2—139
-------
Assessment of the TLC/Salmonella Assay for Screening Hazardous
Wastes
V. Houk, L. Claxton 2—159
Methodology for Evaluating Potential Human Health Effects of
Microorganisms that Degrade Hazardous Wastes
S. George et al. 2—175
Bioassay Determination of Soil Assimulative Capacity
S. Peterson, J. Greene, W. Miller 2—193
ENFORCEMENT
RCRA Land Disposal Restriction Program
V. Hays 3-1
RCRA Laboratory Audit Inspection (L I) Program
E. Pryor 3—5
Operation and Maintenance Guidance for RCRA Ground—Water
K. Jennings 3—7
Unconventional Techniques and Uncommon P nalyses in Hazardous
Waste Analysis
D. Kendall 3—9
LEACHING AND PHYSICAL METHODS
Modification to the TCLP Procedure for Problem Matrices
P. Marsden, L. Williams, G. Hansen 4—1
Further Development of the Liquid Release Test
P. Hoffman et al. 4—7
Validation of Toxicity Characteristic Leaching Procedure
(TCLP) and pplication to Industrial Wastes
R. Ragsdale, R. Meierer 4—9
Performance of the Toxicity Characteristic Leaching Procedures
L. Newcomer, W. Blackburn, G. Hansen 4—25
Determination of Organic Components in Leachates — A Survey
J. Poppiti, E. Johnson 4—47
Evaluation of the TCLP for Determining the Potential of Oily Wastes
R. Truesdale, S. Winters, G. Hansen 4—49
-------
METALS AND MISCELLANEOUS ANALYTES
Evaluation of Microwave Techniques to Prepare Solid and
Hazardous Waste Samples for Elemental Analysis
D. Binstock et al. 5—i
Results of an Interlaboratory Study of ICP Method 6010 Combined
with Digestion Method 3050
T. Hinners et al. 5—11
Preliminary Studies of the Separation and Determination
of Cr(VI) and Cr (VIII) in Waste Water by Solid Absorbent
Extraction and GFM Analysis
R. Stockton et a!. 5—29
Evaluation of SW—846 Cold—Vapor Mercury Methods 7470 and 7471
W. Beckert et a!. 5—33
Factors Affecting EP Toxicity Metals Results
A. Jirka et al. 5—47
Ion Chromatography for the Analysis of Anions in Hazardous
Waste Matrices
R. Kell et al. 5—53
Sample Handling for the Analysis of Cyanides in Solid and
Hazardous Wastes
J. Ritzert 5—67
Determination of Total Sulfide in Solid Waste
LI. I bana 5-73
WX - A Screening Parameter for Environmental Samples
J. Snyder, P. Keliher 5—75
Pyrolysis/MicrocoulometriC Determination of Total Organic
Halides in Solids and Oily Wastes
V. Lieu, V. Woo 5—101
Develo *nent and Evaluation of Test Methods for Total Chlorine
in Used Oils and Oil Fuels
A. Gaskill et al. 5—119
Microwave Acid Sample Decomposition for Elemental Analysis
13. Kingston, L. Gassie 5—121
characterization of Municipal and Household Hazardous Waste
G. Mitchell et al. 5—123
Environmental Screening Methods for Total Organic Halogens
1. Whitechurch, S. Smyers 5—141
-------
AIR AND
GROUND WATER
MONITORING
thai rperson
Jerry Garman
Environmental Scientist
Ground Water Monitoring
and Sampling
U.S. E ’A
401 N Street, S.W.
Washington, D.C. 20460
-------
COMPARISON OF TOX AND GC/MS DATA
FOR RCRA GROUNDWATER MONITORING WELL SANPLES
Steven Pruskin, Leonard Voo, Robert Mason, and Daniel Lillian, U.S.
Environmental Protection Agency, Region II Laboratory, Edison, New
Jersey
ABSTRACT
The method for the analysis of total organic halogen (TOX) was
developed as a means to trace the products of disinfection in
drinking water. It was intended to measure, collectively, compounds
such as trihalomethanes and halogenated ethanes. Briefly, the
method involves passing a measured amount of sample through a column
of activated carbon, followed by pyrolysis of the carbon and
analysis of the gases produced for total halogen (as chloride) by
microcoulometric titration. It has been shown to produce good
recoveries for spiked deionized water and for other simple, fairly
clean waters that are free of turbidity. This method is documented
as Method 9020 in SW—846 (Test Methods for Evaluating Solid Waste).
RCRA regulations require the monitoring of TOX as an indicator of
groundwater contamination at many hazardous waste management
facilities (40 CFR 265.92). Method 9020 is recommended as an
economical method for TOX analysis. Another way of measuring TOX is
to analyze samples for volatile and semivolatile organics by GC/MS
(SW—846 Methods 8240 and 8270) and calculating the sum of the
halogen contents (as chloride) of all compounds that are found.
Samples from RCRA groundwater monitoring wells are generally more
complex than the drinking water samples that were analyzed when
Method 9020 was initially evaluated. In the Region II laboratory we
have analyzed many groundwater samples by Method 9020 and by Methods
8240 and 8270. This report is a presentation of a preliminary
comparison of the results obtained by these two techniques.
BACKGROUND
Dressman, Stevens and co—workers in the Drinking Water Research
Division (DWRD) of the Municipal Environment Research Laboratory, US
Environmental Protection Agency (USEPA), Cincinnati, Ohio, have been
very active in the development of the analysis of Total Organic
Halogen (TOX) as a group parameter. They have thoroughly reviewed
the history of TOX. 1 Briefly, they reported that the use of TOX
began in the early 1970s when Kuhn and Sontheimer 2 developed a
technique for the measurement of total organic chlorine (TOC1) in
activated carbon. Their objective was for this test to be used to
monitor the operation of activated carbon filters in water treatment
plants. These plants were purifying water from the Rhine River for
use as drinking water. A second use for the test was for comparison
1—1
-------
iodide, making the analysis applicable for TOX, not just for TOC1.
This was desirable since organobromides and lodides are also
indicators of water contamination.
The next major improvement was made by Takahashi and Moore. 5 They
packed GAC Into a rnicrocolumn, passed the sample through the column
to adsorb organics onto the GAC, passed a potassium nitrate solution
through the column to displace inorganic halides, combusted the
sample as described above, and analyzed by microcoulometric
titration. They streamlined the analysis step by passing the
combustion gases directly into the titration cell, rather than into
a collection vessel from which an aliquot was taken and injected
Into a separate microcoulometric titration cell.
At this point, the DWRD published USEPA Interim Method 450.1, Total
Organic Halides. 6 The same method was also published in SW—846 as
Method 902O . This Is a method “to be used for the determination of
Total Or&anic Halides as Cl ... in drinking and ground
watera.” 6 ’ Stevens and Dressman reported 1 that the first
commercial TOX analyzers were developed by the Dohrmann Division of
the Xertex Corporation under contract to the DWRD, and that
instruments based on Method 9020 were available from Dohrmann and
from Mitsubishi aiemical Industries Ltd.
Method 9020 was published shortly after the promulgation of a
regulation requiring many hazardous waste management facilities to
monitor the TOX in their groundwater. 8 In order to monitor the
compliance of these facilities with this regulation, we, in the
USEPA Region II Laboratory, have analyzed many groundwater
monitoring well samples for TOX using Method 9020. Many of these
groundwater samples were also analyzed for volatile and extractable
organic compound by GC/MS using USEPA Methods 8240 and 8270
respectively. 9 In this paper, we present a comparison of the
results obtained from the Method 9020 TOX analysis with the TOX that
would be expected on the basis of a summitIon of the halogen
concentrations In the halogenated organic compounds (as chloride)
that were found by GC/MS; this sitmm tion will be referred to as TOX
by GC/MS. This preliminary comparison is being made for 115 wells
at 16 sites for which both Method 9020 TOX and GC/MS analyses were
performed.
Previous Studies
Dressman and StevenslO reported a comparison of Purgeable Organic
Halogen (POX) results for spiked deionized water with those
obtained by analyzing the same samples for Total Trihalomethanes by
GC. They found that the two techniques agreed within 20% for POX
values of 40 ug Cl/L or more, but up to 80% differences were found
at levels below 40 ug Cl IL.
1—2
-------
Williams, Coburn and Bancsi - 2 compared the results obtained for
Method 9020 TOX analyses with TOX by CC/MS analyses of groundwater
samples at ten sites in Canada. At nine of these sites (in which a
total of 25 wells were analyzed) TOX values were 44 ug/L or less and
Method 8240 GC!MS analysis showed volatile organics at 6 ug/L or
less. No further analyses were made at these sites. Comparison of
the results for the Method 9020 TOX with the TOX by GC/MS from these
nine sites showed no correlation between the two sets of values.
The authors attributed the apparent lack of relationship to the low
values. At the tenth site, six wells were sampled and analyzed by
Method 9020 TOX, Volatile Organics by GC/MS (Method 8240) and
Extractable Organics by GCIMS (Method 8270). One well was
upgradient of the site (well U) and five were downgradient of the
site (well Dl, D2, D3, D4, and D5). Well U had a method 9020 TOX
level of 11 ug/L and TOX by GC/MS of 12 ugIL. Well D3 had lower TOX
levels than well Ti by both methods. The four remaining wells all
had substantial levels of TOX by Method 9020 and volatiles by Method
8240. No halogenated compounds were found by Method 8270. In well
D4 they found 560 ug/L of TOX by Method 9020 and 849 ug/L of TOX by
CC/MS. They explained this 66% recovery by the TOX analyzer of
compounds identified by GCIMS as being in agreement with typical
recoveries of volatile organics (at concentrations greater than 200
ugIL) previously reported by Dressinan and co-workers. 13 (Dressman
and co—workers 1 -° later reported that their low recoveries were
caused by the use of a heavily vitrified tube in their pyrolysis
system, and that when a fresh tube was used, nearly complete
recoveries were obtained). Wells Dl and D2 had Method 9020 TOX
values of 1235 ug/L and 1402 ugiL and TOX by CC/MS values of 559
ug/L and 504 ug/L respectively. In an attempt to account for the
difference, Williams and co-workers took the Method 8270 extracts
and solvent exchanged them into cyclohexane. These cyclohexane
extracts were then analyzed by TOX analyzer. This analysis found
947 ug/L of TOX in Dl and 396 ug/L of TOX in D2 from halogenated
compounds that had not been found by CC/MS. Once this is taken into
account, the Method 9020 TOX values agree fairly well with the TOX
value calculated for the CC/MS samples. Well D5 had a Method 9020
TOX of 122 ug/L and TOX by GC/MS of 56 ugIL. Cyclohexane extracts
of this sample were not analyzed because that method was not
sensitive enough to account for missing TOX at this low level. When
the data from all six wells was analyzed, the TOX values from Method
9020 and the value for TOX by CC/MS augmented by the cyclohexane
extract values agreed with each other with a correlation coefficient
of 80%. The authors concluded that TOX is a useful technique for
screening groundwater for contaminants, and that further study was
needed to clarify the relationship between data generated by TOX and
GC/MS.
The intent of this study was to continue the work of Williams and
co-workers in evaluating a new method, Method 9020, by comparing the
results it produces with those obtained using an alternate, more
I—
-------
established procedure, CC/MS. We have added to the database and
confirmed the indication by Williams and co—workers that the data
are not directly comparable. More work is clearly needed to
elucidate and quantify the differences.
METHOD
1. Samples were collected from groundwater monitoring wells at
sixteen hazardous waste management facilities throughout EPA
Region II. The region consists of New York, New Jersey, Puerto
Rico and the US Virgin Islands. Samples were collected by
members of the USEPA Region II Environmental Services Division,
Surveillance and Monitoring Branch. They followed procedures
listed in the RcRA Ground Water Monitoring Technical Enforcement
Guidance Document.’ 4 Samples were collected with bailers except
where the facility had some other type of sampling device
dedicated to their wells. Samples were then transferred from the
bailer to glass sample bottles. Separate bottles were filled for
each analysis (volatile organics, then TOX, then extractable
organics). All samples from a single well were collected within
a three hour period unless the well recharge rate was too slow,
in which case they were collected as rapidly as the well
allowed. Initially 1000 ml samples, enough for quadruplicate
analysis by Method 9020, were collected. It was found that as
consecutive samples were decanted from the sample bottle, the
measured TOX increased dramatically. This was probably due to
organic halides adsorbed on sediment particles, which were more
prevalent lower in the bottle. More reproducible results were
obtained by using 150 ml sample bottles and only making one
determination from each bottle.
2. TOX was determined using USEPA Method 9020 via a Dohrmann DX-20
Total Halogen Analyzer. There are two parts to the analyzer,
Figure 1 is a diagram of the mlcrocolumn carbon adsorption
system, and Figure 2 shows the pyrolysis and microcoulometric
detection system. The method calls for a sample that is free of
undissolved solids because the solids will block the sample flow
at the top of the carbon column. Most of the groundwater samples
that were brought into the lab did not fit this criterion. To
avoid excessive loss of volatiles, sample manipulation was kept
to a minimum. To permit visible solids to settle, the sample was
allowed to sit in its closed container. Then the sample was
carefully decanted from the sample container into the adsorption
module’s sample reservoir. Samples with solids that did not
settle after standing were diluted by pipetting an aliquot of the
sample into a volumetric flask with some water in it and diluting
to volume. This technique was also used for samples that were
expected to be high in TOX or in inorganic chloride. Method 9020
requires that duplicate runs agree within l5 . We found that
this level of agreement could not be achieved for groundwater
1—4
-------
CARBON
MINI-COLUJIN
FIGURE
MICROCOLUMN CARBON ADSORPTION SYSTEM
M
‘“a
SA PL
SAMPLE
RESERVOIR
1—5
-------
FIGURE 2
PYPOLYSIS DETECTION
SYSTEM
M1CROCOULOt1ET C
TITRATION CELL
co 2
OR
-------
samples, so a less rigid standard of 40% difference was adopted.
In many cases where even this standard could not be achieved the
data was reported with a “J” qualifier to indicate that the value
was an estimate.
3. Analysis of Volatile Organic Compounds was performed by USEPA
Method 8240 using either a Finningan 3200 GCIMS system with a
Hewlett—Packard 7675A Purge and Trap sampler or a Finnigan 3300
GC/MS system with a Tekmar LSC—2 Purge and Trap sampler.
4. Determination of Extractable Organic Compounds was performed by
USEPA Method 8270 using a Finnigan 4000 CC/MS system.
5. All analyses reported in this study were performed at the USEPA
Region II Laboratory in Edison, New Jersey. These analyses were
performed between October 1984 and December 1986. The study
includes all wells from which data were reported for TOX (Method
9020), volatile organic compounds (Method 8240), and extractable
organic compounds (Method 8270). Although we are reporting all
of the Method 9020 TOX analyses which have accompanying GC/MS
(8240 and 8270) data, we are not reporting all Method 9020 TOX
analyses performed. Many analyses were not reported because they
were not accompanied by Method 8240 or Method 8270 analyses.
Table 1 shows the number of samples reported at each of the sites
in this study.
6. TOX by GC/MS values were calculated as follows. For each analyte
found by GC/MS, the concentration of the analyte was multiplied
by the weight fraction of halogen, as chloride, in that analyte.
The sum of these halogen concentrations was then calculated for
all analytes found in the sample. This sum is the TOX. Table 2
shows an example of this calculation for one of the sites that
was investigated.
RESULTS AND DISCUSSION
As shown in Table 1, the data fall into four categories; values
below the TOX detection limit for both Method 9020 TOX and TOX by
CC/MS analyses; values above the detection limit where Method 9020
TOX and TOX by CC/MS agree within a factor of two; samples with
Method 9020 TOX values greater than double the corresponding TOX by
CC/MS values and samples with TOX by CC/MS values greater than
double the corresponding Method 9020 TOX values. Table 1 also shows
the number of samples in each category for each site.
In dealing with the first category, samples below the detection
limit, it is essential to first decide what that detection limit
is. A limit of 10 ug/L is generally used in Method 9020 because
that was found in this laboratory to be three times the standard
deviation of a nitrate washed column blank pair. When samples have
1—7
-------
Table I
Number of Samples Number of Samples From Site in Category
Site # Prom Site I II Ill IV
1 11 1 4 6 0
2 6 0 4 2 0
3 4 1 1 2 0
4 3 0 2 0 1
5 7 0 0 7 0
6 7 0 0 7 0
7 6 2 0 4 0
8 5 1 0 4 0
9 13 4 2 7 0
10 9 1 1 1 0
11 8 7 0 0 1
12 6 1 0 4 1
13 4 3 0 1 0
14 5 1 0 4 0
15 4 4 0 0 0
16 17 9 2 5 1
Totals 115 35 16 60 4
Category 1 Samples with Method 9020 TOX and TOX by CC/MS below detection limits.
Category II Samples with TOX by both methods agreeing within a factor of two.
Category III Samples with Method 9020 TOX more than double lOX by CC/MS.
Category IV Samples with Method 9020 TOX less than half TOX by CC/MS.
1—8
-------
been diluted to reduce matrix effects that interfere with the
analysis, the detection limit must be adjusted for this dilution.
When the detection limit of 10 ug/L is used, 12 of the 115 samples
fall into the first category. However, since the summation of
detection limits for TOX by GCIMS is greater than 10 ugIL, we have
raised the detection limit for this study, arbitrarily, to 25 ugIL.
This puts a total of 35 samples into category one.
The second category is where we would hope to find most of the
data. This turns out not to be the case. When the criterion for
agreement is arbitrarily set at a factor of two
TOX by GC/MS
i.e. 0.5 — Method 9020 TOX — 2
it found to contain only sixteen samples.
Category three, where the Method 9020 TOX value is greater than
double the TOX by GC/MS results, contains most of the data. This
category contains 60 of the 115 samples. Of these 60 samples, 36
had TOX by GC/MS results below 25 ug/L and Method 9020 TOX results
above 100 ug/L. There are a number of things that can cause Method
9020 TOX values to be higher than TOX by GC/MS values for a given
sample. Some of the most likely causes are:
1. High levels of inorganic chloride in the sample. At least
15 category three samples are from areas where saltwater
intrusion into the groundwater is a strong possibility. The
potassium nitrate wash can only remove chloride effectively
if the inorganic to organic chloride ratio is less than
20,000 to 1.0
2. High levels of inorganic sulfides in the sample. Six
category three samples had a strong hydrogen sulfide odor.
When they were analyzed the silver electrodes in the
titration cell turned black.
3. Chlorinated organic compounds not readily detected by
GC/MS. Highly polar compounds are not readily purged or
extracted from water. Other compounds may not pass through
the GC column. These and other compounds should be seen by
TOX more readily than by CC/MS.
4. Organic compounds below the GC/MS detection limit. If all
of the halogenated organic compounds that are targeted by
Methods 8240 and 8270 were present just below their nominal
detection limits of 1 to 20 ugiL, the TOX by CC/MS would be
about 200ug CL/L. A combination of target and non—target
compounds below the detection limit could cause a
substantial TOX with no associated GC/MS observations.
1—9
-------
5. Halogenated organics adsorbed onto fine particles suspended
in the solution. These compounds will not be purged or
extracted efficiently for GC/MS analysis, but they will pack
at the top of the carbon column and be combusted in the TOX
analyzer.
6. ContamInation of the carbon columns through contact with
inorganic chlorine after the nitrate washing step.
7. Contamination of the activated carbon by halogenated organic
compounds as vapors during column preparation.
8. Other sources of laboratory or field contamination such as
contaminated sample containers, dilution water or nitrate
wash solution.
Only four of the 115 samples fall In category four, where Method
9020 TOX values are less than half of the corresponding CC/MS
values. TOX is often used as a screening mechanism to identify
contaminated wells. When viewed this way, category three samples
could be considered as false positives and category four samples as
false negatives. Fal8e positives are must less serious than false
negative because a positive result will generally be followed by
additional, more rigorous analyses. A false negative result could
lead to concluding that a hazardous situation is non—hazardous. A
closer look at the four samples in category four shows that they
would not give false negative results. One sample had over 40,000
ug/L of Method 9020 TOX and over 100,000 ug/L of TOX by CC/MS. This
well would certainly receive further attention. The second and
third samples in category four were both collected during the same
sampling trip and the only compound found by CC/MS was methylene
chloride in both samples. The latter sample was labelled as a trip
blank. This indicates that the results for these samples are due to
contamination of the sample containers. (This may also explain the
five category three samples from these sites.) The explanation of
the fourth category four sample is not as clear as for the other
two. The most likely explanation is that the 58 ug/L of
tetrachloroethylene found by CC/MS was introduced into the sample
through contamination; another sample from the same site contained
210,000 ugIL of tetrachioroethylene.
CONCLUSION
Comparison of Method 9020 POX values with calculated POX by GC/MS
values for 115 samples from 16 sites shows that Method 9020 results
do not agree with the level of halogenated organic compounds that
are found by CC/MS. Most of the samples analyzed were found to have
higher POX levels by Method 9020 than by GC/MS. The work of
Williams and co—workers’ 2 suggests that Method 9020 was detecting
halogenated compounds that were not found by CC/MS at the site that
1—10
-------
Table 2
TOX calculated from GC/MS Data for Site Three
Sample 1 Sample 2 Sample 3 Sample 4
Compound %Cl conc ugJL conc ugIL conc ug/L conc ug/L
Determined by GC/MS meas TOX meas TOX meas TOX meas TOX
1,1,1—Trichioroethane 80 U U 2 1.6 U U 2 1.6
Chlorobenzene 31 U U 2 0.6 3 0.9 U U
t—1,2—Dichloroethene 73 U U U U 5 3.7 U U
Dichloromethane 83 24 20 U U 4 3.3 21 17
Trichioroethene 81 U U U U 2 1.6 U U
1,3—Dichlorobenzene 48 U U 2 1.0 U U U U
TOX by GC/MS 20 3.2 9.5 18.6
TOX by Method 9020 a 92 J 264 157 J 24
meas. — measured by CC/MS.
TOX — calculated from concentration measured by CC/MS.
U — undetected
a. The differences between Method 9020 TOX values and TOX by CC/MS are
discussed in the body of this paper.
b. If there had been any brominated compounds in this table, they would
have been calculated using the atomic weight of chloride.
1—11
-------
they investigated. Our work indicates that the observation of
higher values for Method 9020 TOX than for TOX by GC/MS is common
for groundwater samples. It is reasonable to assume that the
differences between the two would be attributable to the nonvolatile
TOX increment observed by Williams and co—workers.
The next step In the evaluation of Method 9020 as a technique for
ground water monitoring is to find the major causes of the “false
positives.” Halogenated organic compounds that are missed by GCIMS
should be identified. It is also necessary to determine what the
effect of sediments is on Method 9020 TOX results. The filtering of
samples must be examined as a sample treatment option. Once these
Investigations have been performed It will be possible to perform a
more thorough analysis of Method 9020.
REFERENCES
Stevens, A A.;Dressman, R. C.;Sorrell, R. K.;Brass, H. J. Organic
Halogen Measurements: Current Uses and Future Prospects, Jour.
AWWA 1985 , 77, (4), 146.
Kuhn, W.;Sontheimer, H. Several Investigations on Activated
Charcoal For the Determination of Organic Chioro—Compounds. Vom
Wasser 1973 , 41, 65—79.
Kuhn, W.;Sontheimer, H. Analytic Determination of Chlorinated
Organic Compounds With Temperature—Programmed Pyrohydrolysis. Vom
Wasser 1974 , 43, 327—341.
Dressinan, R. C. ;McFarren, E. F. ;Symons, J. M. “Evaluation of the
Determination of Total Organic Chlorine (TOC1) in Water by
Adsorption Onto Ground Granular Activated Carbon, Pyrohydrolysis,
and Chlorine—Ion Measurement”; Proceedings AWWA Water Quality
Technology Conference, Kansas City, MissourI, December 1977.
Takahashi, Y,;Moore, R. T. “Measurement of Total Organic Halides
(TOX) In Water by Carbon Adsorption/Microcoulometric
Determination”; Presented at Am. Chem. Soc. National Meeting,
Honolulu, Hawaii, April 1979. Also reported In Takahashi, Y. “A
Review of Analysis Techniques for Organic Carbon and Organic
Halide in Drinking Water”; USEPA publication EPA 570/9—84—005,
1984.
Billets, S.;Llchtenberg, J. J. Method 450.1 — Interim, Total
Organic Halide, USEPA Environmental Monitoring and Support
Laboratory, Cincinnati, Ohio, 1980.
Method 9020, Total Organic Halide, “Test Methods for Evaluation
Solid Waste”; USEPA, Washington, DC, 1982, SW—846.
1—12
-------
USEPA. Hazardous Waste Management System, Part VII: Standards and
Interim Status Standards for Owners And Operators of Hazardous
Waste Treatment, Storage, and Disposal Facilities. Fed. .
1980 , 45, (98) 33239. —
Method 8240, GC/MS Method For Volatile Organics, and Method 8270,
GC/MS Method For Semivolatile Organics: Capillary Column
Technique. “Test Methods For Evaluating Solid Waste”; USEPA,
Washington, DC, 1982; SW—846.
Dressman, R. C.;Stevens, A. A. The Analysis of Organohalides in
Water — An Evaluation Update. J.—Am. Water Works Assoc. 1983 , 75
(81), 431. —
Purgeable Organic Halogen is obtained by purging a ten milliliter
sample for ten minutes and passing the purge gas through the
pyrolyals and detection system of a TOX analyzer.
Williams, D. T.;Coburn J. A.;Bansci, J. J. Study of Total Organic
Halogen as a Means to Detect Groundwater Contamination.
Environment International 1984 , 10, 39.
Dreasman, R. C.;Najar, B. A.;Redzlkowski, R. The Analysis of
Organohalides (OX) as a Group Parameter. “Proceedings AWWA Water
Quality Technology Conference”; Philadelphia, Pennsylvania, Dec.
1979.
USEPA. RCRA Ground Water Monitoring Technical Enforcement Guidance
Document (TEGD). Office of Solid Waste and Environmental
Response, Washington, D.C., September 1986; OSWER—9950.1
1—13
-------
NONVOLATILE ORGANICS AS PROBES FOR CONTAMINATED GROUND WATER PLUMES FROM
HAZARDOUS WASTE MANAGEMENT FACILITIES
Robert, D. Stephens, California Department of Health Services, Hazardous
Materials Laboratory, Berkeley, California; Nancy B. Ball, Thomas S.
Fisher, Raimund Roehi, and William M. Draper, California Public Health
Foundation, Berkeley, California
ABSTRACT
characterization of ground water quality in the vicinity of hazardous waste
disposal facilities is essential in defining site integrity and delineating
plume migration. Ground water contamination by organic compounds at such
facilities is typically indexed by analysis of volatiles or semi—volatiles,
because these parameters can be determined using established EPA analytical
protocols.
Ideal probes for these purposes, however, are substances which are: 1)
unique constituents of disposed wastes; 2) persistent; 3) nonvolatile; and
4) highly mobile or conservative in ground water. Accordingly, certain
non—conventional pollutants may well provide more suitable and accurate
tracers than the current target analytes. This report describes studies on
the identification and quantification of one such non—conventional
pollutant, p—chlorbenzenesulfonic acid (PCBSA), in the BKK landfill ground
water plume.
Ground water adjacent to the BKK Landfill, a major hazardous waste disposal
site in southern California, was found to contain TOX above 200 mg/L, with
less than 4% of the organic halogen accounted for Subsequent analysis of
the ground water samples by liquid chromatography/ nass spectrometry and by
ion chromotography with conductivity and UV absorbance detectors revealed a
plume of PCBSA that accounted for approximately half of the observed ‘lOX.
PCBSA, which occurs as an anion above pH 0.5, was associated with DIYI
manufacturing wastes disposed at the site for many years.
In a further application, PCBSA determinations were instrumental in
resolving the contested mechanism of ground water contamination by volatile
priority pollutants (e.g., TCE, vinyl chloride). Ground water contamination
by these volatile compounds could occur by at least two distinct mechanisms
including: 1) direct leaching of the waste prism during infiltration of
surface water; 2) dissolution of gaseous components from landfill gas
contracting ground water. Engineering controls currently in use at the site
focus on landfill gas extraction based on the assumption that the latter
mechanism predominates. The occurence of PCBSA in the plume provides
unequivocal evidence that leaching of waste prism does contribute to ground
water contamination.
1—15
-------
GAS CHROMATOGRAPHY MATRIX ISOLATION FOURIER TRANSFORM
INFRARED (Gc/MI-IR) SPECTROSCOPY FOR MONITORING
AIR POLLUTANTS
Billy J. Fairless, Thomas T. Holloway, Harry E. Kimball, Richard W.
Tripp, Jody L. Hudson, U.S. Environmental Protection Agency, Kansas
City, Kansas
ABSTRACT
Procedures will be described for collecting and analyzing ambient
air samples on solid adsorbents and in stainless steel canisters
from the vicinity of a suspected source of air pollution. Variables
involving design of the sampling network, collection of
representative samples, sample analyses by GC/MI—IR and other
methods, estimations of data quality and the significance of the
results will be presented. Infra—red spectra from low nanogram high
picogram quantities of matrix isolated pollutants will be shown and
related to vapor phase spectra.
INTRODUCTION
Analytical procedures to monitor for the toxic or non—criteria
pollutants in ambient air have not been studied nearly as much
have the procedures to monitor for the six criteria pollutants.
However, some non—criteria pollutants present in the air are toxic
or carcinogenic or both. The media have become more active in
recent years in reporting air spills, which is one of the causes of
a greater public awareness of both existing and potential air toxic
problems. One result of this increasing public awareness in Title
III of the Superfund Amendments and Reauthorization Act (SARA) of
1986, which requires the Environmental P otection Agency to develop
a more comprehensive air toxics program.
One of the reasons that more work has not been done on air toxics is
that procedures to collect and analyze samples for the extremely low
concentrations required have not been available. Another reason is
that criteria for evaluating the significance of the concentrations
found in the air are also usually available. Any procedure used to
monitor ambient air for a toxic material must be specific for that
material must be sufficiently sensitive to accurately measure the
low concentrations of interest, and must generate data that can be
related to the applicable toxicity criteria.
Several kinds of procedu es are currently used to monitor for
selected toxic pollutants. These procedures generally utilize
bubblers, solid sorbents, or whole air collection devices. Most of
these have one or more deficiencies. Bumblers can be adapted to a
wide range of compounds, but they are difficult to use in the field,
and therefore, expensive when labor is included in the cost. They
also usually have higher detection limits than competing methods.
1— 17
-------
Solid adsorbents require an extensive amount of work to ensure that
the compounds of interest are adsorbed quantitatively, do not break
through under field conditions and then can be desorbed in the
laboratory in a reproducible manner. Many of these procedures are
also less rugged than desired. Grab samples collected in fixed-
volume vacuum containers, or variable volume containers, do not
always allow the collection of sufficient sample to measure the low
concentrations of interest.
We have developed a series of ambient air monitoring procedures and
have completed several studies using polyurethane (PUF) and
stainless steel spheres for sample collection with gas
chromatography and matrix isolated infrared spectroscopy (GC/MI-IR)
for sample analyses. These procedures complement existing
procedures described above and appear to be superior to those
procedures in some respects. We believe this is the first instance
in which GC/MI-IR has been used as a primary analytical tool in a
major ambient air study. Based on our experiences there will be
many such studies in the future.
PROCEDURES
The monitoring network is designed to provide sufficient data to
meet a specific objective. Usually the objective is to determine
whether or not the average concentration of a given pollutant at a
specific site is above an acute or chronic criteria value, or
whether or not the average downwind concentration of a given
pollutant is higher than the average upwind concentration at a
reasonable confidence level. The procedures we are currently using
to determine the number of samples we need to collect a each
monitoring location in the network have been described before. The
most common result is that between seven and fourteen samples are
needed from each sampling location in the network. This means that
a relatively large number of samples are required to satisfy the
network completeness criteria. Therefore, selection of an
analytical method with a correspondingly high sample volume
capability is also required. Once the data are collected, it is
important to review all assumptions made during design of the
network to be sure the valid data actually collected are
sufficiently complete to meet the study objective. Figure 1 is an
illustration of the general approach for selecting the number of
samples to be collected at each monitor location in the network.
The procedures we use to collect semi—volatile materials have been
described by Lewis and Jackson. Our experience has primarily been
with the dioxins and with the polychiorinated biphenyl congeners or
PCB’s. A large volume of air (350 cubic meters per day) is drawn
through a particulate filter and a polyurethane foam plug. The
exposed filter and plug are extracted together in a soxhiet. The
extract is cleaned using conventional chemistry and analyzed by gas
chromatography, gas chromatography—mass spectroscopy and/or by
GC/MI—IR as described below.
1—18
-------
The procedures we use to site a sampler are those described for
locating a Hi-Vol sampler when it is being used to monitor for total
suspended particulates in the air. Therefore, samples are
collected in the breathing zone, away from physical structures, and
are intended to be representative of ambient air.
For volatile pollutants, battery-operated diaphragm pumps are used
to meter ambient air into clean 6—li er stainless steel spheres to a
pressure of approximately 2 atm. The spheres are cleaned by
evacuating with a vacuum pump, filling with clean air and evacuating
again. They can be heated during the cleaning process, but this has
not been necessary to date based on the analyses of field blanks.
Although the battery packs are capable of operating the pumps for
periods of more than 24 hours, our studies to date have been limited
to 8 hours so the data could be more directly related to the
applicable criteria. Figure 2 is a schematic diagram of the
canister sampling apparatus.
For sample analyses, 1,000 milliliter aliquots are taken from each
canister at a sampling rate of 7 50 mi/mm. with a mass flow
controller and metal bellows pump. The sample is transferred to a
manifold which is open to the atmosphere to insure atmospheric
pressure. A 500 ml subsample is simultaneously withdrawn from the
manifold at a rate of approximately 25 mi/mm. and passed through a
dryer to a cold trap which is filled with glass beads and maintained
at a temperature of minus 150 degrees centigrade. Once the sample
is collected in the cold trap, the gas chromatograph carrier gas
(0.6% Argon in Helium) is directed through the cold trap and the
cold trap is then quickly (about 30 seconds) heated to 150 C°. This
transfers the sample to the capillary column of a Mattson GC/MI-IR
Cryolect analytical system. Any stable column that separates the
compounds of interest at a flow of less than 3 mi/mm. may be used.
We usually see a 30M X 0.32 mm dimethylpolysiloxane column with a 1
urn film thickness. The column temperature is maintained at 5 C 0 for
10 minutes, raised to 80 C 0 at a rate of 5 C°/min. and then raised
to 160 C 0 at a rate of 10 C°/min. A splitter at the end of the
column directs approximately 20% of the effluent to a flame
ionization detector and the remaining 80% to the Cryolect where the
pollutants are 4 trapped in crystalline argon at approximately 13
Kelvin degrees. Once an entire chromatograph has been trapped on
the rotating Cryolect gold-plated cylinder, the cylinder is
positioned (via computer) to place each significant peak into the
infra—red beam for collection of the infra—red spectra. We
typically average 32 spectra per compound at a resolution of 4
reciprocal centimeters. These conditions are sufficient to give a
high-quality spectra for approximately one nanogram of each
component injected into the chromatograph. However, in extreme
cases we have averaged over 10,000 spectra which improves the
sensitivity to picogram levels. Figure 3 shows a schematic diagram
of the analytical process.
1—19
-------
RESULTS
Figure 4 shows the first 25 minutes of a typical gas chromatogram of
volatile compounds from an ambient air sample that was collected in
stainless steel canisters downwind of a chemical plant that
manufactures chlorinated hydrocarbons. Figures 5—7 show three of
the resulting infra-red spectra from the ambient samples and
compares them to the respective library spectra. As can be seen,
the sample spectra match the library spectra. They have good
resolution and document the excellent sensitivity that is
characteristic of the GC/MI-IR method. Figures 8-13 show spectra of
three compounds that are of inferior quality, but which are still
considered to be good enough to support a tentative identification
in each case. These spectra are shown to illustrate one of the
deficiencies still remaining in our procedure for volatile organic
compounds. So far, we have been unable to remove all water vapor
from the sample without losing some of the compounds of interest.
This necessitates a spectral subtraction of the relatively high
background and results in lower quality spectra for some compounds
at concentrations where we would otherwise expect to obtain higher
quality spectra. These spectra also illustrate the practical
detection limit using a one-liter sample from a stainless steel
canister and a GC/MI-IR analytical method. Quantitative values are
obtained by comparing the FID response to a calibration curve
obtained by putting calibration standards through the entire
analytical process.
Since the library of argon matrix isolated infrared spectra is very
limited, we have attempted to compare spectra resulting from field
samples with vapor phase library spectra. The results usually
provide either an obviously unreasonable match or what appears to be
a correct identification of the unknown. Figure 14 is one such
example which compares to vapor phase spectra from the library with
an argon matrix spectra from an ambient air sample.
In addition to collecting a whole air sample in a sphere at each
site, a duplicate TENAX sample is also frequently collected. We
collect TENAX samples both to obtain lower detection limits for some
compounds and as a method that compliments GC/MI-IR. The TENAX
samples are thermally desorbed and analyzed by packed column gas
chromatography and low resolution mass spectroscopy. Table 1 shows
a sun nary of those compounds identified by each technique for the
study described above.
Figure 15 shows the GC/MI-IR chromatograni for Archiors 1221, 1016
and 1254. Figures 16-19 show infra-red spectra for two of the many
congeners (biphenyl and 2-chlorobiphenyl) at different
concentrations. Again, the resolution is good and the sensitivity
is much better than that of other infra-red procedures. Figures 20
and 21 are comparisons of the linearity of the responses between the
flame ionization and infra—red detectors for two of the chlorinated
1—20
-------
biphenyls. As can be seen, the linear range of the flame ionization
detector is much larger than that of the infra—red detector for both
compounds. Not obvious from those figures is the fact that the
infra—red detector is more sensitive than the flame ionization
detector for these and other polychiorinated biphenyl compounds we
have analyzed. Figure 22 demonstrates that the GC/MI-IR absorption
versus concentration curve is linear when less than 100 nanograms
are injected into the instrument.
Brasch has demonstrated the utility of electronically subtracting
GC/MI—IR library spectra from the spectra obtained from unresolved
chromatograph peaks and then matching t e remaining or “difference”
spectrum to a known library spectrum. The technique was used to
obtain a positive identification of each of the tetrachiorinated
dibenzofuran isomers which could not be resolved by capillary gas
chromatography. We have used a similar technique to obtain a
tentative identification of a compound where no sample or library
spectrum was available and a positive identification could not be
made from GC/MS data. Figures 23 and 24 show the library spectra of
dimethyl sulfate and of diethy/sulfate. Figure 25 shows a gas
chromatogram of an environmental sample. Figure 26 shows the
spectrum of an unknown compound (largest peak in the chromatogram)
from that sample. Figure 27 is a spectrum formed by electronically
summing the spectra of dimethyl and diethylsulfate. The facts that
the retention time of the unknown was between that of dimethyl and
diethyl sulfate, the GC/MS appeared consistent and the similarity of
the IR spectra shown in FIgures 26 and 27 was the basis of our
conclusion that the unknown compound was methylethylsulfate.
Figures 28 and 29 show the infra—red spectra of carbon—13 labeled
and natural 2,3,7 ,8-tetrachlorodibenzodioxin (2,3,7 ,8—TCDD). The
matrix isolated spectra of all 22 Carbon 12 tetrachlorinated
dibenzodioxin were obtained by Brasch and have been reported by
Gurka et. al. However, the authors did not report detection
limits. We are specifically interested in whether or not GC/MI-IR
is sufficiently sensitive to measure concentrations of dioxin in
ambient air in the one to five picogram per cubic meter range which
we find occasionally in ambient ai samples from the vicinity of
superfund site clean—up operations. Figure 30 shows an infra—red
spectra of one of the less toxic dioxin congeners from approximately
150 picograzns of material injected into the gas chromatograph.
Since we have demonstrated that at least 1,000 cubic meters of air
can be relia ly sampled using polyurethane in a PUF sampler without
breakthrough, it appeared that a total method detection limit of
0.1—0.2 picograms per cubic meter was a feasible target. Therefore,
a polyurethane foam plug was spiked with 2,3,7,8—
tetrachlorodibenzodioxin (2,3,7,8-TCDD) and 320 cubic meters dioxin-
free urban air were drawn through the plug. The foam plug was
extracted and the extract was divided in half to evaluate different
clean—up procedures. One micoliter (approximately 4 nanograms) of
1—21
-------
the extract was injected into the instrument. Figure 31 is the
resulting infrared spectrum. It is clearly of sufficient quality to
support a qualitative and quantitative analysis of that particular
dioxin isomer.
CONCLUSION
It is apparent that the GC/MI—IR analytical procedure, when used as
described above, will provide reliable qualitative and quantitative
information for many of the toxic pollutants in ambient air.
Collection of samples in stainless steel spheres and GC/MI-IR
analytical procedures provide several advantages over other
available procedures. Probably the most important advantage is
greater isomer specificity at much lower concentrations than is
available using other procedures. Another major advantage is the
additional confirmation of tentative identifications made by other
(GC/MS) techniques. There are, however, also many areas needing
improvement. Among these are the ability to eliminate the
interference from water vapor, improvements in software to improve
sample analysis times, and the development of a larger library of
matrix isolated spectra. It is the author’s opinion that these
deficiencies will all be relatively easily solved with time, and
that the GC/MI-IR procedure will become another powerful tool for
environmental monitoring.
1—22
-------
FIGURES
Figure 1 - Graph of the probability that the mean of a given number
of measured concentrations will be different from the true mean at
the 95% confidence level. Typical curves for different kinds of
data are shown and assume a normal distribution.
Figure 2 — Schematic diagram of the apparatus used to collect
ambient air samples in stainless steel canisters.
Figure 3 — Schematic diagram of the apparatus used to transfer a
measured volume of air from a stainless steel canister into the
GC/MI-IR.
Figure 4 - Gas chromatograin from the GC/MI-IR of an ambient air
sample collected downwind from a plant that synthesizes chlorinated
hydrocarbons.
Figure 5 - Argon matrix isolated infrared spectrum of approximately
630 ng of niethylene chloride from an ambient air sample relative to
the library spectrum.
Figure 6 - Matrix isolated spectrum of approximately 650 nanograms
of chloroform from an ambient air sample.
Figure 7 - Matrix isolated spectrum of approximately 400 nanograins
of carbon tetrachioride from an ambient air sample.
Figure 8 - Matrix isolated infrared spectrum of 29 nanograms of 1,2-
dichioroethane from an ambient air sample. The noise is caused by
water vapor.
Figure 9 - Comparison of the infrared spectrum of 1,2-dichioroethane
from an ambient air sample with the library spectrum.
Figure 10 — Infrared spectrum from a compound believed to be
chioroethane in an ambient air sample.
Figure 11 — Comparison of the infrared spectrum of chioroethane
found in an ambient air sample with the library spectrum.
Figure 12 — Infrared spectrum of 1,1,1—trichioroethane found in an
ambient air sample.
Figure 13 — Comparison of the infrared spectrum of 1,1,1-
trichiorethane found in an ambient air sample to the library
spectrum.
1—23
-------
Figure 14 - Comparison of the matrix isolated infrared spectrum of
tetrachioroethylene found in an ambient air sample with a vapor
phase spectrum from the library. Note the better resolution of the
matrix isolated spectrum resulting from less translational and
rotational broadening of the vibrational modes.
Figure 15 — Gas chromatograms from the GC/MI—IR of three aroclor
mixtures. Those peaks caused by the commercially available
polychiorinated biphenyl congeners have been identified using the
GC/MI—IR library search routines. Work is in progress to obtain
spectra of all 209 congeners.
Figure 16 - No legend.
Figure 17 - No legend.
Figure 18 — No legend.
Figure 19 - No legend.
Figure 20 - Graph showing the responses of both the flame ionization
(FID) and infrared detectors on the GC/MI-IR as a function of the
amount of H—chlorobiphenyl injected into the instrument. Note that
the infrared detector becomes nonlinear at approximately 150
nanograms.
Figure 21 - Graph showing the responses of both the FID and IR
detectors as a function of the amount of 2-chiorobiphenyl injected
into the GC/MI—IR.
Figure 22 — Graph showing that the infrared detector is linear with
concentration at low concentrations. The infrared detector is more
sensitive to those polychiorinated biphenyl congeners we have
analyzed than the FID detector on the GC/MI-IR. However, the IR
detector is less sensitive for the PCB congeners than it is for most
other compounds we have run.
Figure 23 — No legend.
Figure 24 - No legend.
Figure 25 - No legend.
Figure 26 — No legend.
Figure 27 - No legend.
Figure 28 — An infrared spectrum of carbon-13 labeled 2,3,7,0-
tetrachlorodibenzodioxin (2,3,7,8-TCDD). There are no peaks above
2,000 reciprocal centimeters.
1—24
-------
Figure 29 — Infrared spectrum of 2,3,7,8—TCDD. There are no peaks
above 2,000 reciprocal centimeters.
Figure 30 - A comparison of the spectrum obtained from 156 picograms
of 2,3,7,8-TCDD with the library spectrum.
Figure 31 — An infrared spectrum of 2,3,7,8—TCDI) from a PUP sampler.
Approximately 350 cubic meters of urban air were drawn through the
PUP prior to extraction, cleanup and analyses.
1—25
-------
160
I
I
Single Sided “t” Test
At 95X Confidence Level
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
U U I U U U U U U —- U — — U- -— 1 I I I
8 9 10 11 12 13 14 15 16 17 18 19 20 30 40 120
(Samples)’
w
-4
Li
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
150
140’
130
120
110’
100
90
80
70 U
60
50
40
30
0
uJ
0
>‘
-o
C
-o
0
0
20
10
0
0
U —I I 1 I I I
1 2 3 4 5 6 7
Number of Data Points
-------
Teflon Tubing
/‘
ori/cff
vu I V C S
/
/
Fl.. :
i’,. ;
I, ,
I .’
It
114
Pu nip
with a vitor,
diciphracnn &
s head
c\J
LU
LU
-------
FIGURE 3
Matrix Gas
Flow Controller
2 m I/mm
LN 2
Sample
Flask
Me t J
Be ho ws
Pump
Flow Controller
50 mI/mm
Drying
Apparatus
?4ani fold
Dewar
Vacuum
Pump
1—28
-------
from Ambient
romatogram Air Sample
lected in stainless steel cannister
downwind of a Chemical Plant
Ch
col
(l4
(3
0
(3
(3
(3
•1
(3
0
0
¼0
H
c)
(3 (3
0
0 (3
- 0
.i . h
(3 0
o
I- - • •
0 (.1
1
0 •
‘3
o I
0
()
‘3
(3
.
-I-
‘3
0
0
‘3
0
L
(3
I —
—4
Retention time
-------
Sample
Methyl ene
Spectrum vs
Chloride
Library
Spec t rum
Library Spectrum
Sample Spectrum
Wavenumb e r
-------
Chloroform
Sample
Spectrum vs Library
Spec t rum
Library Spectrum
Sample Spectrum
L
Wavenumber
-------
Carbon
Tetrachjoride
S’7rlipIe Spectrum vs Library Spectrum
.4
N )
Sample Spectrum
Library Spectrum
1400 1 200 1Q00 800
W v numb r
T1
I-
rn
-------
Sample EJR28 126 * (29 ,ug/M 3 )
r )‘tt
t.). L/C
2. 020
0. OU
0. 016
0.014
0.012
0. 010
0. 00
0. 006
0. 004
0. 002
0. 000
—0. 002
— GC/MI —IR Sp e c t r u in
c i ,
C : ’
L
0
(1
-Q
1400 1200 1000 800
-Ti
I Ti
Co
Wavenumber
-------
1, 2—Dichloroethane
Sample Spectrum vs Library
Spec t rum
q)
()
-Q
0
(I )
I —I
Library Spectrum
Sample Spectrum
1400 1200 1000 800
Wavenumb e r
-n
-4
C,
rn
-------
‘GTfl,L) 1 e
0 .
(1.
U.
0.
.1.
r t,t
,. ,
tI’
0
i. OlE ,
0.
0
f P4t
V. LdIt
0. 010
C I. O0!
0.
‘I.
004
EJR28 126 *
1400
(64 p g/M 3 ) — G C/MI —IR
1200
Spec t runi
1000
q)
U’
‘4
.z
-n
1
Q
Wavenumber
-------
¶1
Samp I e
C/i Jo roe thane
Spec t rum
vs Library
Library Spectrum
1200
Sample Spectrum
Spectrum
L
C)
1400
-n
-4
C,
-4
1000
Wavenumber
-------
L
L )
—1
(;.
0. 04C
‘ .
0. 030’
0. ()2 ’
L’. 020’
0. 0 15
0.010’
0. 005
0. 000
-n
-I
P h
Sarnpl e EJR2 8 126 * (39 p g/M 3) — GC/MI—IR Sp e c t r ‘i in
1200
1000
Wavenumbe r
800
-------
Sample
1, 1,
Spec
trum
1—Trichloroet/-iane
vs
Library
Spec t rum
Library
Spec t rum
Sampie Spectrum
1400 1200 1000 800
rn
(h.)
Wavenumber
-------
S ample
w
=
U-
()
L
(I)
-Q
ckz
Tetrachioroethylene
Spectrum vs Vapor Phase Library Spectrum
1400 1200 1000 00
Wa venumb e r
-------
Polychiorinated Biphenyl Mixtures
-4
U..
2—ChIo
Biphenyi
4—chiorabiphenyi
2, 4 —DichiorobIphenyI
0
.1)
L
Arocior—1221
2—chi orob Jphenyi
2, 2 3, 3’—TetrochiorobiPheflYl
Aroclor—1016
C
-ct.
/
2, 2 5, 5 —Tetrachiorobiphenyi
2, 2 3, 3 4, 4 —hexachiorobipheriyi
/
Arocior—1254
j )
/
Retention time
-------
2—Chiorobip
Aroclor—1227
henyl
nano grams)
11.
c . o
15
0. c
V.
0. U0 i
Wavenumber
-------
2—Chiorobiphenyl — Aroclor 1221 2300 ng/.vl
(140—
-l
0.30—
Q)
- o. 0—
0
C l )
-Q
0. 10—
A A A - ____ L
0.00—
I
4000 3500 3000 2500 2000 1500 1000
Wavenumber
-------
Biphenyl — Aroclor 1221 2300 ng/ul
C)
L i i
0.30—
0. 25
)
( 0.20 —
L
J.a
-C 1
0. 10 —
0.05 —
4000 3500 3000 2500 2000 1500 1000
Wavenurnbe r
-------
Arocior—122
Nanograrns
3500 3000 2500 2000 1500
Biph
en
1
yl
13
(
(,.
i , .
‘,. ,d.
0. 0
0.010
. -1
U-
L
0
)
0. 005
- Zr
0. 000
4000
1000
Wavenumber
-------
46Th
o rob
phen
ISO
Nano grams
-------
H
C
0.5
0 0.4
2—Ch Jo rob iphenyl
40 0.2
20 0.?
0 0.0
0
300 400
Nano grams
I .
r
4,
C’4
‘ -4
700
200
800
700
900
-------
H
(b
L.
0. 9
i. 2
C . 1
(.).
w
-l
Lj.
2 Chiorob
iphenyl
•-zr
0 10
20
30
40
5°
Co
Nano grams
70 80
10(1
-------
Di
me
Su
te
Wavenumbe r
-------
Die thyl
Sul
f a t
/ : r;niL;i
-------
Sample 1K935
794
Sul
fate
Methylethyl
t
L
H
H
N
c-)
çr
Retention time
-------
0. 30
0. O
C. 7 )
ó’O
V .
I.
I P
L .
d. 1 1
‘I.
(\j
U i
-4
U-
Sample No. 1K935794
‘1
I.. )
C::
4000 ‘500
2 U
2000 1500 1 0
Wavenuri’ib e r
-------
Overlay of
Diethyl
Su 1 fate
and
Dimethyl Sulfate
‘. ) i
iJJ
U-
c. 7 i
0.
0. . 0
L .
0.40
0. ‘0
0. 20
0. 10
0.00
4000
Waveriumber
-------
2,3,?,8-TCDD Carbon 13 be1ed
2ø@ t9@@ l8 1.? 1Ø U30@ l5 @ 1 4Ø l3 @ U@ t1. @ U Ø3
9 1 ç ø
4 v e n u. m b r
dxn 9
0.
C
e
0. 50
0.
0.
L
0.1
H—
-i fl
-------
dxri.. 2
2,3,7, 8-TC.OO
L O
r
b
n
C
e
. 80
0.
0.60
0.5i
0. 30
0.20
0.10
tr
‘-.4
1
tgoo IS O 1?4 0 l& 0 1500 140 1300 1200 1 1 0 1000
W v e r, U rn b e
.2
-------
FIGURE 30
di o.4 1.2 ,3, 4-ICOD
. 1 — I I ___ —
R Library Spectrum
0.50—
0. 40—
0.30—
0.20—
0. 10—
156 Picograms
b 0.0014—
S
0
r 0.0012—
b
0.0010—
C
e
0.0008—
0.0006—
0.0002— _____ ________________________________
4ØØØ 3500 3000 2500 2000 1500
Wove number
ci 7
1—55
-------
cY)
LU 0.012
CD
( . C I 10
0. 008
(-)
0. 006
iq 0.004
0. 002
0. 000
2, 3, 7, 8—TCDD (4 Nano grams)
exfracled from PUF
‘C
4000 3000 2000 1500
1000
Waveriumber
-------
REFERENCES
1. U.S. EPA, Code of Federal Regulations, Volume 40, Part 50,
1986, Page 526.
2. Superfund Amendments and Reauthorization Act of 1986, Public
Law 99—499, October 17, 1986.
3. Riggin, R.M.; Winberry, William T., Jr.; and Tilley, Norma V.
Copendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air (EPA-600/4—84—041), Center for
Environmental Research Information, 26 West St. Clair Street,
Cincinnati, Ohio 45268.
4. Bourne, S.; Reedy, G.; Coffee, P.; and Mattson, D.; Matrix
Isolation GC/FTIR, American Laboratory, June 1984.
5. Fairless, Billy; Bates, D.; Hudson, J.; Kleopfer, R.D.;
Holloway, T.T.; Morey, D.; and Babb, T.; “Procedures Used to
Measure the Amount of 2,3,7,8-tetrachlorodibenzo—p-dioxin
(2,3,7,8—TCDD) in Ambient Air near a Superfund Site Clean—up
Operation,” accepted for publication, June 1987, Environmental
Science and Technology.
6. Lewis, R.G. and Jackson, M.D., “Modification and Evaluation of
a High—Volume Air Sampler for Pesticides and Semivolatile
Industrial Organic Chemicals,” Anal. Chem., 54, 592-594 (1982).
7. Bates, Dale and Holloway, Thomas, “Environmental Monitoring and
Compliance Branch Operating and Quality Assurance Procedures
Manual,” U.S. EPA, Region VII, 25 Funston Road, Kansas City,
Kansas 66115 (1982).
8. Gurka, D.F.; Brasch, 3.W.; Barnes, R.H.; Riggle, C.J.; and
Bourne, S.; “Micro—diffuse Reflectance and Matrix Isolation
Fourier Transform Infrared Techniques for the Identification of
Tetrachlorodibenzodioxins,” Applied Spectroscopy, 40, 978—991
(1986).
1—57
-------
EVALUATION OF THE FLUX CHAMBER METHOD FOR MEASURING
AIR EMISSIONS OF VOLATILE ORGANIC COMPOUNDS
FROM SURFACE IMPOUNDMENTS
Alex R. Gholson, John R. Aibritton, R. K. N. Jayanty, Center for
Environmental Measurements, Research Triangle Institute, Research
Triangle Park, North Carolina; Joseph E. Knoll, M. R. Midgett,
Environmental Monitoring Systems Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina
ABSTRACT
Enclosure methods have been used to measure air emissions of a
variety of compounds from soils, water, sediments, and living
organisms. A flux chamber method, which employs the enclosure
method, recently has been used to measure organic air emissions from
hazardous waste treatment, storage, and disposal facilities. Using
a simulated surface impoundment facility (SIS), this flux chamber
method was evaluated. The liquid surface of the SIS was enclosed so
that the total emission rate from a liquid surface could be
determined experimentally. Fmission measurements using the flux
chamber method made at several points on the surface were compared
with the emission rate measured for the total surface inside the
enclosure. Both the accuracy and precision of the flux chamber
method were predicted from these measurements. The influence of
sweep flowrate, emission rate, and different organic compounds on
precision and accuracy were investigated.
The results of this study show that a consistent negative bias
exists for all the flux chamber measurements. This bias became
significantly more negative at a low sweep flowrate (2 LImin). The
bias also was found to be compound dependent. Precision was less
than 5 percent under all conditions for the single component studies
and between 6 and 13 percent for the three component study.
INTRODUCTION
The U.S. Environmental Protection Agency (USEPA) has been instructed
to set air emission standards for hazardous waste treatment,
storage, and disposal facilities (TSDF) by the 1984 Resource
Conservation and Recovery Act amendments. In order to determine the
potential health and environmental effects of these air emissions,
methods to measure or predict them are required. The flux chamber
method has been developed and is being used to measure air emission
from TSDF to provide a data base for regulatory decisionmaking and
to validate proposed models used to predict air emissions 1 ’ 2 ’ 3 ’ 4 .
1—59
-------
C er PEP
Figure 1.
Side view of surface impoundment simulator (SIS).
A m
S’S
Sampling Port s
5,
1-60
-------
In an attempt to define the quality of the data being produced by
the flux chamber method, this study was made to determine the
accuracy and precision of the flux chamber method for use on surface
impoundment facilities. The influence of the experimental parameter
of sweep flowrate and the environmental parameters of emission rate
and volatile organic composition on accuracy and precision was
investigated.
EXPERIMENTAL METHODS
The flux chamber design used for this evaluation was developed for
the USEPA by Radian Corporation. 5 It consists of a stainless steel
cylinder with an enclosure area of 0.13 m 2 . The top is enclosed
with a clear acrylic dome fitted with ports for sweep flow inlet,
sample outlet, temperature probes, and a gas exit. The volume of
the enclosure with a 1 in. insertion depth is approximately 30 L.
The design for the surface impoundment simulator (SIS) is shown in
Figure 1. The surface area of the liquid is 1.86 m 3 with an average
depth of 0.46 in. The surface is enclosed In a shell covered with
Teflon with one end opened and the other end attached to the inlet
of a blower. Sampling ports are provided for two flux chambers,
flow monitoring, and sampling the air before the blower.
The volatile organic compounds (voc) measured for this study were
1,1,1—trichioroethane for single component studies and a mixture of
1,1,1—trichioroethane, toluene, and 2—butanone for the three
component study. The organic components were added to the bottom of
the tank. The density of the compounds or mixtures was always
greater than 1.0 g/niL to prevent forming an organic layer on the
surface. Two immersion heaters were located just above the organic
layer to control the tank temperature and increase the convective
mixing inside the tank. A pump was used periodically to increase
the aqueous concentrations of the organic components.
All analyses were performed using a gas chromatograph (GC) with a
flame ionization detector (FID). Flux chamber samples were
collected in a syringe and injected into a gas sampling valve with a
2—niL loop. SIS air samples were collected in syringes or aluminum
gas cylinders and precoacentrated on Tenax that was thermally
desorbed. SIS water samples were collected In glass vials with no
headspace and analyzed by syringe injection.
RESULTS
Single Component Study
Accuracy and precision of the flux chamber method were calculated
from a series of colocated flux chamber measurements made inside
1-61
-------
the SIS containing a single VOC, 1,1,1—trichioroethane. A series of
nine initial emission mea8urements were made under similar emission
conditions and the same flux chamber conditions. The sweep flowrate
was set at the 5 L/min recommended by Radian’s study. Table 1 shows
the variance found between each duplicate measurement. The
coefficient of variance (Cv) was less than 5 percent for all the
measurements except for two. Using a pooled standard deviation, the
precision was predicted from these data to be 3.0 percent. This
value was chosen to be the control condition precision.
Colocated measurements, conducted at night, were made at an emission
rate approximately one tenth of the control condition measurements.
Table II compares the precision calculated for these conditions with
the earlier conditions. No change in the precision was found for
the low emission rate conditions. The variance found at night was
half that found for the control conditions. The improved precision
at night could be due to an effect of sunlight on the performance of
the flux chamber or the effect of the sunlight on the real emission
rate. Emission rates made in full sunlight were found to be highly
variable both from flux chamber measurements and SIS measurements.
All measurements reported here were made in the shade or on overcast
days.
The precision and accuracy of the method also were determined at a
sweep flowrate of 2 L/min and 10 L/min. Table II ahows the results
for three colocated tests at each flowrate. A decrease in the
precision value (improved precision) was found at the higher
flowrate, and a slight increase In the precision value (poorer
precision) was found for the lower flourate. These results indicate
that precision is improved by Increasing the sweep flowrate possibly
due to improved mixing at the higher flowrate.
The average emission rate calculated from a colocated flux chamber
study was compared with the average emission rate calculated before
and after from the total SIS measurement to predict the accuracy of
the method. Table III lists the average percent bias found between
the flux chamber and SIS values for the control conditions and the
other flux chamber conditions studied. The bias was consistently
negative (the flux chamber values lower than the SIS values). The
average bias values were not significantly different within the 90
percent confidence limit (CL) for all the measurements except the
low sweep flow study, which bad a significantly more negative bias.
This suggests that the accuracy of the method decreases at the lower
sweep flowrate.
Three Component Study
The precision and accuracy study was repeated with three VOC in the
tank. These included 1,1,1—trichioroethane, 2—butanone, and
1-62
-------
TABLE I. FLUX CHAMBER PRECISION AT 5 L/MIN FOR
SINGLE COMPONENT SOURCE
Average
Average Surface emission
Sample surface liquid rate,
number temperature concentration (glmin/m 2 )
% CV
1 23 130 14,300
2.1
2 23 130 12,100
0.14
3 22 130 13,200
1.5
4 21 16.2 9,730
4.6
5 21 16.2 9,750
6.0
6 23 16.2 8,910
0.46
7 23 93.6 7,470
4.8
8 24 83.6 7,350
11
9 24 93.6 7,350
4.8
CV Coefficient of variance.
TABLE II. RESULTS OF FLUX CHAMBER PRECISION
STUDY FOR SINGLE COMPONENT
Variable parameter Number of replicates Pr
ecisiona
Control 9
3.0
Low emission rate 3
2.9
Nighttime 3
1.5
2 L/mln sweep flow 3
4.1
10 L/min sweep flow 3
1.4
I ________
aprecislon = 2n
X
1—63
-------
toluene. Comparing the results of nine colocated duplicate flux
chamber measurements with the control of the single component study
reveals several apparent differences. Table IV shows the precision
e8tilnated for each compound. The values are greater than for the
single component control and vary by almost a factor of two between
themselves. The increased variance may be due to the large
difference in emission rates being pooled.
The accuracy of the three component results showed similar
differences when compared with the single component study. Table V
shows the results of bias calculations between the flux chamber and
the SIS emission results. Three significantly different average
bias values were found, and each one except toluene varied
significantly from the single component results. Of special
interest is that 1,1,1—trichioroethane in the mixture had an average
bias of less than one-half of 1,1,1—trichioroethane in the single
component mixture. The total emission measurement bias showed no
significant difference from the single component results.
CONCLUS IONS
The results of the precision and accuracy study indicate that
precise emission measurements can be made using the flux chamber
method. The consistent negative bias found indicates that the flux
chamber method may underestimate the emission rate from a surface
impoundment. Either the flux chamber depresses the emission rate
over the area it covers or the total emission rate may not be
equally distributed over the surface with higher emission at the
sides.
Of the experimental parameters investigated, only daylight and sweep
flowrate was found to affect the accuracy or precision
significantly. Results suggest that sunlight may affect the
variance between two colocated flux chambers and lower sweep flow
rate (2 Lfmin) increases the variance between measurement and
increases the bias. Both precision and accuracy appear to be
compound dependent and are dependent on the matrix.
Studies are recommended to determine the effect of solar intensity
on both the emission rate and the flux chamber method, the cause of
the compound dependency of the flux chamber accuracy, and the reason
for the consistent negative bias found in the results. Plans for
determining the flux chamber precision in the field currently are
being made.
1-64
-------
TABLE III. RESULTS OF FLUX CHAMBER ACCURACY
STUDY FOR SINGLE COMPONENT
Variable parameter
Number of Average bias,
replicates % + 90% CL
Control
9 —45.1 ± 6.4
Low emission rate
2 L/min sweep flow
10 L/min sweep flow
3 —67.1 ± 16.6
3 —81.5 ± 9.6
3 —49.3 ± 8.3
Nighttime
3 —57.2 ± 21.3
CL = Confidence limit
TABLE IV. RESULTS OF
PRECISION STUDY FOR THREE COMPONENT MIXTURE
Number of Range of
duplicate emission Precision,
Compound
measurements rates
2—Butanone
9 11,000 42,000 6.7
1,l,l—Trichloroethane
9 46,000 100,000 10.6
Toluene
9 5,100 14,000 13.1
Total
9 65,000 160,000 8.6
TABLE V. RESULTS OF ACCURACY STUDY FOR THREE COMPONENT MIXTURE
Compound
Number of
measurements
Average bi
% + 90%
as,
CL
2—Butanone
1,1,1—Trichioroethane
Toluene
9
9
9
—68.1
—21.0
—38.2
±
3.1
4.0
4.4
Total
9
—40.7
3•
CL = Confidence limit
1—65
-------
REFERENCES
C. E. Schmidt, V. D. Baif our, “Direct gas measurement techniques and
the utilization of emissions data for hazardous waste sites,” in
proceedings of the ASCE National Specialty Conference on
Environmental Engineering, 1983, p. 690.
B. R.. Dupont, “Measurement of volatile hazardous organics emissions
from land treatment facilIties,” J.A.P.C.A., 87, 168—176, (1987).
B. M. Ekiund, V. B. Balfour, C. E. Schmidt, Measurement of fugitive
volatile organic compound emission rates with an emission
isolation flux chamber,” in proceedings of the AIChE Summer
National Meeting, Philadelphia, PA, August 1984.
B. R.. Dupont, “A flux chamber/solid sorbent monitoring system for
use in hazardous organic emission measurements from land
treatment facilities,” in proceedings from the 79th Annual
Meeting of the Air Pollution Control Association, Minneapolis,
MM, June 1986.
M. R.. Kienbuach, and D. Ranum, “Developing and evaluating test
methods for quantifying air emissions from hazardous waste
disposal for measuring air emissions from the surface
Impoundments,” U.S. Environmental Protection Agency. Contract
No. 68—02—3889, Work Assignment 42, January 1986.
1-66
-------
FIELD EVALUATION OF THREE METHODS OF SOIL-GAS
MEASUREMENT FOR DELINEATION OF GROUND WATER CONTAMINATION
Henry B. Kerfoot, Lockheed Engineering and Management Services
Company, Inc., Las Vegas, Nevada
ABSTRACT
Three techniques for analysis of soil gases for detection and
delineation of ground—water contamination by volatile organic
compounds were evaluated above two distinct VOC ground—water plumes
(chloroform and benzene/chlorobenzene) in Henderson, Nevada. The
objectives of the studies were to evaluate the correlation of
results with the results of ground—water analyses and to assess the
variability among results from locations separated by short
distances. A driven probe was used to obtain grab samples for
on—site analysis by field gas chromatography, the PETREX SST—Py/MS
passive charcoal sampling/remote analysis system was evaluated, and
the Lockheed passive—sampling system (LPSS) using industrial—hygiene
charcoal passive samplers and remote analysis was tested. Results
from the grab—sample/on—site analysis technique and the LPSS method
showed good correlation with ground-water concentrations above the
chloroform plume. Results from the PETREX SST—Py/MS technique did
not correlate with chloroform concentrations. The short—range
(6—foot) precision of the grab—sampling/on—site analysis method was
characterized by relative standard deviations of 12 to 40 percent
while that of the LPSS method ranged from 10 to 20 percent.
Relative standard deviations of SST—Py/MS results over short
distances were very high. Above the benzene/chlorobenzene plume
none of the three techniques detected those compounds in the
overlying soil gases. The only successful technique for detection
of contamination there was measurement of CO 2 soil-gas
concentrations on the ground-water organic carbon concentrations in
monitoring wells there gave a correlation coefficient indicating a
greater than 95 percent significant correlation. Further studies on
soil—gas network design and data analysis are underway.
Although the research described in this article has been funded
wholly or in part by the United States Environmental Protection
Agency through contract 68—03—3249 to Lockheed Engineering and
Management Services Company, Inc., it has not been subjected to
Agency review and therefore does not necessarily reflect the views
of the Agency and no official endorsement should be inferred.
Mention of commercial products or trade names does not constitute
endorsement of their use.
1-67
-------
IWIRODUCTION
Protection of the nation’s ground water is a national priority.
Contamination from past waste disposal and leaking underground tanks
can destroy aquifers for decades. Some of the worst cases of such
contamination are from organic compounds because of their long
lifetime in the subsurface. For this reason, environmental
scientists have increasingly studied the fate, transport, and
detection of these contaminants.
In order to effectively respond to subsurface organic contamination,
it is necessary to establish the extent of the problem. Because of
the high cost of obtaining and analyzing ground—water or soil
samples, attention has recently turned to indirect methods for
preliminary site characterization. By utilizing a preliminary
reconnaissance—type technique, more efficient sampling networks can
be planned for further work. Soil—gas surveying is an emerging
technology applicable to detection and delineation of subsurface
contamination by volatile organic compounds (VOC’s). Table 1 lists
the twenty—five most frequently encountered substances at Superfund
sites; of these, fifteen are VOC’s amenable to detection by soil—gas
analysis. Because of the potential applications of this technology,
the U.S. EPA has funded an evaluation of soil—gas surveying
techniques.
In this paper, a field evaluation of three different soil-gas
surveying methods is discussed. The three methods were: a grap
sample/on—site analysis technique, the PEThEX SST—Py/MS method, and
a passive—sampling/remote—analysis system. The evaluation was above
a contaminated aquifer in Henderson, Nevada.
EXPERIMENTAL
The field evaluations were performed at the Pittman Lateral in
Henderson, Nevada (Figure 1). At the site unconsolidated gravel
alluvium with discontinuous caliche cement forms an unconfined
aquifer above a clay aquiclude and there is little or no soil
development. The ground surface, water table, and upper surface of
the aquiclude all dip gently towards the Las Vegas Wash to the north
and groundwater flow is to the north. Unconfined groundwater occurs
at 6 to 20 feet and contains two separate contaminant plumes
believed to originate at the industrial complex to the south.
Groundwater is monitored by analysis of samples from an east-west
line of wells separated by 200—foot intervals. Figure 2 is a
hydrogeologic cross section of the site. The groundwater at the
Pittman Lateral contains a wide variety of organic and Inorganic
substances. However, for the purposes of this study, the chloroform
plume on the eastern side (up to 900 ,g/L) and the
benzene/chlorobeuzene plume on the western side (up to 5,000 /2g/L)
1-68
-------
TABLE 1. HOST FREQUENTLY IDENTIFIED SUBSTANCES AT
546 SUPERFUND NATIONAL PRIORITY LIST SITES 8
Henry’s Law
constantb Percent
Substance (ppbv•L/pg of Sites
1 Trichloroethylene 72 33*
2 Lead and compounds NA 30
3 Toluene 56 28*
4 Benzene 71 26*
5 Polychiorinated biphenyls (PCBs) <<1 22
6 Chloroform 40 20*
7 Tetrachioroethylene 123 16*
8 Phenol <<1 15
9 Arsenic and compounds NA 15
10 Cadmium and compounds NA 15
11 Chromium and compounds NA 15
12 1,1,1—Trichioroethane 30 14*
13 Zinc and compounds NA 14
14 Ethylbenzene 59 13*
15 Xylene 43 13*
16 Methylene chloride 23 12*
17 trans—1,2—dichloroethylene 570 11*
18 Mercury NA 10
19 Copper and compounds NA 9
20 Cyanides (soluble salts) <<1 8
21 Vinyl chloride 5100 8*
22 1,2—Dichioroethane 88 8*
23 Chlorobenzene 32 8*
24 1,1 —Dichloroethane 420 8*
25 Carbon tetrachloride 3600 7*
8 Source: Kerfoot, 1987.
b Henry’s Law constant is the equilibrium gas—phase VOC concentration
(ppbv) divided by the concentration in water (pg/L).
* Compounds amenable to detection by soil-gas analysis.
1—69
-------
are of interest because these compounds are VOC’s amenable to
80i1-ga8 surveying.
The three techniques have been described in detail elsewhere
(Kerfoot, 1987; Voorhees et al., 1984; Kerfoot and Mayer, 1986).
The grab—sample method uses a steel pipe with an internal stainless
steel tube connected to sampling ports In the probe tip. The probe
is hammered in to a prescribed depth, a sampling manifold is
attached, soil gas is withdrawn through the manifold, and subsamples
are taken from the manifold with a gastight syringe. The syringes
are then carried to a nearby mobile laboratory for analysis by gas
chromatography. Real—time results for the 0C concentration at the
sampling location are obtained. The PETR.EX SST-Py/MS method uses
a wire with a charcoal—coated tip placed in a glass screweap tube.
The tube is opened and placed with the open end down in a shallow
(ca. 1 foot) hole and buried. The sampler is retrieved after a
measured exposure time, sealed, and shipped to a laboratory for
analysis by pyrolysis/mass spectrometry (Py/MS). Results are
reported In Ion counts and are said to represent the “Integrated
vertical flux of VOC’s at the sampling location. The Lockheed
passive—sampling system (LPSS) uses diffusional charcoal samplers.
The samplers are buried in a manifold at a shallow depth (ca. 1
foot) for a prescribed time, are retrieved and sealed, and are
transported to the laboratory for analysis by gas chromatography.
Results are In an average concentration of VOC’s at the sampling
locations. Carbon dioxide measurements were made using the probe
described above and a carbon dioxide detector tube (Draeger).
Samples were from a depth of 5 feet. Organic carbon analyses were
performed with a Dohrman carbon analyzer.
Sampling above the chloroform—contaminated ground water was
performed at the locations shown in Figure 3. Sampling above the
ground water contaminated with benzene and chlorobenzene was
performed at the locations shown in Figure 4. Grab samples were
taken from a depth of 4 feet except where otherwise noted. Passive
samplers and PE .EX SST—Py/MS samplers were buried at a 1—foot
depth; passive samplers were left in place for 14 days above the
chloroform plume and for 3 months above the benzene /chlorobenzene
plume. All PETREX SST—Py/14S samplers were left in place for 10
days.
Samples were taken at locations 20 feet from monitoring wells to
allow evaluation of the correlation of the soil—gas results with the
results of ground-water analyses. At some sampling locations,
closely spaced (ca. 3 feet) groups of two, three or four samples
were placed to allow evaluation of the short—range variability of
results from each technique. Because of the results obtained above
the benzene/chlorobenzene contamination, the passive samplers that
measure concentration were deployed near only one well and were left
in place for a very long time (3 months) and grab samples were taken
1-70
-------
within 1 foot of highly contaminated ground water, in an effort to
detect those compounds in soil gas there. PETREX samplers were
left in place for ten days.
RESULTS AND DISCUSSION
chloroform Plume
Table 2 lists the ground—water chloroform concentrations, the mean
soil—gas chloroform concentrations of grab samples from a 4—foot
depth as measured by on—site gas chromatographic analysis, the
results from the Lockheed , assive sampling system (LPSS) at a 1—foot
depth, and the PETR.EXW SST—Py/NS results. The 4—foot
grab—sample and LPSS data correlate well with the ground-water data,
and the two data sets correlate very strongly with each other. The
SST—Py/MS results (ion counts) do not correlate with the ground-
water data; the highest results (chloroform ion counts) were
obtained from samples 800 feet away from the closest location where
chloroform was detected in the underlying ground water. Figure 5
shows the mean results from each technique and ground—water analyses
as a function of the east-west coordinate.
Table 3 lists the mean values and associated standard deviations for
the closely spaced samples analyzed by each method. The precision
shown among these measurements helps indicate how much of a
variation in results can be attributed to inherent variability of
the method. This variability is a combination of sampling and
analysis imprecision along with real variations in the chloroform
concentrations. The precision of the grab—sample/on—site analysis
method, as indicated by the relative standard deviation (RSD; 100 x
standard deviation / mean), was 7 percent and 43 percent; the
passive—sampling method gave RSD values between 12 and 25 percent;
the SST—Py/MS results had RSD values of 98 percent (3 samples) and
140 percent (2 samples).
On the basis of the results of evaluation above the chloroform
plume, both the grab—sample/on—site analysis technique (Kerfoot and
Barrows, 1987) and the passive—sample/remote—analysis method
(Kerfoot and Mayer, 1986) gave results that correlated well with
ground—water concentrations. In addition,, acceptable precision was
obtained with both methods. The PETREX°’ SST—Py/MS method gave
results that did not correlate with ground-water chloroform
concentrations and the precision of results from closely spaced
samplers was unacceptable.
BENZENE/CHLOROBENZENE PLUME
Results of grab sampling with on—site analysis above the benzene and
chlorobenzene contaminant plume did not indicate the presence of
these compounds anywhere in soil gases above the contaminated ground
1—71
-------
TABLE 2. GROUND-WATER AND MEAN SOIL-GAS ANALYTICAL
RESULTS ABOVE THE CHLOROFORM PLUME
Ground- Mean Soil—Gas Concentration (ppb) 8 SST-Py/MS
Water Grab Sample Passive Sampling Results
Well Concentration (u /L) (4 ft)a (1 ft) (ion counts)a
635 __d 5482
631 5 C
629 11 23 2.0 _ d
627 175 68 7.8 __d
625 866 370 19.2 3121
150 14.3
623 555 40 2.4 2835
621 l0.5C _..d
a Mean of 4 locations around the well (see Figure 2) unless otherwise noted; source:
ICerfoot and Barrows, 1987.
B ND = not detected; ground—water detection limit = 5 pg/L, passive sampling detection
limit 0.02 ppb.
C Only one location.
d Not sampled
e Location 624 is halfway between 623 and 625.
variability is a combination of sampling and analysis imprecision along with real variations
In the chloroform concentrations. The precision of the grab—sample/on—site analysis method,
as indicated by the relative standard deviation (RSD; 100 x standard deviation ÷ mean), was
7 percent and 43 percent; the passive-sampling method gave RSD values between 12 and 25
percent; the SST-Py/MS results had RSD values of 98 percent (3 samples) and 140 percent (2
samples).
1—72
-------
TABLE 3. MEAN AND RELATIVE STANDARD DEVIATIONS OF CHLOROFORM
MEASUREMENTS FOR CLOSELY SPACED SAI4PLESa
Location
Passive—Sample
(ppbl;1—ft depth)
Grab—Sample
(ppb;4—ft depth)
SST—Py/MS
(ion counts; 1—ft depth)
635
4593 ( 140 )b
631
N.D.
--
--
629
2.6 ( 14 )C
627
6.7 (] 4 )C
124 ( 43 )d
625
15 ( 2 0 )C
624
14.3 (] 2 )e
189 (7)1
——
623
2.6 ( 25 )C
——
9092 (98)
621
N.D.
--
- -
a Relative standard deviation (7 ,) in parentheses
b Two samples separated by 3 feet
C Four samples in a square pattern 3 feet on a side
d Four samples in a trapezoidal pattern 3 ft x 3 ft x 6 ft x 4.2 ft
e Ten samples in a north—south line at 3—foot intervals
Three samples in a line at 3—foot intervals
1—73
-------
water. Although grab samples were taken 1/2 to 1 foot above the
water table adjacent to well 641, neither benzene nor chlorobenzene
was detected in them. LPSS samplers exposed for three months did
not show any benzene or chiorobeuzene, either. Table 4 lists the
concentrations measured in ground—water samples from wells in that
plume as well i soil—gas results from nearby sampling locations.
Because PETREX’ ’ SST—Py/MS results were the only ones to detect
berizene or clilorobenzene in soil gases above that contaminant plume,
an evaluation of the precision of results among closely spaced
samplers was performed. At the two locations with multiple samplers
the RSD values were 81 percent (two samplers) and 170 percent (three
samplers) for benzene. Chlorobenzene ion counts had an RSD value of
58 percent (2 samples at 641), with all non—detect results at the
other multi—sampler location (649).
In consideration of the contrasting results obtained above the
chloroform plume and the benzene /chlorobenzene plume, under the same
hydrogeologic circumstances it was postulated that compound—specific
properties must be responsible for the differences observed.
chloroform has Henry’s Law and diffusion constants intermediate to
those of benzene and chlorobenzene and the maximum concentrations of
the aromatic compounds are nearly an order of magnitude higher than
chloroform, so that differences in partitioning Into soil gases or
diffusing through the vadose zone are not the causes of the
situation encountered. It has been noted that hydrocarbons
biodegrade readily in shallow soil gases (Evans arid Thompson, 1986),
and chlorobenzene has also been shown to biodegrade aerobically
while chloroform is resistant to that process (Bouwer, 1984).
Therefore, the disappearance of the two aromatic compounds from the
soil gases could be due to this process.
As a test of this hypothesis, carbon dioxide concentrations in soil
gases were measured at a 5—foot depth several locations at the site
and dissolved organic carbon concentrations of ground—water samples
were measured. Table 5 shows the two data sets. Linear regression
shows that there is a 90 percent significant correlation between the
two variables. This is in agreement with, but does not prove, the
hypothesis that aerobic biodegradation is occurring at a rate
limited by the concentration of organic carbon present.
CONCLUS IONS
Soil—gas surveying can be a valuable field reconnaissance technique
to delineate VOC contamination. Both active sampling with on-site
analysis and passive sampling with remote analysis methods gave
results that adequately delineated ground—water VOC contamination In
this study, although not all the passive sampling methods used gave
the same quality of data. However, the methods that were used
successfully were not able to detect nearby subsurface contamination
at the same site. Although evidence indicates that aerobic
1-74
-------
TABLE 5. GROUNDWATER DISSOLVED CARBON AND SOIL—GAS CARBON
DIOXIDE CONCENTRATIONS
Well
631
64].
643
645
647
649
Organic
Carbon (mg/L )
2.52
13.01
12.27
14.79
12.54
9.20
Inorganic
Carbon (ing/L )
40.80
75.99
73.20
76.80
73.65
57.86
Soil—Gas
Carbon Dioxide (% )
0.089
0.452
0.318
0.352
0.410
0.096
TABLE 4. GROUNDWATER AND SOIL—GAS MEASUREMENTS
FOR BENZENE/CHLOROBENZENE PLUME
B Mean of samplers around a well.
b Not sampled.
-0
Soil—Gas “ *1r ( ppbv )
Grab Sample Passive Sample
C H Cl C 6 H. C H 5 cl
ND ND ND ND
_.b ._b ..._b
ND ND ND ND
ND ND _b __b
ND ND . .._b _ .b
ND ND ..b _. .b
ND ND b _ . .b
ND ND ND ND
Ground-
Water
Concentration (pg/L)
c 6 w. cH 5 C 1
<10 <10
340 520
4700 3200
3200 4520
3100 4880
1300 2400
<10 <10
Well
623
635
639
641
645
649
653
Blank
C 1 jç
Mean 8 SST-Py/MS
Results
(ion counts)
C i i i ’-
218
793
529
380
218
189
353
ND
103
575
139
121
133
148
133
ND
1—75
-------
biodegradation may be responsible for that, further work would be
required to prove the hypothesis. However, these results indicate
that workers using the technology should be careful to demonstrate
its correlations with subsurface contamination and its short—range
precision at each site as routine quality assurance precautions.
ACKNOWLEDGEMENTS
The assistance of C.L. Mayer and J.C. Curtis of Lockheed Engineering
and Management Services Company and of P.B. Durgin of the U.S. EPA
Environmental Monitoring Systems Laboratory, Las Vegas, Nevada, were
crucial in this work.
REFERENCES
Bouwer, E.J., 1984. “Biotransformation of Organic Micropollutants
in the Subsurface,” In Proceedings of the NWWA/API Conference on
Petroleum Hydrocarbons and Organic Chemicals in Ground Water —
Prevention, Detection, and Restoration , 5—i November 1984,
National Water Well Association, Worthington, Ohio, pp. 66—80.
Evans, O.D., and G.M. Thompson, 1986. “Field and Interpretation
Techniques for Delineating Subsurface Petroleum Hydrocarbon
Spills Using Soil Gas Analysis,” In Proceedings of the NWWA/API
Conference on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water — Prevention, Detection, and Restoration , 12—14
November 1986, pp. 444—458.
K.erfoot, H.B., 1987. “Shallow—Probe Soil-Gas Sampling for
Indication of Ground Water Contamination by Chloroform,”
Environmental Science and Technology , In Press.
Kerfoot, RB., and L.J. Barrows, 1987. Soil—Gas Measurement for
Detection of Subsurface Organic Contamination , National Technical
Information Service, U.S. Department of Commerce, Springfield,
Virginia, Accession Number PB87 174884/AS.
Kerfoot, H.B., and C.L. Mayer, 1986. “The Use of Industrial Hygiene
Samplers for Soil—Gas Surveying,” Ground Water Monitoring Review ,
Fall, 1986, pp. 74—78.
Voorhees, K.J., J.C. Hickey, and R.W. Kiusnian, 1984. “Analysis of
Groundwater Contamination by a New Surface Static Trapping/Mass
Spectrometry Technique, Analytical Chemistry 56: 2604—2607.
1-76
-------
Figure 1
Study Site Location.
LAS VEGAS WASH
PITTMAN LATERAL TRANSECT
•1 —
INDUSTRIAL
COMPLEX
.—.--.-I
o 1 2 3
N
KILOMETERS
PITTMAN
1—77
-------
BENZENE/CHLOROBENZENE
0
I-
EAST
SCALE IN FEET
0 500
SCALE IN METERS
200
0
I
LU
1660
0
1640
1620
I-
I.’
C ,,
z
0
I-
>
14J
—S
w
CHLOROFORM
1
FACE
1560
655
650
645
640
WEST 148 FT. DEEP
635
630
-n
I
(D
I\)
STATIONS
625
•1
0
09
‘9
0
0
1
0
09
‘ 9
C,
‘4.
0
09
09
S .”
C 4.
(V
620
615
610
TEST WELL
.1.
-------
Figure 3
Sampling Locations Above the Chloroform Plume.
624 e Q
fl
o 5 e U D
635 633 631 629 627 625 623 621
NOTES:
a 2 Passive samplers separated by 3 feet o
bsingle_monitor and Dual—monitor passive aamplers 0
separated by 3 feet
C 3—foot square pattern of passive samplers o
dlrapezoidal (3ff a 3ff a 6ft x 4.2ff) pattern of grab 0
samples and 3—ft square pattern of passive samples
3 SST samplers at 3—foot intervals along a line
LEGEND
Well
• Grab—sampling location
0 Passive—sampling location
X SST—sampling location
200 feet —.4
Figure 4
Sampling Locations Above the Benzene/Chlorobenzene Plume.
653 649 645 641 639
a a
• • e • oeo
sos
400 feet 400 feet 400 feet 200 feet— 1
LEGEND
NOTES: Well
5 Three Samplers at 3—foot intervals 0 3—ft square pattern of passive samplers
along a line (3—month exposure)
bTwo Samplers separated by 3 feet • Two grab—sampling locations 3 feet apart
.. 3—ft square pattern of grab—sampling locations
X SST samplers
0 Single grab—sampling location
1-79
-------
Figure 5
Soil—Gas and Ground—Water Measurement Results.
o GROUNDWATER CHLOROFORM CONCENTRATION (pg/L)
L SOIL-GAS PASSIVE-SAMPLING RESULTS (ng/pL X 10)
o SOIL-GAS GRAB-SAMPLING RESULTS (ppbv)
800
700
600
500
400
300
200
/
/
/
/
/ ‘ ‘
I
/
/
100
50
0
629
WEST
627
625
623
621
EAST
1-80
-------
PROCEDURES USED TO MEASURE THE AMOUNT OF
2,3,7, 8-TETRACHLORODIBENZO-P-DIOXIN
(2,3,7,8-TCDD) IN THE AMBIENT AIR
NEAR A SUPERFUND SITE CLEAN-UP OPERATION
Billy J. Fairless, Dale I. Bates, Jody Hudson, Robert D. Kleopfer,
Thomas T. Holloway, Debra L. Morey, U.S. Environmental Protection
Agency, Region VII, Kansas City, Kansas; Tony Babb, IT Corporation, Air
Quality Services, Knoxville, Tennessee
ABSTRACT
Sampling and analytical procedures are described that were successfully
used to monitor for 2,3,7,8-tetrachlorodibenzo—p—dioxin (2,3,7,8-TCDD)
in air samples collected near a superfund site clean—up operation.
Measured concentration of 2,3,7,8—TCDD in air samples are related to
both an action level (3.0 picograms per standard cubic eter of air)
and to a calculated no observed effect level (5.5. pg/N ). The study
concluded that it is possible to collect reliable data for 2,3,7,8—TCDD
in air at concentrations that are below the action level specified by
the Centers for Disease Control. Data quality was defined relative to
the quality control procedures described in the study. There was no
apparent relationship between particulate matter in the air and
2,3,7,8—TCDD in the air.
INTRODUCTION
For some animal species, 2,3,7,8—tetrachlorodibenzo-p—dio in (2,3,7,8-
TCDD) is one of the most toxic synthetic compounds known. The LD-50
has been reported to be 0.6 ug per Kg of body weight for the guinea
pig, 115 ug per Kg for the rabbit and 22 ug per Kg for the rat.
However, 2,3,7,8-.TCDD does not ap ear to be nearly as toxic to humans
as it is to other animal species. Some of the chlorinated dioxin
isomers are, for all practical purposes, not toxic at all. The LD—50
for 2,7—dichlorodib nzo-p-dioxin is approximately 2.0 g/Kg of body
weight for the rat. These facts require that any analytical procedure
used to measure for the concentration of 2,3,7,8—TCDD must be isomer
specific.
In addition to being toxic, 2,3,7,8-TCDD is also a suspected
carcinogen. Recently, Dr. R. Kim ough and others suggested the
extremely low number of 2.8 X 10 g/Kg body weight per day a being
the dose responsible for a cancer risk of one in one million.
Previous studies have d ected concentrations of 1 labeled dioxin in the
air in the range of 10 grams per cubic meter. Several studies have
shown the presence of dioxins in emissions from incinerators at
similarly low concentrations. Attempts to measure dioxin
concentrations in ambient air at these levels require very sensitive
analytical procedures.
1-8 ].
-------
In the early 19 7 0’s, 2,3,7,8—TCDD was formed as a byproduct in the
production of hexachiorophene and/or Agent Orange in a small facility
in southwest Missouri. It is known that, unlike some other sources of
dioxin, this process resulted in a very high fraction of the 2,3,7,8-
TCDD isomer being formed. It is now believed that the 2,3,7,8-TCDD was
produced by condensation of 2,4,5-trichiorophenol. The phenol resulted
from alkaline treatment of l,2,4,5—tetrachlorobenzene as shown below in
equation 1.
NaOH
1) 2C 6 H 2 C1 4 ————) 2C 6 H 3 0C1 3 ————> C 12 H 4 O 2 C1 4 + 2 HC1
Waste materials containing the dioxin byproduct were mixed with used
oil and subsequently applied to roads and other surfaces for dust
control.
One of the places where the waste material from southwest Missouri was
applied was a trailer park near St. Louis, Missouri. The site covers
an area of approximately 11 acres. It has an irregular shape as shown
by the site map (Figure 1). The surface and subsurface soils at the
site were tested and were found to contain 2,3,7,8-TCDD at
concentrations above 1.0 ug per Kg of soil (1.0 ppb) at most locations
within the site. Therefore, a mit gation plan was prepared to control
exposure to the contaminated soil. The plan called for the removal of
the contaminated soil from the surface and for storage of the removed
material in a safe location on site until detoxification procedures are
available -
The EPA has a responsibility under the comprehensive Environmental
Response, Compensation and Liability Act (CERCLA or Superfund) to clean
up hazardous waste sites known to be contaminated with toxic or
hazardous materials. A second responsibility is that agency clean—up
activities must not constitute a source of pollution for members of the
general public who are in the immediate vicinity of the clean—up
operation. At the beginning of this work, it was not apparent that it
would be possible with available technology to meet both of these
responsibilities for the site in question. This paper describes
considerations which went into the design and operation of the network
and our findings during the first five months of operation. Some of
the specific questions we had to address and hoped to answer are listed
below:
Can the concentration of 2,3,7,8—TCDD in ambient air be measured at
levels that cause an insignificant risk to the public?
Can a cost—effective network which measures the average amount of
dioxin in the air be implemented over time, with a high degree of
confidence? If not, are other options available that will provide
acceptable answers?
1-82
-------
If dioxin is found in the air, will it be absorbed to particulate
matter that is trapped on a filter or will it be in the vapor phase or
both?
Relative to local weather conditions and the normal scatter in any
measurement, how many sampling sites should the network contain? How
many samples should we collect at each sampling site each day? How
long should the network be operated.
What are the best available procedures to collect a sample? What are
the best analytical procedures? What reliability can be expected from
these procedures?
Is there any relationship between the amount of particulate matter in
the air and 2,3,7,8—TCDD in the air?
Is there a measurable cause—effect relationship between clean-up
operations and measured concentrations of 2,3,7,8—TCDD in the air? Is
any of the 2,3,7,8—TCDD in the air coming from background sources?
If significant concentrations of 2,3,7,8—TCDD are found in the air, can
models be used to predict the path of any plume that might be caused by
the clean—up operations?
Are any of the pollution- abatement actions more effective in
controlling the release of dioxin than others?
PROCEDURES
1. Design of the Monitoring Network
When designing the monitoring network, we assumed that we needed
measurement detection limit in the range of 0.1 to 1.0 pi ograxn/M in
order to obtain reliable measurements at the 5 picogram/M level. The
response from the Centers for Di ease Control (CDC) contained their
recommendation of 5.5 picogram/M as an estimated no observed effect
level (NOE ) and their concurrence with our 6 recommendation of 3.0
picogram/M as a “warning” or action level. Realizing that the total
cost of sample collection and analyses would probably exceed $1,000 per
sample, our objective was to collect the minimum number of samples that
would have a good (95% confidence level) probability of showing an
exceedance of the action level if one occurred and of showing that no
exceedance occurred if indeed one did not occur. Based on prior
experience, we estimated that we could monitor for the concentration of
2,3,7,8—TCDD in ambient air at these levels with an accuracy of 25%
relative standard deviation based on the recovery of field spikes. We
assumed that the upwind and downwind site data would each have a normal
distribution. The number of measured concentrations (samples) needed
at each sampling location was then calculated as a function of the
1-83
-------
difference between the action level (true mean) and the mean of a given
number of measured values.
The statistical calculation is outlined as follows. The average
concentration of all samples collected at a site is used as an
approximation of the true average concentration (which would be
obtained from continuous monitoring of the air over a long period of
time). The uncertainty of that approximation is indicated by the 95%
confidence limits as calculated by the following formula.
t S
95% confidence limits Y + .975, n—i n
where Y = the sample average
n = the number of samples at each sampling location
s = the standard deviation of samples collected at that sampling
location
t• 975 , _] = the tabulated student’s t value for n—i degrees of freedom
The width of the 95% confidence interval for a site depends on the number of
samples collected and the scatter among the data from those samples, as is
shown in Table 1.
A comparison of the experimental data with the action level will result in a
conclusion that no exceedance occurred if the 95% confidence interval is
completely below the action level. The data will demonstrate that an
exceedance did occur if the 95% confidence interval is completely above the
action level. The data will demonstrate that more samples would be needed in
order to show conclusively whether or not the true average concentration
exceeded the action level if the action level is within the 95% confidence
level.
We elected to initially collect 14 samples at each monitoring site. Using
14 data values and the assumptions described above, a definitive conclusion
would then be reached if the data were either 14% above or below the action
level.
In order to obtain sufficient data to be able to assess the effects of on-
site activities on the off—site ambient air under variable wind conditions,
the monitoring network had to be designed to provide long-term monitoring of
the air at or near the property boundaries of the site. Based on the
physical configuration of the site (see Figure 1), which is long relative to
its width, we decided that a minimum of six fixed monitoring locations would
be needed to insure consistency throughout the study and to have one upwind
1-84
-------
on the appropriate number of monitoring sites in the network, was to minimize
the probability that contaminated air would pass between the monitors without
being detected. The three largest variables effecting this probability were
the size of the area within the site that might be releasing dioxin, the
local variation in the wind direction during a sample collection time period
and the duration of the study.
Considering that dioxin might be released from approximately one-fifth of the
site at any given time from truck traffic and construction work, that the
wind direction was very seldom in one direction for 80% of a 24-hour period
and that the study would require approximately 200 working days, we concluded
that six monitoring sites would be sufficient. An on-site meteorological
monitoring station was incorporated into the network design for the purpose
of obtaining adequate wind speed and direction data.
A site visit was made prior to finalizing the network design to obtain
detailed information on the topography, and to choose the specific sampling
locations relative to anticipated activities at the site. The specific
locations for the air samplers were selected so they would be near but just
inside the perimeter fence for security purposes, would be consistent with
accepted siting guidance for criteria pollutant monitoring, would provide
permanent placement throughout the life of the project (i.e., the samplers
would not have to be relocated during the course of the excavation activitie
at the site), and would provide adequate coverage for most wind directions.
The specific sampling locations which were selected are shown in Figure 1.
2. Collection of Representative Samples
To collect representative samples, we used commercially available modified
high—volume air samplers that employ both a filter for collecting particulate
matter and a solid adsorbent for collecting vapors. A diagram of the air
sampler (General Metal Works, Inc., Model PS—i) is shown in Figure 2. In
operation, a known volume of air (calculated from the flow rate and time of
sampling) is drawn through a dual-chambered sampling module (see Figure 3)
and exhausted to the air via a 10—foot exhaust duct. The upper portion of
the sampling module holds a 4—inch diameter glass fiber filter which collects
the particulate matter and the lower portion consists of a cylindrical glass
cartridge (65 mm x 125 mm) containing a 3-inch-long solid adsorbent which
entraps selected vapor phase compounds. Polyether—typ, polyurethane foam
(PUF) plugs were utilized as the solid adsorbent material.
To measure the very low concentrations of dioxin required in this project, a
large sample volume of air was required. Since the air samplers are only
capable of providing a flow rate of approximately 0.280 cubic meters/minute,
it was determined that the samples should be collected on a 24—hour basis (+
15%) . This time and flow will give a sampled air volume of 300—400 cubic
meters of air.
1-85
-------
The samplers were placed on 1—meter high platforms to obtain samples of the
ambient air in the breathing zone. The samplers were operated daily except
on Sunday. If excavation occurred on Sunday, the samplers were operated on
that day as well.
We assumed the wind direction could not be predicted for the 24-hour sampling
period, nor would it be constant over that period of time. Therefore, all of
the samplers were operated to collect samples each day.
During the first 14 days of sampling, all of the samples were analyzed for
both particulate matter and 2,3,7,8—TCDD to obtain baseline data for the
site. Pollution abatement activities were occurring at the site during this
time period. Subsequent to this initial sampling period, only one upwind and
one downwind sample was submitted for 2,3,7,8—TCDD analysis each day. The
selection was based on the prevailing wind direction and the amount of
particulate matter collected on each filter for the sampling period.
3. Comparability
For a risk assessment, we estimated the maximum amount of dioxin a person
just off—site would experience during the time of the clean-up activities.
We decided to use a daily averaging process because an average is more
comparable to 8 an action level that is based on chronic effects than is a
single value. We decided to average the data values from each monitor
separately (rather than average data from different monitors) because that
would be more representative of exposure for someone living near that
monitor. Since we wanted to be able to take pollution abatement actions as
quickly as possible after the data were available, we decided to use a 14—day
running average (average concentration of the most recent 14 days) which
would be calculated daily. In the interest of safety, we elected to use the
detection limit as a measured value when calculating the running averages for
all samples that did not contain a measurable concentration of 2,3,7,8—TCDD.
Finally, since none of the analyses of non—downwind samples showed any
measurable 2,3,7,8—TCDD, we decided to use an average of these numbers for
those days when data were not available. A daily data point might be missed
for a specific monitor if that monitor was neither upwind nor downwind of the
site, or if no work was occurring at the site due to bad weather or for a
non—working Sunday.
4. General Procedures
Particulate Matter (PM) and 2,3,7,8—TCDD sampling utilized samplers as
previously described. Polyurethane foam (PtJF) cartridges were precleaned by
the laboratory and shipped to the on-site sample coordinator in sealed, glass
sample jars. Filters were obtained from the PUF sampler manufacturer in lots
of 100. Five percent of the filters were run as blanks to ensure acceptable
detection limits. Sample modules were collected and new samples started each
morning prior to the start of any remedial activity on site. Surgical gloves
1-86
-------
and Teflon tipped forceps were used to remove the glass fiber filter and the
PUF cartridge from the sample module. The glass fiber filter was placed on a
calibrated balance and weighed to the nearest 0.1 milligram. The glass fiber
filter was then folded in half twice (sample side inward) and placed in the
glass cartridge on top of the PUF plug. The sample cartridge was then
wrapped in aluminum foil (to shield it from sunlight) and placed in the
sample jar from which it came.
The concentration of the particulate matter (PM) for each sample was
calculated immediately and used in the selection of samples for 2,3,7,8—TCDD
analyses.
5. Analytical Procedures
All samples were analyzed by an EPA contractor in accordance with a Region
VII standard method titled “Determination f 2,3,7,8—TCDD in Air Samples
Using Gas Chromatography-Mass Spectrometry.”
T 9 samples were spiked with internal (13C 12 —2,3,7,8-TCDD) and surrogate
C Cl 4 2,3,7,8—TCDD) standards of isotopically labeled 2,3,7,8—TCDD. The
samples (filter, PUF and glass cylinder) were then extracted with methylene
chloride in a Soxhlet apparatus. The extracts were cleaned using silica gel,
modified silica gel, alumina and carbon prior to the analyses by high
resolution gas chromatography and low resolution mass spectrometry.
The gas chromatography column was a 30m x 0.32 mm I.D. fused silica capillary
DB—5 with a 0.25 u film thickness. Calibration was done by tabulating the
peak heights or peak areas from triplicate injections of 2,3,7,8—TCDD versus
the internal standard. Quantification is based on the response of native
TCDD relative to the isotopically labeled TCDD internal standard.
Performance is assessed based on extensive quality assurance requirements.
These include a requirement for accuracy of a surrogate analyses on each
sample.
6. Quality Assurance
To minimize sample handling and/or contamination in the field, two sampling
modules equipped with quick release connectors were acquired for each air
sampler. The availability of two complete sampling modules, which were
numbered for ready identification, made it possible to simply exchange a
clean sampling module for the module containing the sample in the field. The
modules were transported to and from the sampling sites individually wrapped
in plastic bags and stored in a closed container (an ice chest customized to
hold the modules upright). This practice enabled all sample handling
(disassembly of module and placement of the filter, PUF and glass cartridge
in a sample container) and sampling module preparation (cleaning and
assembly) to occur in a controlled environment.
1-87
-------
A five—point calibration curve (equation) for each sampler was obtained
initially using a magnehelic gauge and a calibrated orifice. The equation
for each sampler was then used to calculate sample volumes. The calibration
was repeated monthly and a one—point check was performed every other week. A
flow audit on 25% of the air samplers in use was performed monthly by an
individual other than the normal operator and with a different calibrated
orifice. The maximum acceptable difference between the reported (sampler)
and actual (calibrated) flows was established as ± 7%.
Laboratory CC/MS instrument calibration consisted of an initial 3—point
calibration conducted in triplicate. The mean, standard deviation, and %
Relative Standard Deviation (RSD) of the Relative Response Factor (RRF) for
2,3,7,8—TCDD was calculated at each of the three concentration levels, as
well as for the overall. Acceptable calibration required an RSD of 10% or
less at each individual level, as well as for the overall. Every 8 hours,
the calibration was to be verified through the analysis of a low level
standard solution. The percent difference of the RRF for the continuing
calibration from that of the overall mean RRF of the initial calibration
could not be more than 10%.
Both system and performance audits were included in the air monitoring plan
to ensure that the established procedures were actually being followed. The
audit process provided the means for continually evaluating the quality of
the data being generated, identifying apparent problems quickly, and making
in-process changes to correct apparent problems.
The use of quality control samples was included as a routine means of
tracking the precision and accuracy of the data generated and detecting
problems relating to the quality of the data reported. Throughout the
sampling effort, one field blank, one blank to be spiked by the laboratory,
and one field—spiked sample were submitted for analysis with every 17 actual
samples. In addition, during the initial 14-day sampling period, a second
air sampler was collocated with one of the perimeter samplers to collect
duplicate field samples.
RESULTS, DISCUSSION AND CONCLUSIONS
Data Quality
We have grouped the data quality information into qualitative information and
quantitative data.’ Based on the results of our field and laboratory
audits, we concluded that the procedures described above were being followed
as written, and that, with the exception of two data points, all data were
acceptable relative to the qualitative variables.
Two concentration levels were utilized in the performance evaluation (PE)
samples. A high level PE of 11.6 ng was used to monitor bias above the
exposure limit. The three data points generated at this level gave a mean
value of 10.99 ng or 95% recovery. Control limits at
1-88
-------
the 95% confidence interval were 9.65—13.6 ng. Twenty—six medium level
performance audit samples containing 5.8 ng were analyzed. Mean recovery was
5.29 ng or 91% recovery. A 95% confidence interval of 4.35 — 6.09 ng was
determined.
The field blanks consisted of a filter, PUF and glass cylinder to which only
surrogate and internal standard solutions were added. The 19 field blank
analyses resulted in no detectable quantities of 2,3,7,8—TCDD.
The laboratory fortified 19 samples with 5 ng of 2,3,7,8—TCDD. The mean
recovery of these was 4.75 ng or 95% recovery.
Nine audits of the sampler flows were performed during the study and gave a
standard deviation of flow difference of 1.72%.
Considering the variances in the measurements of both the volume of air
sampled and the amount of dioxin found in the sample, we estimate the data
are accurate to within + 12% of the reported values at the 95% confidence
level.
Is 2,3,7,8—TCDD in the Vapor Phase or on Particulate Material ?
Two experiments were conducted to try to determine what fraction of the
2,3,7,8—TCDD and 2,3,7,8-TCDF would pass through the filter and what fraction
would remain on the filter under the sampling procedures described above. We
assume that the materials were in the vapor phase when they passed through
the filter.
The first experiment was designed to evaluate the potential for analyte
breakthrough of 2,3,7,8—TCDD, l,2,3,4-TCDD and 2,3,7,8—TCDF. The experiment
consisted of spiking clean filters with 10 ng each of 2,3,7,8-TCDD and
2,3,7,8—TCDF and 13.8 ng of 1,2,3,4—TCDD in 200 ul hexane. Ambient air was
then drawn through the samplers for varying lengths of time using identical
flow rates. At least three samples were collected for each time period. The
PUF and filter were analyzed together as described above. The results of the
experiment are shown in Table 2.
From these data we conclude that the sampling procedures are effectively
collecting all of the 2,3,7,8—TCDD in the sampled air and that our field
spikes can be used for an accurate estimate of method bias. These
conclusions are supported by work recently r ported by F.L. DeRoos, J.E.
Tabor, S.E. Miller, S.C. Watson and J.A. Hatchel.
In the second experiment, a solution from a standard containing both 2,3,7,8—
TCDD and 2,3,7,8-TCDF (tetrachloro—p-dibenzofuran) was placed on the glass
fiber filter. Uncontaminated air was then passed through the sampler as
described above for 17.5 minutes, 2 hours and 24 minutes, and 24 hours. The
filters and PUF’s were then analyzed separately. The average total (filter
and PUF) percent recovery for
1-89
-------
dioxin was 112% with a standard deviation of 1.3%. The average percent
recovery of Furan was 90% with a standard deviation of 11%. The resulting
data are shown in Table III. It was obvious from this experiment that the
dioxin was very slowly migrating from the filter to the PUF. We conclude
that an analyses of only the particulate matter or only the vapor phase in
the sample would both give erroneous results. It also appears that the
Furans are more easily transferred from the filter to the PUF than are the
diox ins.
Effectiveness of Sampling Methodology
The data described above provides a basis for concluding that the sampling
equipment and procedures were adequate for this study. Clearly, the sampler
must capture both particulate and vapor phase materials if the data are to be
representative of the air sampled. The volume of air sampled was just large
enough to give adequate data precision (the maximum RSD at the monitor
showing highes concentra-tions of 2,3,7,8-TCDD was approximately 22%) in the
3—5 picogram/M concentration range. The samplers were reliable (no down
time) and they maintained the initial flow characteristics very well (maximum
change in measured flow was approximately 2%). The ability to very easily
change the sampling modules made field work much more convenient and probably
improved data precision considerably.
Are there any Relationships between Weather Conditions and 2,3,7,8—TCDD in
the Air ?
At no time during this study did we observe a measurable concentration of
2,3,7,8-TCDD at a non-downwind monitor. Approximately 200 samples were
collected and analyzed for 2,3,7,8—TCDD. Ten of those samples contained
concentrations of 2,3,7,8—TCDD above the detection limit. Each of the
positive concentrations occurred on days when clean—up activities were
occurring in the immediate vicinity of the monitor and when the clean—up
activities were upwind of the monitor for most of the day. From these data
we conclude that there is a measurable relationship between wind direction
and concentrations of 2,3,7,8—TCDD in the air. We also conclude that the
2,3,7,8—TCDD we found in the air was originating at the location of the
clean—up work and that the contamination remains in the air only for a short
time period. Any background of 2,3,7,8-TCDD is below the method dection
limit -
We were unable to find a model that would, in our hands, accurately give the
experimental results we obtained.
Can Particulate Matter Concentrations be Used to Predict 2,3,7,8—TCDD
Concentrations ?
Figure 4 is a graph showing the relationship between particulate matter
concentrations and 2,3,7,8—TCDD concentrations. Although there may be a very
general relationship, it is apparent that one could not reliably predict
2,3,7,8—TCDD from measured particulate matter
1-90
-------
concentrations. It is also obvious from the data in the graph that there is
no particulate matter concentration that could be used as a control limit for
2,3,7,8—TCDD. The particulate matter concentrations referred to were
obtained by the procedures described in this paper. They are not equivalent
to the TSP values usually obtained with a Hi-Vol sampler. However, a good
relationship between TSP by the EPA-approved procedure and the measured PZ
concentrations has been demonstrated for this study.
Since dioxin analyses are so expensive relative to TSP analyses, it may be
beneficial in future studies to look for a 2,3,7,8—TCDD/TSP relationship
under different conditions.
Was the General Public Exposed to Significant Concentrations of 2,3,7,8-TCDD
as a Result of this Action ?
Figure 5 is a graph showing the 14-day running average and the associated 95%
confidence levels of those averages for the monitoring site with the highest
concentrations. As can be seen, all of the 14—day averages were well below
the warning level and the NOEL at all times. We would emphasize that the
averages were calculated using the assumption that the detection limit was a
measured concentration for all samples that did not have a measurable
concentration of 2,3,7,8—TCDD. Therefore, the actual exposure was probably
even less than the maximum possible indicated by the graph. From this data,
we conclude that concentrations of 2,3,7,8—TCDD that cause an insignificant
risk to the public can be measured in ambient air using the procedures
described above. We also conclude that during this stu y, the public was not
exposed to a significant concentration (5.5 picogram/M for a “few months”)
of 2,3,7,8—TCDD at any time.
Can Monitoring Costs be Reduced in Future Studies ?
We estimate the total cost of the study described above was approximately
$295,000 or approximately $1,500 per sample. After reviewing the data from
this study, we see no way to substantially reduce the costs without
additional data. It does not appear to be possible to use particulate matter
as an indicator parameter for dioxin, thus, reducing the analytical costs.
We would not recommend reducing the size of the monitoring network for a site
of this size. We believe the network should be in operation any time
pollution abatement activities are occurring at the site. We doubt that
efforts to improve data quality would be cost—effective. The authors are
investigating sample breakthrough for longer sampling times both for the PUF
and other solid absorbents. An option that might result in a reduced cost
would be to reduce the time spent to obtain baseline data from 14 days to a
shorter time period. This kind of decision is properly a role of management
since costs must be balanced against the increased risk of obtaining a false
positive and of unknowingly exposing the public to a significant amount of
2,3,7,8-TCDD. The results of this work should make that decision easier to
reach.
1-91
-------
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Morris Kay and John
Wicklund for providing resources and encouragement, and to the many EPA and
EPA contractor employees who have participated in this effort by doing their
assigned work in an efficient and professional manner.
1-92
-------
COMPLETE REFERENCES
1. Esposito, M.P.; Drake, H.M.; Smith J.A.; Owens, T.W. “Dioxins: Sources,
Exposure, Transport and Control,” Volume I, Report No. EPA-600/2—80-156,
Industrial Environmental Research Laboratory, Office of Research and
Development, U.S. EPA, Cincinnati, Ohio (Pages 147-199).
2. Tschirley, Fred H. “Dioxin,” Scientific American, February 1986, Volume
254, Number 2, February 1986, Pages 29—35.
3. Kimbrough, Dr. R., as referenced in April 1, 1985, letter from Dr. Vernon
Houk, Center For Environmental Health, Centers for Disease Control to Mr.
Morris Kay, Regional Administrator, EPA, Region vii.
4. Kleopfer, R.D. “Chemosphere,” 14, 739, (1985).
5. Emergency Planning and Response Branch, U.S. EPA, Region VII “Quail Run
Mitigation Plan,” 25 Funston Road, Kansas City, Kansas.
6. Code of Federal Regulations 40, Part 50.11, Appendix B.
7. DeRoos, F.L.; Tabor, J.E.; Miller, S.E.; Watson, S.C.; Hatchel, J.A.
“Evaluation of an EPA High-Volume Air Sampler for Polychlorinated
Dibenzo—P-Dioxins and Polychiorinated Dibenzofurans”; Contract Number 68-
02—4127, U.s. Environmental Protection Agency, Methods Development and
Analyses Division, Environmental Monitoring Systems Laboratory, Research
Triangle Park, North Carolina 27711.
8. Margolis, Dr. Stephan “Health Consultation, Dioxin Ambient Air Levels
Off—Site,” Memorandum to Mr. Ed Skowronski, Public Health Advisor, EPA,
Region VII, 1985.
9. U.S. EPA, Region VII, “Determination of 2,3,7,8—TCDD in Air Samples Using
Gas Chromatography — Mass Spectrometry,” 25 Funston Road, Kansas City,
Kansas 66115, August 1985.
10. Bates, Dale; Holloway, Thomas T.; Environmental Monitoring and Compliance
Branch, Operating and Quality Assurances Procedures Manual, U.S. EPA,
Region VII, 25 Funston Road, Kansas City, Kansas 66115, November 11,
1982.
1—93
-------
TABLE I
Estimation of Number of Samples Needed
to Meet Study Objective
Number of Samples
per Site
1
t 975 , n—i
RSD = 12% RSD = 24% RSD = 36%
95% confidence limits = sample average +
7
14
28
56
2.447
2,160
2.052
1.96
11%
7%
5%
3%
22%
14%
9%
6%
33%
21%
14%
9%
1-94
-------
TABLE II
Breakthrough Study Results
0/0
12/ 180
24/ 360
48/720
72/1080
2,3,7,8-TCDD
1,2,3,4-TCDD
2,3,7,8-TCDF
MPR
RSD
MPR
RSD
MPR
RSD
89.6%
9.1%
92.8%
5.47%
92.4%
6.20%
64.7%
2.95%
70.8%
5.92%
74.4%
6.86%
81.2%
1.91%
93.6%
7.91%
99.1%
5.43%
93.7%
5.71%
100%
2.59%
112%
9.20%
88.9%
3.90%
92.9%
11.2%
99.4%
8.95%
MPR = Mean Percent Recovery
RSD = Relative Standard Deviation
1—95
-------
TABLE III
Distribution of TCDD and TCDF between the Filter and PIJF
Percent Found on the Filter and PUF
Sampling Time (Minutes)
1 1
17.5
144
1440
TCD O Filter 100
PUF 0
TCDF Filter 63
PUF 37
97
4
10
90
82
18
0
100
1-96
-------
Figure 1
A diagram of the site. During pollution abatement activities, approximately
6” of soil would be removed from a section (5,000 square feet) with a backhoe
and placed in plastic bags. The new surface would then be tested for dioxin
and be either declared clean or another 6” of soil would be removed. Air
pollution could originate either from the soil handling operations or truck
traffic within the site. The six monitoring locations relative to the
contaminated area are shown in the figure (M, 0, D, K, L & E). The
prevailing wind is from Monitor E to Monitor 0.
Figure 2
Modified Hi—Vol Sampler (General Metal Works, Inc., Model PS—l) used to
collect 2,3,7,8—TCDD samples.
Figure 3
A detailed illustration of the sampling head module (see Figure 2) showing
the fiberglass filter used to collect particulates, and the glass
cylinder/PUF used to collect vapors.
Figure 4
Data from samples collected from a single monitor with both valid PM and
2,3,7,8—TCDD concentrations above the respective detection limits are shown.
There is no reliable correvlation.
Figure 5
The data points are representative (we only show every 5th day for clarity)
averages of measured or estimated (see text) concentrations for the most
recent 14 days. A downwind sample is one in which the monitor (0 in Figure
1) was downwind of the clean—up activity at the site during the last day of
the 14—day averaging period. A non-sampling value is shown for those days
when the concentrations at site llol were estimated rather than measured on
the last day of the averaging period.
The average of all measured (or estimated) concentrations for 14 consecutive
days is plotted on the Y axis against the last day of the 14 day averaging
period on the X axis for monitor “0” (see Figure 1). Detection limits were
taken as measured concentrations. If a measurement was not taken for a given
day because the monitor was not downwind or upwind on that day, an estimate
of the concentration was made by averaging all of the non—downwind values to
date. Since conc ntrations above the detection limit (usually in the range
0.4-0.8 picogram/M ) were obtained only when a monitor was downwind of the
site, all estimated values were averages of detection limits. The data are
presented in this manner to illustrate our best estimate of the maximum
exposure of any of f— ite population that might have resulted from this clean-
up activity.
1—97
-------
I
MONITORS
CONTAMINATED
AREA
S
ITE
MAP
0
D
.
.
K
L
M—E
E
SAMPLE
LOCATION
-------
M a gn e he tic
Gauge 0.100 in.
Exhaust
Duct
(6 j 10
Voltage Variator
Elapsed Time Meter
Modified Hi-Vol Sampler
(General Metal Works, Inc., Model PS-i)
used to Collect 2,3,1,8,-TCDD Samples.
Sampling
Head
Venturi
Flow
Control
Value
7-Day
Timer
Base Plate
1-99
-------
A Detailed Illustration of the Sampling Head Module Showing the Fiberglass Filter
Used to Collect Particulates and the Glass Cylinder PUF Used to Collect Vapors.
3
Lower Canister
Glass Cartridge
Retaining Screen and Put Plug
Filter Holder Support
0
0
Filter Holder
With Support
Screen
PUF
4” Diameter Filter
Silicone Rubber
Gaskets
FILTER
Filter Retaining Ring
Silicone
Rubber
Gasket
-------
I -
a)
a)
C)
0
I -
a)
C l)
E
Cu
I .
I .-
C)
I -
Cu
a)
(‘3
C-)
I-
Cu
a-
SAMPLES FROM QUAIL RUN WITH 2,3,7,8-TCDD
VALUES ABOVE THE DETECTION LIMIT
C.C = 0.5021
2.0
7.0
350
300
250
200
150
100
2,3,7,8-TCDD
*
*
*
*
*
*
*
0
*
1.0
3.0
4.0
SM
6.0
(Picograms per Cubic Meter)
-------
5
14 DAY RUNNING AVERAGE
7.00
6.00 —
NO OBSER VED EFFECT LEVEL
QN0N-
500 DOWNWIND
0 DOWNWIND
I x
CZ i • NON—
SAMPLING
-.cv,.
3.00 — ACTION LEVEL — — —
0
04 rn 1.00 _ I 0 + I
Ann , I I I I __ I I I I I I I I
4’
DATES
-------
VOC EMISSION RATES FROM SOLID WASTE LANDFILLS
W. Gregory Vogt, Senior Project Scientist, Lisa Y. Montague, Staff
Scientist, Peter J. Carrico, Associate Staff Scientist, Joyce K.
Hargrove, Associate Staff Scientist, SCS Engineers, Reston, Virginia
ABSTRACT
Regulatory agencies and the public have directed increased attention
at the health risks associated with exposure to trace constituents
present in landfill gases. Direct field measurements of surface
emissions are difficult and often inconclusive, indicating a need
for quantitation by alternative means To this end, a study was
conducted to estimate the maximum surface and subsurface emissions
for volatile organic compounds (VOCs) from RCRA Subtitle D disposal
sites. The purpose of these estimates was to provide baseline data
for assessing potential air quality impacts.
Estimates were based on measurements from pump test programs common
to the landfill Gas industry. Field pump tests performed from 1982
through 1986 at 20 selected landfill sites provided gas generation
rates for wet and dry climates per volume of landfilled solid
waste. The median LFG generation rates for wet and dry geographic
regions were 207 cubic feet of gas per cubic yard of refuse per year
(cf/cy/yr) and 81 cf/cy/yr, respectively. In addition, VOC
concentrations were derived from a compilation of laboratory
analyses from sites between 1982 and 1986. Various landfill
configurations were evaluated on the basis of surface area, waste
depth, location of disposed wastes (above ground, below ground, and
above and below ground), and region.
Results from the study indicated that the solid waste landfill
emissions are greater for sites in wet climate regions due to the
amount of available moisture required for biological activity.
However, VOC concentrations do not change considerably with
climate. Total gas emissions in wet climates ranged from 124
million to 2,960 million cubic feet per year for the various
landf ill configurations. VOC mass surface emissions for the same
landfills ranged from 1 to 22 tons/year.
APPROACH
Regulatory agencies and the public have directed increased attention
at the health risks associated with exposure to trace constituents
present in landfill gases. Correlations of landfilled waste types,
landfill gas (LFG) characteristics, and direct field measurements of
surface emissions are difficult and often inconclusive, indicating a
need for quantitation by alternative means. To this end, a study
was conducted to estimate the maximum surface and subsurface
emissions for volatile organic compounds (VOCs) from RCRA Subtitle D
1—103
-------
disposal sites. The purpose of these estimates was to provide
baseline data for assessing potential air quality impacts.
Estimates were based on measurements from pump test programs common
to the landfill gas industry. Field pump tests performed from 1982
through 1986 at 20 selected landfill sites provided gas generation
rates for wet and dry geographic regions per volume of landfilled
solid waste. In addition, VOC concentrations were derived from a
compilation of laboratory analyses from sites between 1982 and
1986. A range of landfill configurations was assigned and evaluated
on the basis of surface area, waste depth, location of disposed
wastes (above ground, below ground, and above and below ground), and
region. Estimates could then be made to calculate VOC emission
rates from all landfill configurations and for wet versus dry
climatic areas. Our approach for determining these landfill
emission rates is depicted In Table 1.
COMPILE LANDFILL GAS GENERATION RATES
Landfill test data were compiled to determine typical LFG generation
rates and VOC emission rates per volume of in—place refuse. Pump
test programs, which provide empirical data for sustained LFG
production rates over a period of several days or weeks, were
surveyed from field studies conducted from 1982 through 1986 at
municipal solid waste landfills nationwide. On the basis of refuse
moisture contents, methane content, estimated landfill life (for LFG
production), and refuse density, the gas generation rate was
determined, expressed in units of cubic feet (cf) of raw landfill
gas per cubic yard (cy) of in—place waste per year (ef/cy/yr).
Available moisture within the landfill Is required for biological
activity to occur and hence, the moisture content of the in—place
wastes directly affects LFG generation rates. As a result,
geographical regions having different net precipitation!
infiltration rates will produce LFG at different rates. The field
studies were separated into wet and dry regions based on geographic
location and precipitation rates. For instance, southern California
is a relatively dry region with an average rainfall of about 20
Inches per year. By comparison, the Midwest and East Coast are
relatively wet regions with an average annual rainfall over 35
Inches. Wet regions have generally higher gas generation rates and
refuse moisture contents, but have estimated production “lives”
lower than those for dry regions. A general relationship is that
the greater the net precipitation/infiltration rate, the greater the
gas generation rate, and the shorter the overall duration of LFG
production (or landfill life).
Results from the compilation provided median gas generation rates
for wet geographic regions (207 cflcy/yr) and dry geographic regions
(81 cf/cy/yr).
1-104
-------
TABLE 1
1 10W VOC EMISSION RATES WERE DETERMINED
1. Compile LFG Generation Rates from Field Test Data
o RCRA Subtitle D disposal sites
o Nationwide pump test programs
o Wet versus dry geographic regions
2. Compile Laboratory Analyses for VOCs Present in LFG
o RCRA Subtitle D disposal sites
o Major VOCs found in landfill gas
o Median concentrations
3. Select Range of Landfill Configurations
o Surface Area
o Waste Depth
o Location of disposed wastes (above or below ground, or both)
o Emissions pathway (surface or subsurface, or both)
4. Calculate VOC Emission Rates
o All landfill configurations
o Total gas and total VOCs
o Wet versus dry geographic regions
1—105
-------
COMPILE LFG ANALYSES FOR VOCs
A literature search was conducted to compile laboratory analyses for
VOCs present in landfill gas. While published reports provided some
pertinent data, most of the data were collected directly from
laboratory analyses. Sources for many of these analyses were
engineering firma which had conducted nationwide landfill gas test
programs, monitoring of gas extraction systems, or general
environmental investigations. Various references were assembled and
reviewed with regard to site descriptions, landfill gas collection
systems, sampling and analytical methods, detection limits, and
other information. These provided an initial compilation of all
trace constituents sampled in landfill gas
Selection criteria were established for references and specific
sampling events in order to derive a reliable data base; i.e., one
representative of solid waste landfills with active biological gas
production. The analytical results included in the data base met
the following criteria:
o Landfill gas samplea were taken from landfills that
reportedly received only non—hazardous solid wastes and
essentially were regulated as R RA Subtitle D disposal
sites. However, hazardous wastes from small quantity,
generators or within household hazardous wastes may have been
disposed at these landfills.
o Landfill gas samples were taken after 1982. Sampling and
analytical results obtained prior to that time were not
included in the data base due to questionable or inadequate
sampling methods and limitations of available analytical
techniques.
This data base identified 87 VOCs present in LFG on the basis of
samples taken and frequency of detection. However, many of these
compounds were sampled and detected at less than five of the 35
surveyed sites. Consequently, a list of the typlcal” LFG
coast ituents was derived for those compounds found in at least 50
percent of the total samples analyzed and in at least 20 percent
of the sites surveyed. This list includes the 16 VOCs presented
in Table 2. As shown, the range, median, and mean concentrations
are provided for comparison purposes. Note that the median values
are always less than the mean values, indicating that the reported
data are skewed, presumably due to one or more high readings for
each compound.
The permissible exposure level (PEL) established by OSHA Is
included in Table 2 for each listed compound. Vinyl chloride has
the lowest PEL at 1 volume part—per—million (Vppm) and It is the
only compound where the PEL was exceeded by most of the samples.
1-106
-------
TABLE 2
I .
CD
TYPICAL TRACE CONSTITUENTS IN LANDFILL GAS
Compound
Range of
Concentration
(Vppm)
Median
Concentration
(Vppm)
Mean
Concentration
(Vppm)
Standard
Deviation
(Vppm)
PEL
(Vppm)
Methylcyclohexane
Acetone
0.02 - 19
0.05 - 12
3.3
4.2
7.2
4.8
7.7
3.5
500
1000
Trichiorofluoromethane
0.26 - 8.8
1.4
2.5
2.8
1000
n-Hexane
0.001- 31
4.9
7.8
9.4
500
1,1-Dichioroethane
0.06 - 20
0.65
4.2
6.4
100
1,2-Dichioroethene
0.11 - 95
1.7
9.1
21
200
1, 1-Dichioroethene
0.03 - 4.2
0.21
0.8
1.2
5
Methylene Chloride
0.079-380
3.8
32
74
500
Xylenes
0.001-110
5.8
16
25
100
11,1-Trichioroethane
0.003- 31
0.057
1.8
5.8
350
Ethylbenzene
0.012- 91
1.7
6.2
16
100
Trichioroethylene
0.011- 44
0.51
3.1
6.8
100
Vinyl Chloride
0.03 - 50
4.7
6.9
8.9
1
Tetrachioroethylene
0.006-190
0.79
11
27
100
Benzene
0.016- 24
0.54
2.0
3.5
10
Toluene
0.004-360
1.4
27
53
200
* Permissible [ xposure Level prescribed by OSFIA for workplace exposure.
-------
In addition to compound—specific identifications and concentrations
from the analytical data base, ‘values for total non—methane VOCs
were obtained from the surveyed sites. The median concentration of
non-methane VOCs was found to be 58.5 Vppm, which includes compounds
In addition to those listed in Table 2. VOC emission rates are
typically reported in units of mass per volume of air (or gas).
Consequently, the median value of 58.5 ppm was converted to a
mass/volume basis for VOC emissions calculations. A value of 58.5
ppm is equivalent to 0.000015 pounds (lba) of total VOCs per cubic
foot of landfill gas, assuming an average VOC molecular weight of
100 grams/mole.
SELECT lANDFILL CONFIGURATIONS
The next step for the determination of VOC emission rates was to
select representative landfill configurations. Twelve landfill
scenarios were selected to provide a general range of landfill sites
that typify existing disposal sites. The configurations differed on
the basis of surface areas (20 and 100 acres), total waste depth
(20, 40, 50 and 100 feet), and location of disposed wastes (above
and below ground, all below ground, and all above ground). The
configurations were consistent on the basis of shape (each landfill
was square, with length equal to width) and side slopes (above
ground side slopes were 3:1, and below ground side slopes were 2:1).
FIgure 1 illustrates the generic landfill configurations, including
side slopes and locations of wastes. Note that subsurface areas
with regard to emissions are based on the vertical projection of
each subsurface side slope. This is because subsurface emissions
were assumed to travel horizontally and generally not in a downward
direction. The landfill volumes range from 592,000 to 14,320,000 cy
of refuse.
Except for the above ground landfills, surface areas greatly exceed
subsurface areas, principally because gas emissions through the
bottom of sub8urface portions of the landfills were not considered.
For the above ground landfills, however, the subsurface gas emission
pathway did consider migration through the bottom of the landfill
and thus, the surface and subsurface areas for emissions were about
the same.
CALCULATE EMISSION RATES
A series of algebraic calculations was performed in order to
determine total emissions (raw landfill gas and VOCa) from the 12
landfill configurations. After a surface area (In acres) and a
depth were assigned for a particular landfill configuration, areas
and volumes relative to gas emissions were calculated. Total gas
emissions (cffyr) were determined for each configuration by the
1-108
-------
Area 2
L fll dI
I
b
C
11
2
ABOVE AND BELOW GROUND
length of top surface
length of landfill at natural ground level
length of bottom of landfill
depth of disposed waste where d 1 + d 1 = d
subsurface area of vertical projection of each side slope (below ground)
surface area of each side slope (above ground)
Figure 1. Landfill Configurations
3
ABOVE GROUND
Area 1
BELOW GROUND
Ii
2
)‘
WHERE:
a =
b =
C =
d, d 1 =
Area 1 =
Area 2 =
1-109
-------
product of the in—place waste volume (cy) and the gas generation
rates (cflcy/yr) for both wet and dry regions. Because there Is a
direct relationship between total LFG emissions and total VOC
emissions, the total VOC emission (ibs/yr) were determined simply as
the product of the total LFG gas emissions (cflyr) and the median
VOC generation rate (lbslcf).
Algebraic calculations were also performed to determine surface and
subsurface flux rates. Flux rates are volumes of landfill gas
emitted per surface (or subsurface) area of the emissions pathway
per year (cf/sf/yr) for the VOCs, flux rates are in units of
lbs/sf lyre Surface flux rates were defined as the maximum potential
emission rates through surface routes or pathways. When determining
surface flux rates, we assumed that all of the gases generated were
emitted through the surface pathway. That is, no portion (zero
percent) of the landfill gases or VOCa are emitted through
subsurface pathways Similarly, the determination of subsurface
flux rates assumed that all the landfill gas and VOCe are emitted
only through subsurface pathways, with zero percent surface
emissions.
Tables 3 and 4 present estimated surface and subsurface flux rates
for landfill gas emissions and for the non—methane VOCs. At most
landfills, a portion of the gases are emitted through the surface,
and a portion are emitted through subsurface pathways. Furthermore,
a percentage of the VOCs are absorbed by the soil on the landfill
cap and/or the subsoil (for subsurface migration). Accordingly, the
numbers presented in Tables 3 and 4 should be considered “worst
case” values for typical solid waste landfills.
RESULTS
Table 3 presents estimated LPG emissions and potential flux rates
for the 12 landfill configurations. In addition, emissions were
calculated for both wet and dry climates. Based on in—place waste
volumes, total gas emissions ranged from 49 million to 1,160 million
cy/yr for the dry climates. Total gas emissions for the same
landfill configurations In wet climates increased by a factor of
about 2.6 (range 124 million to 2,960 million cf/yr). The
greatest gas emissions were generated in the largest landfills (100
acres with 100 feet of total waste depth).
These total gas emissions are derived for raw landfill gas, which
consists of about 50 percent methane and 50 percent carbon dioxide.
Thus, a reliable estimate for methane emissions Is simply half the
value shown in Exhibit 3. As an example, the landfill configuration
with wastes disposed 20 feet below ground, 20 acres In a wet
climate, would emit approxImately 62,000 million ef of methane gas
per year.
1-110
-------
TABLE 3
Surface
Area
(ac)
Waste
Volume
(NM cy)
ESTIMATED GAS EMISSIONS FROM SELECTED LANDFILL CONFIGURATIONS
Dry Climate
Waste
Depth
(It)
Total Gas
Maximum
Potential Flux
Emissions
Surface Only’
Subsurface
Only”
(MN cf/yr)
(cf/sf/yr)
(cf/sf/yr)
I. Wastes Disposed Below Ground
Wet Climate
Total Gas
Maximum
Potential Flux
Emissions
Surface Only’
Subsurface
Only”
(NM cffyr)
(cf/sf/yr)
(cf/sf/yr)
20
20 0.6
49
56*
685
124
142*
1,730
20
50 1.3
105
121
629
269
309
1,610
100
20 3.1
251
58
1,530
642
147
3.910
100
50 7.3
591
136
1,490
1.500
344
3 7ao
II.
Wastes Disposed Above
Ground
20
20 0.6
49
56*
56”
124
141*
142”
20
50 1.2
97
108
111
248
277
285
100
20 3.0
243
55
56
621
142
142
100
50 7.0
567
128
130
1,450
328
333
III.
Wastes Disposed Above
and Below Ground
20
40 1.2
97
110*
1,360”
248
281’
3,47Ø**
20
100 2.5
203
227
1.220
518
578
3,100
100
40 6.2
502
115
3,060
1.280
292
7,800
100
100 14.3
1,160
262
2,920
2,960
670
7,500
* Assumes all emissions are through landfill surface.
** Assumes alt emissions are through sides of landfill.
Assumes all emissions are through bottom of landfill.
-------
Surface
Area
(ac)
Waste
Vo I
(NM cy)
TABLE 4
ESTIMATED VOC EMISSIONS FROM SELECTED LANDFILL CONFIGURATIONS
Dry Climate
Waste
Depth
(ft)
Total VOC
Maximum
PotentIal Flux
Emissions
(Ibs/yr)
Surface Only
(tbs/sf/yr)
Subsurface Only
(Lbs/sf/pr)
I. Wastes Disposed Below Ground
Wet Climate
Total VOC
Emissions
(Lb S/pr)
Nax mum
Potential Flux
Surface Only
(lbs/sf/pr)
Subsurface OnLy
(lbs/sf/pr)
20
20
100
100
20
50
20
50
0.6
1.2
3.0
7.0
20
20
100
100
20 0.6
50 1.3
20 3.1
50 7.3
740
i ,ooo
3,800
8,700
0.OOO85
0.0018
0.00087
0.0020
0.010*0
0.010
0.023
0.022
1,860
4,040
9,600
22.500
0.0021*
0.0046
O.0ofl
0.0052
0.0260*
0 024
0:059
0.057
II.
Wastes Disposed Above
Ground
740
1 500
3,600
8,500
O.00084
0.00167
0.00082
0.00192
O.O0085
0.00172
0.00083
0.00195
1,860
3 ,720
9,320
21,750
0.0021*
0.0042
0.0022
0.0069
0.00210*0
0.0043
0.0021
0.0050
III.
Wastes Disposed Above
and Below Ground
20
20
100
100
40 1.2
100 2.5
60 6.2
100 14.3
1,370
3,050
7,530
17,400
0.0016*
0.0034
0.0017
0.0039
0.019*0
0.0 18
0.046
0.044
3.720
7.770
19,200
44,400
0.0042*
0.0087
0.0044
0.0100
0.0520*
0.047
0.117
0.112
* Assumes all emissions are through Landfill surface.
0* Assumes all emissions are through sides of landfill.
• a Assumes all emissions are through bottom of landfill.
-------
Flux rates shown in Exhibit 3 differed significantly relative to
emissions pathways, differed to a lesser degree relative to landfill
configuration, and differed by a factor of about 2.6 for wet versus
dry climates. In all cases, the subsuface emissions pathway yields
the higher potential flux rate. Except for the above ground
landfills, the subsurface flux rate greatly exceeded the surface
flux rate, generally by factors from 5 to 26. For the above ground
landfills, however, the surface and subsurface flux rates were
similar.
Landfill configuration affects the resultant emissions flux rate but
not to a large extent. That is, if waste depth is increased, there
is a direct proportional increase in surface flux rates and, for
above ground landfills, in subsurface flux rates. If surface area
is increased as much as five—fold, the surface flux rate remains
about the same. However, this same five—fold increase in landfill
surface area resulted in a more than two—fold increase in the
subsurface emissions flux rate. The exceptions to this relationship
were the above ground landfill configurations, where the subsurface
flux rate did not change significantly with surface area.
Exhibit 4 presents estimated VOC emissions and potential flux rates
for the same landfill configurations. Because there is a direct
proportional relationship between estimated LFG emissions and total
VOC emissions, the impacts of variables such as surface area, waste
depth, and climate are the same as those discussed above. The VOC
emissions were calculated on the basis of mass emitted annually
(lbs/yr), and ranged from 740 to 17,400 lbs/yr for the landfill
configurations located in dry climates. Total VOC emissions for the
same landfills in wet climates increased by a factor of about 2.6
(range = 1,860 to 44,400 lbs/yr).
CONCLUSIONS
Compilation of literature sources and empirical data from landfill
test programs provided estimates of landfill gas and VOC emission
quantities for a variety of landfill configurations. Subsequent
calculations provided corresponding surface and subsurface flux
rates. Our conclusions are as follows:
1. Total landfill gas emissions ranged from 49 to 2,960 million
cf of gas per year. Since landfill gas consists of 50
percent methane, annual methane emissions can be estimated to
range from 24 to 1,480 million cf of methane per year.
2. Total VOC emissions ranged from 740 to 44,400 pounds per year
(0.37 to 22.2 tons per year), assuming a molecular weight of
100 for the non—methane VOCs present i t t landfill gas.
1—113
-------
3. Landfill emissions are greater for sites located in wet
climate regions due to the amount of available moisture
required for biological activity. In general, gas and VOC
total emissions were estimated to be about 2.6 times higher
in wet climates versus dry climates.
Our emission calculations were based on several simplifying
assumptions which were applied to a limited number of landfill
configurations. For example, a single gas emission rate for VOCs
provides a rough measure of actual emissions at landfill sources.
Similarly, the use of only two gas generation rates (for a wet and
dry climate) is a generalization. In addition, surface and
subsurface flux rates normally do not occur in the absence of the
other.
1—114
-------
A CONTROL CHART STRATEGY FOR
GROUND WATER MONITORING
George T. Flatman, Exposure Assessment Research Division,
Environmental Monitoring Systems Laboratory, Las Vegas, Nevada;
Thomas H. Starks, Environmental Research Center, University of
Nevada — Las Vegas, Las Vegas, Nevada
ABSTRACT
Monitoring the groundwater around a waste impoundment site presents
a statistical problem in testing compliance or leaking. Two types
of errors are present in statistical tests, and both are
undesirable. If the test is designed for early detection, there may
be too many false positives. This means an out—of—control state is
declared when It does not exist and unnecessary increased sampling
Is started. In contrast, if the test is designed for conservative
detections, there may be too many false negatives. That means an
in—control state is declared when an out—of—control state exists and
pollution takes place undetected.
One commonly used statistical procedure that suggests an answer for
this type of problem is control chart methodology. This poster
session explains control chart decision logic and applies two of the
more promising types to actual data. The Shewhart chart is designed
to detect a large, binary leak from a seasonal or random
fluctuation. The Cusum chart is designed to detect a growing leak.
A monitoring strategy that uses a combination of these two charts is
discussed.
INTRODUCTION
Under the Resource Conservation and Recovery Act of 1976 (RCRA), the
U.S. Environmental Protection Agency has developed regulations for
landfills, surface impoundments, waste piles, and land treatment
units that are used to treat, store, or dispose of hazardous
wastes. These regulations include requirements for the monitoring
of ground water in the top aquifer below the hazardous waste site
(HWS). This monitoring Involves the drilling of wells into the
aquifer up—gradient and down—gradient of the HWS, and the sampling
and analysis of well water at regular time Intervals to help
determine whether leachate from the HWS has entered the aquifer.
There are several as yet unsolved problems in this monitoring
program. These problems include determination of appropriate
methods for obtaining accurate measurements of some constituents
such as volatile organics, specifications for well construction, and
detection and accommodation of shifting direction and rate of
aquifer flow. However, this paper discusses the problem of
developing good decision rules for determining when additional
regulatory action may be required and recommends that the
1—115
-------
development be based on a realistic model for the ground—water
measurements. Industrial quality control schemes are considered in
terms of their possible application to the ground-water monitoring
decision problem.
OUND-WATER CHARACTERISTICS
To develop a reasonable model for the measurements of GQP, some
discussion of ground-water characteristics is required. The
horizontal velocity of water in an aquifer is slow (a few meters per
day is considered a high velocity), and lateral dispersion of a
contaminant is considerably slower than fluid flow; this means that
plumes do not widen to any great extent as they extend through the
aquifer (Freeze and Qierry, 1979). At a particular sampling time,
the water at different wells near a HWS will have entered the
aquifer at different times and therefore will carry different
concentrations of various monitored constituents. Hence, the value
of a monitored parameter may be increasing at one well while it is
decreasing at another. Concentrations of contaminants in samples of
ground-water may also change because of changes in water table. A
high water table resulting from recent rains or flooding may cause a
reduction in measured concentrations of a pollutant because of
dilution of the contaminant or because the contaminant is floating
at the top of the aquifer which is now above the well screen.
Therefore, changes in the water table may have similar and virtually
Simultaneous effects on all wells near a HWS.
PROBL 4S ASSOCIATED WITH THE ANALYSIS OF GROUND-WATER DATA
There are several characteristics of ground—water monitoring data
that Beverely complicate the development of statistical decision
procedures. They are as follows:
1. Qiemlcal analysis of water samples is expensive (sometimes on
the order of $1,000 per sample);
2. Several water quality parameters are monitored;
3. anges in aquifer water quality and flow characteristics are
effected by human intervention off the HWS such as
intermittent pumping of water from the aquifer or accidental
spills of pollutants into the aquifer;
4. For some monitored substances such as volatile organic
compounds, measurement error variance and bias tend to be
large
5. For some monitored substances, the chemical analyses will
result in “below instrument detection limit” or “not
detected” for most of the water samples submitted; and
1-116
-------
6. It is difficult to obtain consecutive samples from a well at
a particular sampling time that are true replicates because
of a tendency for such observations to follow trends.
High costs of analyses and large measurement variances make it
difficult to obtain an adequate number of background measurements to
characterize the system. Similarly, it is costly to obtain a
sufficient number of samples and aliquots at a monitoring well to
estimate current GQP values with high precision or to provide tests
with good power characteristics.
The fact that the values of several water quality parameters are
being monitored at each of several down—gradient wells implies that
in each sampling period, many decisions must be made as to whether
parameter values at down—gradient wells have increased sufficiently
to justify some form of regulatory action. If a separate test of
hypotheses is performed for each of these decisions, the probability
of falsely rejecting at least one null hypothesis is likely to be
quite high even though the probability of such a false—positive is
low for each test. While the simplest and most typical reaction to
a large concentration that signals the occurrence of an unusual
event is to take another sample to confirm the measurement, one
cannot be complacent about false alarms since the cost of taking
unnecessary additional samples can be a heavy burden for the HWS
owner. To reduce the number of tests, one might think of using a
multivariate procedure such as Hotelling’s T 2 —test (Anderson, 1984),
but such tests usually require estimation of the covariance matrix
of the vector of measurements. Unfortunately, the number of
sampling periods required to obtain the data to estimate the
covariance matrix is greater than the dimension of the observation
vector. Costs and start—up time restrictions make it unlikely that
the covariance matrix could be adequately estimated from
measurements taken prior to start—up of the HWS. In addition, human
intervention is likely to change the covariance matrix of the
observations during monitoring. Beyon 1 these problems with the
covariance matrix, a significant large T —value does not necessarily
imply an unusual increase in any of the monitored chemical
concentrations.
Another and probably better approach to the multiple tests problem
is to use the Bonferonni simultaneous inference technique (Miller,
1966) which prescribes the use of a significance level of o /m for
each of the m tests to be performed at a given time of sampling so
as to keep the overall probability of making at least one false
positive (Type 1) error at or below o Naturally, the reduction of
the significance levels of the individual tests from to &m also
reduces the power of the tests.
1-117
-------
The ‘below instrument detection limits” measurements of a water
quality parameter are impossible to employ in the types of
statistical analyses and decision rules mentioned above. While In
many cases these qualitative statements can be replaced by
instrument values obtained in the chemical analysis, these numbers
are likely to have a considerably different error distribution than
will observations that are above the instrument detection limits.
The decision as to whether an above—detection—limit measurement
represents an increase in the mean concentration of the chemical
over previous “not detected” readings Is often a question which must
Involve the analytical chemist because of his knowledge of the
quality of the current measurement. Statistical methods proposed by
Aitchison (1955) for positive continuous random variables with
positive probability mass at zero may be useful here; these same
statistical methods were considered in terms of air monitoring by
Owen and DeRouen (1980).
An inability to obtain true replicate samples makes it impossible to
estimate sampling variance and difficult to obtain desired levels of
precision for quality estimates.
CONTROL CHART SCHEMES
If the slow flow of water and the narrowness of plumes in the
aquifer make the tests of means and interactions Inappropriate and
It outlier tests are lacking In power, what other decision
procedures might one apply? An approach suggested by Vaughan and
Russell (1983) for monitoring effluent from waste treatment plants
is to use industrial quality control schemes. Such schemes compare
an observation with the observations that came before it at that
location.
A prime consideration in using industrial quality control methods to
monitor ground-water quality at a HWS Is that It separates effects
of location and well construction from the decision process.
Instead of comparing the concentration of a monitored chemical at a
particular well with measurements at other wells, one compares the
current measured concentration with the past history of measured
concentrations of the chemical in water from this well. Some other
advantages and also some disadvantages related to the earlier list
of problems with ground-water measurements, will become evident as
the nature of quality control schemes is presented.
Discussion In this section is restricted to one—sided control
schemes because the common concern In monitoring is to detect an
Increase in pollutant concentration. The extension to two—sided
schemes Is straightforward if they are needed for an indicator such
as pH.
1-118
-------
SHEWHART CHARTS
The Shewhart (1931) quality control chart is one of the oldest and
simplest of the industrial quality control procedures. The chart is
simply a graph of time of sampling, or sample number if samples are
equally spaced in time, versus the sample mean value for the quality
parameter being monitored. Time, or sample number, is the abscissa
and sample mean value is the ordinate of a point on the graph.
Typically the horizontal axis is positioned so as to intersect the
vertical axis at the steady—state mean value, p for the quality
parameter. A horizontal line is also drawn to intersect the
vertical axis at ,u+Z where Z is the upper cC quantile of the
standard normal distribution and cY is the long—run standard
deviation of the sample means. This line is called the upper
control limit, and when a point falls above the line, the process is
declared out of control. The average in—control run length (i.e.,
the average number of samples between declarations that the system
is out of control, when in fact it is in control) is l/ . if the
sample means have a normal sampling distribution. The commonly used
value of Z is 3, for which the corresponding value of o - is
0.0013. In industrial quality control, the sample sizes are usually
somewhere between 5 and 10 depending on cost and internal
variability between members of a sample.
Loreazen and Vance (1986) give a procedure for determining on an
economic basis the sample size n, Z, and the time between samples.
It would appear that their approach could be generalized to other
control schemes and to ground—water monitoring situations.
A second control chart is often kept for the variability of the
product. It is similar to the Shewhart chart for the sample means,
only now the ordinate is the sample range, or standard deviation,
and the horizontal lines representing upper and lower control limits
are located on the basis of the distribution of the statistic
(sample range or sample standard deviation) under the assumptions of
a normal distribution for the quality parameter measurements. In
practice the lower limit is seldom used (Guttman et al., 1982).
This chart is not nearly as robust with respect to the assumption of
normality as is the chart for the sample means, and so
out—of--control situations for variability must be viewed with more
Skepticism than similar results on the means chart. If the
variability of measurements of the quality parameter changes, the
height of the upper control line on the Shewhart chart for means is
adjusted accordingly, or action is required to bring the variability
back to its previous level.
THE CUSUN QUALITY CONTROL SCHEME
The CUSIJM (for cumulative summation) control scheme derives from a
paper by Page (1954) and is somewhat more complicated than the
1-119
-------
Shewhart chart. (A review article by Lucas (1958a) gives the
current state of development of this procedure). The CUSIJM control
scheme makes use of information in the present sample and in the
previous samples in reaching decisions as to whether the process is
In control, whereas the Shewhart chart makes decisions on only the
current observation (i.e., the Shewhart chart Is a graphical
representation of a sequence of individual tests of the mean,
whereas the CUSIJN scheme Is a sequential probability ratio test of
the mean). The one—sided CUSUM scheme involves the computation of a
cumulative sum S which for the Ith sample is given by the formula.
Sj maxf 0, z 1 — k + si —i
where zj Is the standardized ith sample mean (I.e., zj ij — u3/cr)
and k is a parameter of the control scheme. When Sj exceeds a
specified value h, the process is declared out of control (i.e., In
pollution monitoring a decision is made to begin additional
monitoring activity). The values of h and k are chosen to obtain
desired average run lengths (ARL) under in—control and specified
out—of—control situations. For a scheme designed to be sensitive to
changes in mean quality of size 0 , k is usually chosen to be D/2,
and ii is selected to give the largest in—control ARL consistent with
an adequately small out—of—control ARL consistent with an adequately
small out—of—control ARL. Typically SO Is taken to be 0; however,
Lucas and Crosier (1982a) have suggested that by employing a
slightly higher value of h and starting with Scj=h12, one can
decrease out-of-control ARL while maintaining in-control ARL. They
also give tables of ARL for In—control and out-of—control situations
and various values for h and k. However, if one is reasonably
confident that there Is no contamination coming from the HWS at the
start, then one can start with SçjO and can somewhat decrease the
chance of an early false positive.
Example 2:
To illustrate the Shewhart and CUSUM schemes, random normal
(u =1.0, 62=4) deviates were drawn from a table of such numbers.
To illustrate an out-of-control situation, 2 was added to each of
the numbers drawn after the fourth (1=4). These numbers are to
represent the sample means of a process that went out of control
between the fourth and fifth samplings. The in—conttgl situation
has sample means distributed N (10,4), 80 zj (X 1 — u)/o
(Xj — 10)/2. The US1JM scheme (Table 2) indicates that a
decision to take action (process is out of control) should be
made at ilO. The CUSUI4 chart (Figure 1) gives a visual
impression of when the process went out of control. The
corresponding Shewhart chart with an upper control limit of
(u +36)16 is given in Figure 2.
This example illustrates the weakness of the standard Shewhart chart
in detecting small changes in the value of the mean. One way to
1—120
-------
TABLE 2. CUSUM QUALITY CONTROL SCHEME
18
16
34
X 12
10
8
6
(k=0.’ . h=5)
S S
S
1 I I I I I
5 6 7 8 9 10
1—1 1 1 I I I I
2 3 4 5 7 8 9 10
Figure 1. Th€ CUSUN ccrntroi chart
(k 0.5, h-5)
1—121
1
x
In—Control
z
S
x
Out—of--Control at 1=5
z
0
0
1
1 ) 1.50)1
2.252
1.752
1 14.5O l
O.55 l
2
11.108
O.55 4
1.806
11.108
—1.203
3
7.59)1
—1.203
0.103
7.5914
—1.210
14
7.580
—1.210
0.000
7.580
5
11.588
0.79)1
0.29 14
13.588
1.79)1
2.001
6
12.002
1.001
0.795
114.002
1.217
7
10.14314
0.217
0.512
12.14314
0.689
8
9.378
—0.311
0.000
11.378
1.3514
9
10
10.708
11.278
0.3511
0.639
0.000
0.139
12.708
13.278
1.632
S
0
1.752
1.806
0.103
0.000
1.29)1
2.595
3.312
3.501
14.355
5.L 19 ) 1X
— a — —
S
liii
1 23 i
• .
Figure 2. Shewbart qua]ity control chart
(1110, 2)
— a a a a a
S
S
• S
•
6
5—
4—
3
1
I
3
-------
reduce this weakness in the Shewhart chart is to declare the process
out of control whenever there are r successive sample means with
value above u. The value of r is usually chosen to be 7 or 8.
However, this procedure also reduces the in—control ARL.
THE COMBINED SHEWRART CUSTJM SCHEME
The Shewhart scheme is better than the CUSUM scheme in quickly
detecting large ( 3o ) shifts in the mean ,a, whereas the CUSUM
scheme is usually faster in detecting a small change in that
persits. Bisseji (1984) has also shown that the CUSUM scheme is to
be preferred when the mean is increasing In a linear time trend. To
take advantage of the good properties of both tests, Lucas (1982)
suggested combining the two procedures. This is accomplished by
declaring the process out of control if the JSUM S 1 is above
specified upper Shewhart limit or if the CUSUM Sj is above a
specified limit h. To keep a reasonable in—control ARL, Lucas
suggests using an upper Shewhart control level of (p +4 ). Lucas
calculated that if this upper Shewhart control limit is used with
CUSTJM parameter values k’0.5 and h”5, the In—control ARL is 459
while the out-of—control ARL is 10.4 if the true mean shifted upward
by o and only 1.6 if the mean shifted upward by 4o • If the nature
of the data Is such that there are occasional outliers (perhaps
because of contajpjnat Ion of samples during handling and processing),
Lucas and Crosler (1982b) suggest the use of two—in--a—row rule.
That is, require values in two successive samplings above the upper
Shewhart control limit before declaring an out-of--control situation
based on the Shewhart control limit being exceeded.
For measurements that are typically near or below the instrument
minimum detection limits, it may be possible In some cases to treat
the measurements, or some transformation of the measurements, as
Poisson count data (Ingameus and Switzer, 1973). Lucas (1985b) has
discussed how the CUSUM quality control approach can be employed
with Poisson Count data.
MULTIVARIATE QUALITY CONTROL SCHEMES
In monitoring ground water at a HWS, the concentrations of a several
GQP may be measured at each of several monitored wells. If a
quality control chart Is kept for each parameter at each well, then
one has the problem that while the chance of a false alarm Is kept
small for each chart, the overall probability of a false alarm
becomes large (i.e., the in—control ARL for the whole set of charts
may be quite small in spite of large in—control ARL’s for each
individual chart). This problem is addressed In survey papers by
Jackson (1985) and Alt (1985). One approach is to use a control
scheme that in effect performs a Hotelling’s T 2 test on each
sample. This is essentially a two—sided procedure in that
abnormally low concentrations or unusual combinations of moderate
1-122
-------
concentrations as well as abnormally large concentrations of some
pollutants may trigger the alarm. The statistic calculated at each
time of sampling is
T 2 = (X — uo)T l (X_ u 0 )
where X is the vector of (transformed) parameter measurements at the
monitored wells, u 0 is the steady state mean vector and is the
(positive definite) covariance matrix for the vector of the random
variables corresponding to X. Under assumptions of multivarlate
normality for the measurement vector, T 2 will have a chi—square
distribution with degrees of freedom equal to the dimension of the X
vector when the system is in steady state. Hence the control chart
is similar to the Shewbart chart in that each observed T 2 —v-alue is
plotted against the number of its sampling period, and the system is
declared out of control T 2 exceeds an upper control limit which is
the upper 100 —percent point of the appropriate chi—square
distribution. Jackson and Mudholkar (1979) have suggested
additional statistics that might be monitored to determine whether
outliers or changes in have occurred. Montgomery and Klatt (1972)
have determined 2 optimum sample size and interval between sampling
times for the T control procedure.
A procedure, call MCUSUM for multivarlate CUSUM, proposed by Woodall
and Ncube (1985) advocates running individual CUSUN charts on the
different GQP or on principal components of the GQP values and
choosing h and k so that the combined ARL under in—control
Conditions will be acceptable. Unfortunately, when running CUSUM’s
on measurements of individual GQP where measurements on different
parameters or on the same parameter from different wells are
correlated, the calculation of appropriate values for h and k Is
extremely difficult. If principal components are used, computation
of appropriate h and k values is easier to perform, but the control
charts for principal components win have to be two sided in most
cases, and when action levels are exceeded, It is difficult to
Interpret the cause, and it may not be due to excessively high
concentrations of a m nitored parameter. (This is the same problem
encountered with the T procedure.)
A DECISION PROCEDURE USING A CONTROL CHART SCHEME
A decision procedure for ground-water monitoring must be adaptive
and pragmatic. It must be in accord with available knowledge about
the aquifer being sampled. At the beginning of sampling when little
Information about the system Is available, simple procedures such as
outlier tests should be employed. As more information about the
mean and variance of measurements taken at the various wells becomes
available, changes to more powerful procedures should be possible.
1—123
-------
At the end of a year of sampling (at least once per quarter at each
up— and down-gradient well), the temporal variance for each GQP may
be estimated by using the analysis of variance procedure discussed
earlier. With this variance estimate and the sample mean for the
GQP at a well, one can initiate a combined Shewhart—CUSIJM scheme
with possible inclusion of the robust two—in—a—row rule of Lucas and
Crosier (1982b) that was mentioned above. At this stage It would be
well to choose a larger than normal CUSUM threshold value h, because
of the lack of precision in the estimation u and o . At the end of
each year, provided a basic change in the system (I.e., the aquifer)
has not occurred, the year’s data should be combined with the data
from previous years to obtain improved estimates u and o . The
quality control scheme should then be updated by using these new
estimates. If experience indicates a fairly steady system and that
the estimates of u and a. are reasonably accurate, the value of the
threshold h in the CUSUM may be reduced to increase the power of the
scheme in detecting changes in u that may be due to leakage from the
HI’S.
When the control scheme for a GQP at a down—gradient well indicates
an out-of—control situation, the pattern of change in measurements
of the GQP at this well should be compared to the patterns observed
at the up—gradient wells. If a similar pattern was observed earlier
or is currently at one of more up-gradient wells, it would Indicate
that the out-of—control situation is being caused by something other
than the HWS and that increased monitoring activity would not be
required. The Importance of monitoring up-gradient wells and
keeping corresponding control charts on them cannot be
overemphasized. They are the principal means of determining when
out—of —control situations down-gradient are caused by of f—HWS
activities. In addition, when similar patterns cannot be found
up—gradient, they give greater confidence to a call for additional
monitoring action in down-gradient out—of—control situations.
The combined Shewhart—ajsuM procedure was selected here because
different types of leakage (e.g., small or massive) are possible at
a HI’S. The robust two—in—a—row rule should be considered because
gross errors In GQP measurements are a problem which is usually due
to Contamination of some stage of the measurement system. It would
be useful in situations where the slow flow of water in the aquifer
allows the wait for a decision till the next sampling period without
significantly Increasing health risks or remedial costs if a leak
has occurred. However, in most cases, it will be necessary to
follow the current procedure of checking for measurement error by
taking a second sample at the well as soon as possible after an
unusually high sample measurement is obtained. Multivariate control
schemes do not appear appropriate to ground-water monitoring because
of the amount of Information required to accurately estimate the
COyarjance matrix.
1-124
-------
PROBLEMS ASSOCIATED WITH THE QUALITY CONTROL APPROACH
There are some difficulties to be overcome in the application of
quality control methods to the monitoring of HWS. Before a quality
control scheme can be instituted, the system must be sampled in a
steady state situation over a long enough period to obtain good
estimates of its first and second moments. Intervention by humans
and nature may make it difficult to study a system in steady state
condition for a long enough period to establish a control scheme.
The control schemes are based on the assumption that measurements
are approximately normal in distribution. In Industrial quality
control, several samples are taken and their measurements are
averaged so as to obtain a measurement for the sampling period that
is approximately normal and has reasonable small variance. The high
cost of sample analysis and possible redundancy of measurements of
successive ground—water samples make it difficult to employ this
procedure here. Data transformations may provide near normality of
the data, but they cannot help in considerations of power of the
text implicit in the quality control scheme.
Cyclic trends in GQP values are hard to detect, measure, and
remove. However, failure to remove such trend can have serious
effects on the quality control scheme.
SUMMARY
It is essential that a realistic workable model for the measurements
be formulated and used both in construction and evaluation of
decision procedures. Tests based on unrealistic models will not
succeed in providing an answer to the ground—water monitoring
decision problem no matter how simple and elegant the test may be.
The formulation of good decision procedures (i.e., procedures that
give few false positives and few false negatives and that are based
on an affordable data base) for determining when increased
monitoring activity is needed at HWS is extremely difficult because
of the high cost, low precision, and multivariate nature of
ground-water monitoring data along with system instability which is
due to intrusions of the aquifer caused by man outside the HWS.
The application of Industrial quality control schemes to each type
of water quality measurement from each well or to a vector of the
measure ntg from each well or from all monitored wells was
discussed. The advantages of the quality control approach are that
It provides a method for looking at what is happening to water
quality over time and that it gives users a better feel for the
nature of the data that they have obtained than a simple test of
means does. A sequence of decision rules starting with outlier
tests and ending with combined Shewhart CLJSUM control schemes is
suggested as a reasonable approach to monitoring ground-water at a
1—125
-------
HWS. Unfortunately, quality control schemes do not solve all the
problems involved in determining a statistical decision rule for
when an alarm should be sounded. Problems Involving cost and
precision of measurements will ultimately have to be solved by
improvements in measurement technology.
NOTICE
Although research described in this article has been supported by
the United States Environmental Protection Agency, under Cooperative
Agreement CR 812189—01, it has not been subjected to Agency review,
and therefore does not necessarily reflect the views of the Agency,
and no official, endorsement should be inferred.
REF .ENCZS
Aitchison, J. On the Distribution of a Positive Random Variable
Having a Discrete Probability Mass at the Origin. J. American
Statistical Assoc., 50:901—908, 1955.
Alt, F. B. Nultivarlate Control aiarta. In: Encyclopedia of
Statistical Sciences, Vol. 6, S. Kotz and N.L. Johnson, eds.
John Wiley & Sons, New York, 1985. pp. 110—122.
Anderson, T.W. An Introduction to Multivariate Statistical
Analysis (2nd Ed.). John Wiley & Sons, New York, 1984. 675 pp.
Bisaell, A.F. The Performance of Control (barts and Cusums Under
Linear Trend. Applied Statistics, 33(2) :145—151, 1984.
Freeze, R.A., and J.A. (berry. Groundwater. Prentice—Ball, Inc.,
Englewood Cliffs, NJ, 1979. 602 pp.
Outtaan, I., S.S. Wilks, and J.S. Hunter. Introductory Engineering
Statistics (3rd Ed.). John Wiley & Sons, New York, 1982. 580
pp.
Ingamells, C.O., and p. Switzer. A Proposed Sampling Constant for
Use in Ceochemical Analysis. Talanta, 20(6): 547—568, 1973.
Jackson, J.E. Multivariate Quality Control. Communications in
Statistica — Theory & Methods, l4(11):2657—2688, 1985.
Jackson, J.E., and G.S. Mudholkar. Control Procedures for
Residuals Associated with Principal Component Analysis.
Technometrica, 21(3): 341—349, 1979.
Lorenzen, T.J.., and L.C. Vance. The Economic Design of Control
(harts: A Unified Approach. Technometrics, 28(1):3—1O, 1986.
1—126
-------
Lucas, J.M. Combined Shewhart—CIJSUN Quality Control Schemes. J.
Quality Technology, 14:51—59, 1982.
Lucas, J.M. cumulative Sum (CUSUM) Control Schemes.
Communications in Statistics — Theory & Methods,
14(11) :2689—2704, 1985a.
Lucas, J.M. Counted Data CUSUM’s. Technoinetrics, 27(2):129—144,
1985b.
Lucas, J.M., and R.B. Crosier. Fast Initial Response for CUSUM
Quality Control Schemes. Technometrics, 24(3):129—l44, 1982b.
Lucas, J.M., and R.B. Crosier. Robust CUSUM: A Robustness Study
for CUSUM Quality Control Schemes. Communications in Statistics
— Theory & Methods, 1l(23):2669—2687, 1982b.
Miller, R.G. Simultaneous Statistical Inference. McGraw—Hill,
Inc., New York, 1966. 272 pp.
Montgomery, D.C., and P.J. Klatt. Economic Design of T 2 Control
cuarts to Maintain current Control of a Process. Management
Science, l9(1):76—89, 1972.
Owen, W.J., and T.A. DeRouen. Estimation of the Mean for Log
Normal Data Containing Zeros and Left—censored Values, With
Applications to the Measurement of Worker Exposure to Air
Contaminants. Biometrics, 36(4):707—7l9, 1980.
Page, E.S. Continuous Inspection Schemes. Biometrika,
41(l):l0O—l14, 1954.
Shewhart, W.A. Economic Control of Quality of Manufactured
Product. Van Nostrand, New York, 1931.
Vaughan, W.J., and C.S. Russell. Monitoring Point Sources of
Pollution: Answers and More Questions from Statistical Quality
Control. American Statistician, 37(4, Part 2):476—487, 1983.
Woodall, W.H., and N.M. Ncube. Multivariate CUSUN Quality—Control
Procedures. Technometrics, 27(3):285—291, 1985.
1—127
-------
THE USE OF GEOSTATISTICS FOR CONTOUR MAPS
Evan J. Englund, George T. Flatman, Exposure Assessment Research
Division, Environmental Monitoring Systems Laboratory, Las Vegas,
United States Environmental Protection Agency, Las Vegas, Nevada
ABSTRACT
Monitoring data are spatial variables, called regional variable in
geostatistical text books. This means they have locations as well as
amount of pollutant, and they are correlated in space. Structural
correlation means that samples that have locations close together
usually have similar amounts of the pollutant, and samples with
locations farther apart have less similar amounts of the pollutant.
Experience in monitoring confirms that monitoring data have these
characteristics and need spatial analysis or geostatistics.
Spatial variables are not adequately described by a mean and a standard
deviation, but rather they must be described by a control map of means
and a contour map of interpolation errors or kriging errors. This
poster session will show contouring of actual pollution plumes with
discussion of problems such as outliers.
MONITORING ACTIVITIES BASED ON GEOSTATISTICS
The number and location of sampling sites and the frequency of sampling
are essential decisions in all monitoring activities. To improve the
statistical basis for monitoring activities, researchers at EPA’S
Environmental Monitoring Systems Laboratory in Las Vegas are adapting
geostatistical methods and times series analysis to the space and time
problems of monitoring, with particular attention to soil and
groundwater pollution. These approaches will improve the design of
sampling programs and the interpretation of monitoring data.
Pollution plumes generally spread in a contiguous manner. Thus, samples
taken close to one another in space or time should be more alike than
samples taken further apart. Such samples are called “correlated” or
regional variables. The contaminant levels in samples of such closely
associated variable obey different statistical laws than “random”
variables.
Sampling of random variables assumes that each sample value is
unrelated to the proximity in time or space of other sample values.
Thus, the best sampling schemes for this kind of variable are random
sampling designs. The variability of the pollution concentrations at
the site and the uncertainty in the data that is acceptable determine
the sample size. The larger the variation at the site or the smaller
the acceptable uncertainty, the larger the number of samples that must
be taken.
For correlated variables, the best sampling schemes are systematic
approaches, such as sampling on grid, instead of random sampling. The
number of samples required for correlated variables as for random
1-129
-------
varibles is determined by the variation in pollutant levels at the site
and the acceptable uncertainty. But analysis of correlated variables
can use the relationships between samples to reduce the number of
samples needed.
— __
H
Ii ______ ______ I
> Rsngeofcorr&atjon
A B C D E F
Diitanc. b.tw..n pun of umpi. points
Figuae 1. A semi-variogram
S 1I—VARIOGRAM
The semi—variogram is a geostatistical tool based on information about
time or space relationships between sample observations. The
seini—variogram in Figure 1 graphically shows the relationships between
observations in terms of the distance from each sample to each of the
other samples. The x—axis represents the distance between sample
points, and the y—axis represents the “variance” of the difference in
pollutant levels between pairs of samples which are equal distances
apart.
Since both the distance and the difference between two samples of
correlated variables taken at the same point are zero, their point of
intersection on the graph corresponds to the intersection of the x and
y axis. When the distance between two samples increases and their
correlation weakens, the difference in their values also increases,
resulting in a rising curve or semi—variogram. As the difference
between samples becomes sufficiently great, the sample values become
independent of each other. Their difference becomes more nearly
constant, and the curve becomes a horizontal line. The distance along
the x—axis where the semi—variogram is rising represents the “range of
correlation,” the distance within which samples are related. This range
is used for determining the grid design for sampling, for contouring
pollutant concentrations on a map, and for calculating the
uncertainties in contouring.
SAMPLIM STATEGY
In practice an initial semi—variogram can be constructed on the basis
of preliminary sampling undertaken in the area of suspected
contamination. A semi—variogram can be computed from the sample data by
analyzing all possible pairs of the data. For example, in Figure 2
S
1-130
-------
point 1 through 5 might represent soil samples. There are four pairs (1
and 2, 2 and 3, 3 and 4, 4 and 5) of these samples separated by a
distance equal to A. Thus, for the x—axis value of A, the y—axis value
would be the variance of the differences of these four pairs. In
practice, even for a small preliminary sample, the calculations of all
possible pairs can be so long that a computer program must be used If
the correlation among samples is weak——that is, samples close together
or far apart have similar differences——the semi—variograin approaches a
horizontal line. A horizontal semi—variogram implies that the variable
is random and can be represented simply be a mean and a variance. If
the correlation is strong, then the differences between samples taken
close together are smaller than differences between samples taken
further apart, and the semi—variograxn is a rising curve as shown in
Figure 1.
B
A
1 2 3 4 5
. .
C
D
Figure 2. Possible pairs of sampling points.
The spacing of sampling points for intensive sampling can be determined
using the semi—variogram as a guide. A grid designed to cover the area
to be monitored provides a tool for determining sample sites, with
samples taken at each grid intersection. Usually a grid spacing of
about two—thirds the range of correlation insures that the sampling
points in the intensive sampling phase are close enough to each other
to have correlated values. To sample at closer distances would provide
little new information while greater spacing could miss a change in
pollution levels. The results of the intensive sampling can be used to
refine the semi—variogram and confirm the appropriateness of the
sampling distances.
KRIGING
A technique called “kriging” interpolates pollution levels at points
between the sampling sites so that uisoplethsI of pollution levels can
be drawn on a map. The kriging estimate of the pollutant level at any
particular point is the weighted average of the nearest neighboring
sample values. The size of the neighborhood is determined by the range
of correlation. The result in an isornap, a contour map with isopleths
tracing the lines of equal values. Figure 3 presents pollutant levels
for a study area in Southeast Ohio. Isomaps can be used as overlays for
1-131
-------
aerial iotographs or maps of the study area. The three—dimensional map
show in Figure 4 is another way of presenting data shown in Figure 3
arKi clearly distinguishes the same high and low pollutant levels.
Kriging also can coi çute the standard errors of estimation for the
estimated samples values using the range of correlation. These error
estimates can also be mapped. Figure 5 contours the standard errors of
estimates or the concentration levels shown in Figures 3 and 4.
By using these three types of geostatistical outputs, a decision maker
can identify areas requiring cleanup, ii re sampling, or no action. A
cleanup area would be delineated by pollution averages which are above
a chosen action level and standard errors which are below a chosen
acceptable level. An area for nore investigation would have high values
for both pollution averages and standard errors, and an area for no
action would have low values for both.
To date, the Laboratory’s efforts have been directed toward applying
geostatistical techniques to two-dimensional problems, such as
contamination of surface soil. Future efforts will investigate the
feasibility of using similar techniques in addressing three—dimensional
problems, such as those encountered in groundwater contamination.
? C ICE
Although the research described in this article has been supported by
the United States Environmental Protection Agency, it has not been
subjected to Agency review, and therefore does not necessarily reflect
the views of the Agency and no official endorsement should be inferred.
1—132
-------
Figure 3. Kriging of benzo-a-pyrene concentration in soil.
Lv&s in ppb
Sc1s I I I 1 1 1 L - Low concentration
5 km H - High concentration
• - Sampling pointi
Figure 4. Three dimensional representation of data used for Figure 3.
1—133
-------
Sc&.lit I I I
km
Figure 5. Standard errors of estimates for benzo.a-pyrene concentration in soil.
By using these three types of geostatistical outputs, a decision maker
can identify areas requiring cleanup, ziore sanpling, or no action. A
cleanup area would be delineated by pollution averages which are above
a chosen action level and standard errors which are below a chosen
acceptable level. n area for n re investigation would have high
values for both pollution averages and standard errors, and an area for
no action would have low values for both.
To date, the Laboratory’s efforts have been directed toward applying
geostatistical techniques to two—dimensional problems, such as
contamination of surface soil. Future efforts will investigate the
feasibility of using similar techniques in addressing three—dimensional
problems, such as those encountered in groundwater contamination.
1-134
-------
BIOLOGICAL TEST
METHODS
thai rpe rsons
Reva Rubenstein Liewellyn Williams
thief Deputy Director
Health Mses nt Section Quality Assurance and
Office of Solid Waste Methods Research
U.S. EPA Environmental Research
401 N Street, S.W. Laboratory
Washington, D.C. 20460 U.S. EPA
P.O. Box 15027
Las Vegas, NV 89114
-------
UTILITY OF IMMUNOAS SAY FOR TRACE ANALYSIS
OF ENVIRONMENTAL CONTAMINANTS
Jeanette M. Van Einon, Research Molecular Biologist, EMSL, U.S.
Environmental Protection Agency, Las Vegas, Nevada
ABSTRACT
Immunochernical technology is being rapidly applied to solve problems
in environmental chemistry. Immunoassays such as the enzyme—linked
immunosorbent assay (ELISA) are sensitive bioanalytical procedures
applicable to many classes of compounds. Immunoassay development
must be guided by the basic principles of analytical chemistry.
Sample collection and preparation, and data acquisition and handling
all need to be optimized.
INTRODUCTION
Improvements in selectivity, sensitivity, speed, and simplicity are
important goals in analytical chemistry (Milby and Zare, 1984). The
evolution and application of new analytical methods will help meet
these requirements. Since their introduction, immunoassays have
become widely used for the analysis of proteins, hormones, and
drugs, and are rapidly being extended to pesticides, and
environmental contaminants (Table I). Immunoassays utilize specific
antibodies and a detectable label on either antibody or antigen.
The high selectivity of immunoassays is assured by the inherent
selectivity of immunological reactions, their sensitivity is
determined by antibody affinity and by the detection limit of the
label.
Immunochernical methods of analysis offer advantages of sensitivity,
specificity, and speed of analysis (Hammock and Mumma, 1980).
Compounds which are most difficult to analyze by classical
procedures are frequently amenable to analysis by immunochemistry.
Immunoassays are especially applicable to compounds that: 1) must
undergo extensive derivatization for gas chromatography or
high—pressure liquid chromatography; 2) are water soluble; 3)
require a multi—step sample preparation, and 4) are chemically
complex. In several of these instances, immunoassays not only
complement other analytical techniques but may be the most practical
method. Products of biotechnology such as genetically engineered
microorganisms, plant gene manipulations, and fermentation products
are particularly amenable to immunochemical detection.
ASSAY FORMATS
Several different immunoassay formats are currently in use. As the
technology becomes more commercialized many other assay permutations
can be expected. Some of these new assay designs will be developed
2-1
-------
to solve a particular need, while others will be introduced as
slight modifications to existing procedures. Whatever the assay
design, each is dependent on the highly specific antigen—antibody
reaction. Although antibodies are biologically derived reagents,
limaunoassays must not be confused with bioassays. These two
techniques are fundamentally distinct and are based on different
principles.
ANTIBODIES
Polyclonal antibodies are adequate for most immunoassays, however,
monoclonal antibodies could be developed against pesticide and
environmental contaminants yielding an analytical reagent that is
physically, chemically, and immunologically homogeneous. The large
supply of monoclonal antibodies could facilitate the standardi-
zation of Immuaoassays as several laboratories would be using the
same antibody clone. Although a very small amount of antibody Is
needed per assay, having a large supply of monoclonals could
alleviate fears of eventually exhausting the supply. Yet, it is
important to consider that, in some cases, polyclonals will be
superior to monoclonal antibodies (Hammock and Numma, 1980).
However, as lilmunoassay moves into the private sector, there will be
compelling administrative and legal pressures to employ monoclonal
antibodies (Van Emon et. al., 1985).
ADVANTAGES OF IMMUNOAS SAYS
For a chromatographic analysis, unless a large number of expensive
instruments are employed, usually just one analysis can be done at a
time by one person (Chait and Ebersole, 1981). With iimnunoassays,
many analyses can be performed at once by a single Individual
without elaborate apparatus, and the results can be monitored
visually in some cases, eliminating an instrumental detector.
I *unochemica1 methods, traditionally unfamiliar to the residue
chemist, offer exciting possibilities for newer cost—effective
approaches for residue analysis (Hemingway, 1984). Affinity columns
loaded with antibody could selectively remove and clean—up an
analyte from solution before analysis by a conventional method.
Imaunoassays could also be used as selective detectors for
cbroaatographic separation techniques.
The analytical use of antibodies enables substances to be
quantitated at low levels in crude extracts. The costly and time—
consuming steps of extraction and isolation required of many other
analytical methods can often be avoided. This elimination of steps
reduces the amount of error in the procedure. Due to the minimal
sample preparation and speed of analysis, Immunoassays are excellent
screening methods for the more expensive analytical techniques such
as gas chromatography/mass spectrometry.
2—2
-------
Since the assay is performed in an aqueous media, slightly different
methods of extraction and sample preparation should be expected.
Such changes should result in an overall simpler procedure.
However, extraction and procedural efficiencies must be determined
for each assay.
Antibodies which will detect parent compound plus toxic metabolites
could be used in combination with one or more highly specific
antibodies to guantitate several compounds of interest. If the
antibody is of moderate specificity, one can develop a general
screening assay for the presence of a class of compounds or the
presence of a specific functionality. However, if one needed to
analyze several different compounds simultaneously in one matrix,
immunoassay may not be the method of choice, due to the large amount
of controls and standards needed. Immunoassays could be
successfully used for the rapid screening of a large number of
samples for the presence of specific types of compounds and for
confirmatory tests (Ercegovlch, 1971).
Immunoassays can be optimized for speed, portability, or
sensitivity. Highly sensitive assays can be developed for research
purposes, but by using the same reagents one can develop an assay
that will give a quick qualitative answer in real time. Such rapid
assays can be performed on site. Immunoassays can be adapted as
simple field screening procedures, since reagents are stable, have
long shelf—lives, and present no health hazards (Anonymous, 1979;
Voller et al., 1976). Assays that are optimized for speed will be
done so at the expense of sensitivity. However, the sensitivity of
the antigen—antibody reaction is so great that a small loss in
detection limit will still afford a usable assay. Qualitative
immunoassays could be used to determine sources of contamination,
direct the collection of samples, and monitor cleanup operations.
CONCLUSIONS
An immunoassay consists of the specificity of an immunological
reaction with a sensitive indicator system. Improvements in all the
components of immunoassay are certain, and will yield improved
assays. These improvements include new enzyme labels; sensitive
methods of enzyme determination such as cofactor cycling; and better
cross—linking reagents, especially of the hetero—bifunctional type
(Wisdom, 1976). Many modifications can be expected In the delivery
systems for these assays. For instance, one can envision ELISA
systems coupled to photon counters or luminometers if enzymes such
as luciferin were used or to silicon chip if redox enzymes are
used. Such assays could provide a dynamic output of concentration
of analyte in real time.
The field of immunoassay is very broad with new techniques
constantly being introduced. The various types of Immunoassays are
2-3
-------
complimentary and there is no one ideal label for use in all
immunoassays. The choice of assay format and label should be based
on the analyte to be measured and the purpose for which the
measurement will be made. There is a wide variety of circumstances
and environments in which Immunoassay and other similar ligand
binding assay techniques could be employed. Although there appears
to be an overindulgence in pursuing novel methods, alternative
methodologies based on the antigen—antibody reaction will
continually emerge, many designed for a particular circumstance.
Table 1
I14MIJNOAS SAYS WITH ENVIRONMENTAL APPLICATIONS
Compound References
1. Aidrin and Dieldrin Langone and Van Vunakis, 1975
2. Bacillus thuringiensis Wie et al., 1982
3. Benomyl Lukens et al., 1977
4. Chiorsulfuron Kelley et al., 1985
5. 2,4—D and 2,4,5—T Rinder and Fleeker, 1981
6. Paraoxon Hunter and Lenz, 1982
7. Paraquat Van Emon et al., 1986
8. Parathion Ercegovich et al., 1981
9. S—Bioallethrin Wing and Hammock, 1979
REFRRRNCES
Anonymous, Lancet . 780—781 (1979).
Chait, E.M., and Ebersole, R.C., Anal. Chem . 53, 682A—692A (1981).
Ercegovich, C.D. Pesticide Identification at the Residue Level ,
Gould, R..F. ed., p. 162—177, American Chemical Society Pubi.,
Washington, D.C. (1971).
Ercegovich, C.D., Vallejo, R.P., Gettig, R.R., Woods, L., Bogus,
E.R., and Muama, R.0., J. Agric. Food Chem. , 29, p. 559-563
(1981).
Hammock, B.D., and Mumma, R.O. Pesticide Analytical Methodology ,
Harvey, J., and Zweig, G., eds., p. 321—352, AmerIcan Chemical
Society, Washington, D.C. (1980).
Hemingway, R.J., Aharonson, N., Greve, P.A., Roberts, T.R., and
Thier, H.P., Pure & Appi. Chem . 56, 1131—1152 (1984).
2-4
-------
Hunter, K.W., and Lenz, D.E., Life Sd . 30, 335—361 (1982).
Kelley, M.M., Zahnow, E.W., Petersen, W.C., and Toy, S.T., J.
Agric. Food Chem . 33, 962—965 (1985). —
Langone, J.J., and Van Vunakis, H. Res. Commun. Chem. Pathol.
Pharmocol . 10, 163—171 (1975).
Lukens, H.R., Williams, C.B., Levison, S.A., Dandliker, W.B.,
Murayaina, D.., and Baron, R.L., Environ. Sd. Tech . 11, 292—297
(1977).
Milby, K.H., and Zare, R.N., Amer. Cliii. Prod. Rev . Jan, 12—19
(1984).
Kinder, D.F., and Fleeker, J.R., Bull. Environm. Contain. Toxicol .
26, 375—380 (1981).
Van Einon, J., Hammock, B. and Seiber, J.N., Anal. Chem . 58,
1866—1873 (1986).
Van Emon, J.M., Seiber, J.N., and Hammock, B.D., Bioregulators for
Pest Control , Hedin, P.A., ed., p. 307—316, AmerIcan Chemical
Society, Washington, D.C. (1985).
Voller, A., Bidwell, D.E., and Bartlett, A., Bull. World Health
Org . 53, 55—65 (1976).
Wie, S.I., Andres, W.R., Hammock, B.D., Faust, R.M., and Bulla,
L.A., Appl. and Environ. MIcrobiol . 43, 891-894 (1982).
Wing, K.D., and Hammock, B.D., Experientia , 35, 1619—1620 (1979).
Wisdom, G.B., Cliii. Chem . 22, 1243—1255 (1976).
2-5
-------
DERIVATION AND USE OF MONOCLONAL ANTIBODIES
FOR ENVIRONI 1ENT L MONITORING
Alexander Karu, Head, Hybridoma Center, College of Natural Resources,
University of California, Berkeley, CA
ABSTRACT
Over the past 12 years, monoclonal antibody technology has proven
valuable in numerous biomedical applications. This talk will
summarize unique advantages and difficulties of extending this
technology to environmental toxicology, chemical quality assurance,
and risk assessment, where it has evolved more slowly.
Most analytes of environmental interest are of low molecular weight,
somewhat nonpolar, sparingly soluble in aqueous media, and at low
concentration in the test sample. Monoclonal antibodies (MAbs) for
detecting these substances must have the highest possible affinities.
The analytes (haptens) are made immunogenic by covalent coupling to
“carrier” molecules. Conjugates must be designed with regard to
sterism, linkage chemistry, linker length, hapten density, and choice
of carrier. Immunization and screening procedures should make use of
different conjugates. Immunization protocols should promote T—helper
cell responses and minimize T—suppressor cell responses to the hapten.
Different mouse strains should be inanunized, as they can differ
significantly in their antibody responses. Because any set of MAbs
will have a broad spectrum of affinities and specificities, it is
highly desirable to clone hybridomas during initial selection, and to
screen a great many. Robotic sampling systems and computer—assisted
screening are essential for this purpose. Newer techniques such as
splenocyte transplantation and antigen—focused cell fusion may improve
the yield of useful hybrodomas, but these methods have not yet been
widely tested.
Enzyme and radioimmune assays are the most common analytical methods
in current use, but several others, ranging from immunoelectrodes and
solid—state immunosensors to disposable cards and “dipsticks,” are
rapidly being developed. MAbs are also potentially useful agents for
extracting and concentrating the substances of interest from complex
mixtures or from solid surfaces. Problems of analyte concentration
and elimination of interfering substances are different for
inimunoassays than for standard analytical procedures. Techniques to
allow immunoassay of substances that are nearly insoluble or
chemically unstable in aqueous solvents remain to be developed.
Immunoassays are ideal for rapid screening of large numbers of
samples, e.g., for on—site monitoring of the degradation and dispersal
of toxic materials. For the near future, most applications will
involve monitoring of air, water, and soil pollution and residue
levels in foods. Other potential uses include quality analysis in
chemical manufacturing, determination of worker exposure and safe
field re—entry intervals, and tracking of organisms and recombinant
gene products used in biological and other conventional analytical
methods.
2 —7
-------
NAb—based environmental monitoring presents new regulatory problems.
Agencies must develop criteria for standardizing diverse
iinmunochemical assays and relating them to instrumental analysis.
Policy makers will have to factor immunoassay results into their
decision—making processes. Inununoassays for adducts or metabolites of
genotoxic substances have great potential for dosimetry and risk
assessment, but their implementation will raise legal and ethical
issues.
2-8
-------
IMNUNOAS SAY FOR THE DETERMINATION OF PENTACHLOROPHENOL
P ND RELATED COMPOUNDS IN WATER SAMPLES
T.C. Chiang, L.R. Williams, S.D. Soileau, R.F. Schuxnan, and K.W.
Hunter, Jr., Lockheed Engineering and Management Services Co., Inc.,
Las Vegas, Nevada
ABST1 ACT
The use of immunochemical methods has been wide spread in clinical
laboratories, and the techniques have also been applied to the
determination of a variety of low molecular weight compounds such as
drug residues, mycotoxins, pesticides and plant hormones.
Immunochemical analytical methods are generally based on the principle
of competition between analyte and labeled form of the analyte for a
specific receptor. In a cooperative effort between Westinghouse
Bio— nalytical Systems, EPA, and Lockheed—EMSCO, a new monoclonal
antibody—based immunochemical analytical procedure for the
determination of pentachlorophenol (PC?) levels in water has been
evaluated.
The method is a competitive inhibition enzyme immunoassay using
anti—PCP antibodies prepared by the monoclonal antibody technique.
The free PC? in solution binds to the anti—PCP antibodies and inhibits
their subsequent binding to the solid—phase adsorbed PCP protein
conjugate. After an incubation, the anti—PC? antibodies not bound to
the surface of the microtiter wells are removed by washing. The
number of antibodies bound to the solid—phase is determined by adding
an enzyme—labeled second antibody with binding specificity for the
anti—PC? antibody. After an incubation and final washing, the
substrate of the enzyme is added and the amount of anti—PC? antibodies
bound to the solid—phase is directly proportional to the intensity of
color produced by the enzymatic reaction. The intensity of this color
is therefore inversely proportional to the concentration of PC? in the
sample. Quantitation is achieved by comparing the mean optical
density obtained with the sample to a standard curve generated from
known concentrations of PCP on the same plate.
Thirty—six water samples from various sources (tap, ground and surface
water) were spiked with different levels of PC? and analyzed by two
independent laboratories. The samples were also analyzed using an EPA
approved GC/NS method. In addition, aliquots of the sample extracts
prepared for GC/MS analysis were also analyzed by the immunoassay.
For water samples that were analyzed directly, the method detection
limit was 88.8 ppb (parts per billion). The method also provides for
the use of an internal standard addition to determine possible
interferences. During the course of this study several hardware
problems associated with this type of assay were encountered (mainly,
the plate effect and the reader effect). These unexpected effects
have resulted in considerable delay in finishing this study.
Nevertheless, the results indicate that this irnmunoassay can be used
to detect PC? in selected water samples, and that once the hardware
problems are resolved, the precision, accuracy and the speed of this
procedure will be much improved. This study is the first of several
2-9
-------
planned to evaluate methods based upon inuuunochemical techniques that
offer potential for low—cost, sensitive and selective detection of
target chemicals of Agency interest. The study was developed in
coordination with the OSW staffs to demonstrate the equivalency of
proposed alternatives to Agency—accepted methods.
2-40
-------
ULTRA SENSITIVE BIOASSAY FOR DIOXIN
Richard F. Schuman and Kenneth W. Hunter, Westinghouse Blo—Analytic
Systems Co., Rockville, I4aryland
ABSTRACT
2,3,7,8—tetrachlorodibenzo--p—dioxin (TCDD) has been shown to induce
ethoxyresorufln—O—deethylase (EROD) activity or other biological
responses in a number of cell cultures. Using a new cell line, WBAS
86—163B, an ultra sensitive bioassay for the detection of TCDD and
related compounds has been devised. The target cells can detect
TCDD concentrations as low as 2 X 10—1314, which is at least 1.5 logs
more sensitive than other bioassay systems. With the WBAS cell
line, an induction response can be obtained after treatment of the
target cells for only 18—24 hours, four to 10 days less than
required by other bioassays. These cells have been used
successfully to detect TCDD in soil and water samples.
Several dibenzo—p—dloxins, chlorinated phenols and polychiorinated
biphenyls were tested for their abilities to induce EROD in the
target cells. TCDD was at least several thousand fold more potent
in elliciting a response than were the other chemicals. TCDD was 8
times more active as an EROD inducer than the next most effective
chemical, 2,3, 7,8—tetrachlorodibenzo—furan.
INTRODUCTION
2,3,7,8—tetrachlorodlbenzo—p—dioxln (TCDD) contamination in the
environment has become a major public health issue. TCDD and
related compounds such as polycb.lorinated dibenzofurans are produced
as by—products of the manufacture of industrial organic compounds
(eg. 2,4,5—T) and also as the result of several combustion
procedures (1—3). TCDD is an extremely stable compound which has
been shown to cause liver and kidney damage, chloracne, mutagenesis
of bacterial and mammalian cells as well as inimunosuppression In
mice (1,4—7). Thus, rapid and reproducible assays are needed to
meet regulatory requirements for the detection of TCDD.
TCDD analysis is conventionally done by gas chromatography/mass
spectrometry (GC/NS) analysis. To detect low concentrations of
dioxin usually requires using the mass spectrometer in its most
sensitive mode. However, under these conditions, other compounds of
interest (eg. chlorinated dibenzofurans) may not be detected. An
alternative method for detection of TCDD and related compounds is
the use of an assay based upon the biological response of cells to
the chemical.
Several bioassays from the detection of TCDD have been described,
including In vitro systems based upon Inhibition of cell division,
2-11
-------
altered cellular morphology and enhanced differentiation of specific
cell types, as veil as Induction of enzymes In vivo and In vitro (2,
8-10). In this paper we describe a detection method for TCDD which
is quicker and more sensitive than other bioassaya. The assay has a
dynamic range of 0.060 pg/mi to 1.0 pg/mi and has been used to
detect TCDD in water samples, sediment from water samples, and
benzene extracts of soil samples.
METhODS
Target Cells : WBAS cell culture 86—163B (patent pending) was used
in all experiments. The parent cell line was derived from mouse
liver cells. After mutagenesis and cloning, cells were Isolated
based upon the Inducibility of the enzyme ethoxyresorufin—0—
deethylase (EROD) by TCDD. The cells were cultured in William
Medium E (WME) with 10% fetal bovine serum (FBS), 2 mM L—glutamine
and 10 mM Hepes buffer (p11 7.4).
Chemicals and Sample Preparation : TCDD was obtained from the
National Bureau of Standards (Galthersburg, MD) or Cambridge Isotope
Laboratories (CIL; Woburn, MA). Other dlbenzo—p—dioxlns,
chlorinated phenols and polychiorinated biphenyls were purchased
from Chem Services (West Chester, PA) or were synthesized at WBAS.
2,3,7,8—tetracblorodibenzofuran (TCBDF) was obtained from GIL.
For routine use in the assay, benzene extracts of TCDD—containing
soils are reduced to near dryness under a flow of nitrogen gas. The
samples are then brought to volume (1 ml) In dimethylsuif oxide
(DMSO). Sediment Is removed from water samples by centrifugatlon,
then dried, weighed and resuspended in 250 ul of DMSO. The
sediment—free water is then passed over a C18 reverse—phase column.
The resulting water fraction (“stripped” water) Is retained for
testing. Organic material on the column is eluted with acetone and
taken to near dryness as described above and brought to volume In
DMSO.
Assay : WBAS 86—163B cells are planted into 35 mm tissue culture
dishes 20—24 hours before treatment. The medium is then removed
from the cells and they are refed with 3 ml WME as described above,
except that the FBS concentration is reduced to 1% (1% WME). For
comparison of Induction of EROD, the cells are treated with stock
solutions of dibenzo—p-dioxin and 12 selected chlorinated compounds
(Table i) prepared in DMSO. Duplicate wells are treated with 10 ul
of sample and an additional 10 ul of DMSO, resulting In final DMSO
concentration of 0.67%. A standard curve Is generated by treating
duplicate wells with appropriate concentrations of TCDD to yield
final doses of 1, 0.5, 0.25, 0.125, 0.0625, 0.03125 and 0.0 pg/mi,
all in a final concentration of 0.67% DNSO.
2-12
-------
For benzene extracts which have been exchanged into DMSO, duplicate
wells of target cells are treated as described above. An additional
set of duplicate wells also treated with 10 ul of each sample and 10
ul of TCDD spiking solution (150 pg/mi). The final concentration of
the TCDD spike in these wells is 0.5 pg/mi. If further dilutions of
a sample are required, they are made in DMSO, so that the solvent
concentration remains constant.
Stripped water samples are initially mixed with 5X WME at a ratio of
4 parts sample: 1 part 5X WHE and this mixture is used to treat the
cells. However, If significant toxicity is Induced by the samples
under these conditions then they are tested at lower concentrations,
usually at dilutions of 1:10 through 1:50.
Approximately 20 hours after treatment the cells are analyzed for
EROD activity induced by the TCDD standards or by Inducer8 in the
samples. The general details of the enzyme assay have been
described elsewhere (14).
RESULTS
Comparison of Inducers : Thirteen compounds have been tested for
their abilities to Induce EROD In target cells (Table 1). The
concentration of each chemical needed to induce an EROD response is
compared to the amount of TCDD needed to induce a similar response
(Table 1). For example, 20,000 pg/mi of 2,4,5—trlchiorophenol
(2,4,5—TCP) induces a response equivalent to that induced by 0.063
pg/mi of TCDD. By this method, TCDD is estimated to be 3.2 X lO
more potent (20,000/0.63) than 2,4,5—TCP as an inducer of EROD in
the WBAS target cells. Table 1 shows that none of the dibenzo—p—
dioxins, polychiorinated blphenyls or chlorInated phenols are nearly
as effective as TCDD In provoking a response in the target cells.
Only TCDBF is within a log of TCDD in its ability to induce the
enzyme In the target cells.
Sample Toxicity : Biological response systems for the detection of
chemicals such as TCDD are particularly susceptible to
toxicity—related phenomenon. The toxicity is sometimes overt,
leading to altered morphology in the target cells; more subtle
effects are also noted. For example, a sample which causes no overt
cellular toxicity might induce a calculated TCDD equivalent value of
0.127 at a dilution of 1:20 and a calculated value of 0.164 at
1:50. The higher EROD response at the higher dilution indicates
some residual toxicity associated with the sample that prevents full
expression of the induced enzyme. In such cases the calculated
value for the TCDD spike usually falls below the acceptable limit of
0.375 pg/mi. Further dilutions are required to bring the toxic
substances down to an acceptable level and allows complete induction
of the enzyme.
2-13
-------
TABLE 1
Induction of EROD in WBAS Target Cells by
Chlorinated Compounds
Test Inducing Ecauiva lent Relative
Chemical Dose (pg/mi) TCDD Dose (pg/ml)a TCDD Effectb
Dibenzo—p—dioxin (DD) 40,000 0.20 2.0 X 1O 5
7—Cl—DD 40,000 0.32 1.3 X 10
2,7—dichloro—DD 40,000 0.20 2.0 X 1O
1 2,3,4—tetrachloro—DD 20,000 2.00 1.0 X lO
1,2,3,4,7,8—hexachloro--DD 4,000 1.00 4.0 X
Pentach].orophenol 40,000 0.25 1.6 X i0
2,4,5 trichiorophenol 20,000 0.063 3.2 X 10
2,4 dichiorophenol 20,000 0.063 3.2 X 1O 5
2,2’,3,3, ‘6,6’ Po].y—
chlorinated biphenyl(PCB) 40,000 0.125 3.2 X 1O
2,3,3’,4,4’5 PCB 10,000 2.00 5.0 X
2,2’,4,5,6’ PCB 40,000 0.125 3.2 X 10
2,2’,4,5,5’ PCB 40,000 0.125 3.2 X 1o
2,3,7,8—tetrach].oro— 8 1.00 8
dibenzofuran
a Dose of TCDD required to give EROD response equivalent to that
of the test chemical
b Dose of the test chemical divided by the dose of TCDD required
to give an equivalent EROD response.
2-14
-------
Another problem area, although not necessarily related to toxicity,
is high concentrations of inducers in a sample. Thus, the results
obtained from an unknown sample might be 2.03 pg/mi at a dilution of
1:21600 with a similar value at a dilution of 1:43200. This lack of
dilution effect indicates that such high concentrations of inducing
substances are present that they are beyond the range of analysis.
Performance of the Assay : Figure 1 shows the results of a typical
TCDD bioassay. Each point is the average of two replicates. The r 2
for this set of data is 0.997. In general, r 2 below 0.950 are rare,
but for comparative purposes the assays are normalized with respect
to the luminescence response of the 1.0 pg/mi TCDD control sample,
setting this value to i00 . All other values are determined by
dividing the luminescence at each dose by the luminescence at 1.0.
(e.g., if the luminescence at 1.0 pg/mi is 400 and at 0.25 pg/mi is
130, the relative response at 0.25 pg/mi is 0.325). A summary plot
of seven assays is presented in Figure 2. The r 2 of the data is
0.985; the midpoint of the curve is approximately 0.5 pg/mi
(approximately 1.6 X 10 - 2 M). Inclusion of values for 2.0 p /ml
(which is outside the dynamic range of the system) lowers the r to
approximately 0.900. However, if the 2.0 pg/mi point is included,
4—parameter logistic analysis can be used and an r 2 of 0.990 can be
obtained (not shown). Generally, the lowest concentration of TCDD
used as a standard is 0.0625 pg/mi. However, the inclusion of
0.03125 pg/mi does not adversely affect the r 2 of the standard
curve, as demonstrated in Figure 1. This lack of effect on the r 2
indicates that the 0.03125 pg/mi standard is within the dynamic
range of the curve. However, at present, 0.0625 pg/mi
(approximately 5 X 1O 3 M) Is used as the limit of detection.
Samples falling below that level are retested at higher
concentrations.
DISCUSSION
In this paper we describe a rapid, sensitive and reproducible
bioassay for the quantification of TCDD and other EROD inducers In
environmental samples. The target cells, WBAS 86—163B was selected
for its hyper inducibility by TCDD. Like other cell culture—based
assays for TCDD (2,10,11), related compounds also induce a response
in the target cells. Except for TCDD, TCDBF was the most potent
inducer of EROD In the WBAS cells. Approximately an 8—fold increase
in TCDBF, compared to TCDD, was required to obtain the same
response. These results are similar to those obtained by Glerthy
and Crane (2) using the “flat cell assay” (FCA) and by Safe (10)
using H4IIE cells, both of whom found that approximately 10—fold
more TCDBF than TCDD was required to produce the same response.
However, the sensitivity of the WBAS target cells is greater than
reported for other systems. The midpoint of the dynamic range for
our cells Is approximately 1.6 X l0 2 M, as compared to 1.9 X lO -°M
reported by Safe (10). The limit of detection in the FCA is l0 - 1 -M;
2—15
-------
Figux e 1
I
TCDD Bioassay, W S Cells
Figure 1: A standard curve from a typical TCDD bioassay. Each
point is the average of two replicates. Standards were teste 1 in
two-fold dilutions from 1.0 down to 0.03125 pg/mi. The r of
this line was 0.997.
rEço we
0 02 0.4
iWO E 14*SÜS pg/rf
06
2-16
-------
Figure 2
Cummulative Results, TCDD Bioassay
C
+ —SD
Figure 2: A standard curve generated from the normalized data of
7 TCDD bioassays. Standards were tested in two-fold dilutions
from 1.0 to 0.0625 pg/mi. The r 2 of the line is 0.985.
90
aD
E
J
40
20
10
0
0 0.2 0.4 0.6 0.3 1
C
2-17
-------
with the WBAS cells it is below 2.0 X 10 3 M. Another advantage of
the WBAS bioassay is the rapid turnover rate. The bioassay data
presented here are for 20—24 hour exposures. The H4IIE system used
by Safe includes a 4—5 day incubation of cells with inducer; the FCA
assay requires 14 days.
TCDD and related compounds are currently detected by the use of
CC/MS. The major advantage to CC/MS analysis is its ability to
detect different isomers within a single sample. However, under
conditions required for maximum sensitivity, multiple analyses may
be required to detect similar compounds within a singe sample.
The bioassay, while not able to distinguish between similar isomers,
provides many advantages not available with the GC/MS. In general,
the assay does not require highly skilled personnel or complex
equipment. The assay is inexpensive and rapid, allowing for
multiple testing and the generation of statistical data for
samples. Although isomer specificity is not provided by the
bioassay in a single test, the presence or absence of toxic
chemicals which induce EROD, can be determined and expressed as
TCDD equiva1enta. ’ Since many such compounds are subject to
regulatory cleanup processes, the data obtained from bioassays can
be used to establish priorities in toxic chemical decontamination
programs within industry and government.
REFERENCES
flindsill, R.. D., D. L. Couch and R. S. Spiers. J. Env. Path. Tox.
4:401—425, 1980.
Gierthy, J. F. and D. Crane. Fund. Appi. Tox. 5:754—759, 1985.
Safe, S., et al. airomosphere 14:675—683, 1985.
Girl, A. K. Nut. Res. 168:241—248, 1986.
Zack, J. A. and ft. ft. Suskind. J. Occup. Med. 22:11—22, 1980.
Cook, R. R. Lancet 1:618—619, 1981.
Pitot, H. C., et al. Canc. Res. 40:3616—3620, 1980.
Osborne, ft. and W. F. Greenlee. Tox. Appi. Pharm. 77:434—443, 1985.
Hudson. L. G. et al. Bioch. Biophy. Res. Comm. 115:611—617, 1983.
Safe, S. Ann. Rev. Pharm. Tax. 26:371—399, 1986.
Sawyer, T. and S. Safe. Tax. Let. 13:87—94, 1982.
2-18
-------
STRATEGIES FOR USING BIOASSAY METHODS FOR THE IDENTIFICATION
OF HAZARDOUS COMPONENTS AND COMPARATIVE RISK ASSESSMENT
OF COMPLEX MIXTURES
JoEllen Lewtas, U.S. Environmental Protection Agency, Health Effects
Research Laboratory, Research Triangle Park, North Carolina
ABSTRACT
Two strategies particularly useful for approaching the toxicology of
complex mixtures are: (1) bioassay—directed chemical
characterization and (2) comparative bioassay studies.
Bioassay—directed fractionation and chemical characterization is a
strategy for identifying biologically active compounds or compound
classes in complex mixtures. The identification and assessment of
mutagens and carcinogens in complex mixtures has been significantly
advanced by the use of short—term genetic bloassays.
Bioassay—directed fractionation coupled with new organic
characterization methods has provided the tools needed to more
efficiently identify potential carcinogens in complex mixtures. A
comparative potency strategy for evaluating the relative toxicity,
mutagenicity and carcinogenicity of a series of different mixtures
has been used to provide comparative potency data for risk
assessment. The comparative mutagenicity and carcinogenicity of a
series of combustion emissions has been assessed using dose—response
studies in bacteria, mammalian cells and rodents. This data base
has been used to develop a comparative risk assessment methodology
for combustion emissions which is being extended to the evaluation
of hazardous waste incinerator emissions and other complex mixtures.
INTRODUCTION
Combustion emissions consist of a very complex mixture of thousands
of chemicals resulting from incomplete combustion of either
vegetative sources Ce. g., tobacco, wood, etc.), fossilized fuel (e.
g., oil, coal, gasoline, diesel fuel) or synthetic sources (e.g.,
syn—fuels, other chemical products). This complex mixture generally
contains a high percentage of organic chemicals including aliphatic
hydrocarbons, polycyclic aromatic hydrocarbons (PAils), substituted
PAils and other polar organic compounds. The emission mixture itself
is heterogeneous and consists of very volatile gases (e.g., CO, NOR,
SO 2 , hydrocarbons and aldehydes), semi-volatile material (e.g., 2—3
ring aromatics), condensed organic and inorganic matter adsorbed
onto very small (usually submicron) carbonaceous particles often
called soot.
These very complex mixtures cannot now or in the forseeable future
be completely characterized chemically since they contain thousands
of components. Therefore, toxicological studies on the Individual
components is an unsatisfactory means of assessing the toxicology of
2-19
-------
the total mixture. These mixtures, however, can in most cases be
reproducibly generated by the emission source (e.g., automobile on a
dynainometer) to generate the mixture to which humans are exposed.
Considerable engineering effort has been devoted to standardizing
combustion conditions and in some cases even the fuels, so that
these emissions can be reproducibly generated. This has been done
in order to chemically and physically characterize the emissions,
however, these same techniques may be applied to the generation of a
representative emission source for toxicological studies.
The toxicological data base on combustion emissions varies from
being very extensive and good for some sources such as cigarette
smoke and certain automotive emissions (e.g., diesel emissions) to
very limited for wood combustion and synfuels. The more extensive
data base for certain sources, including human data, should make it
possible to draw some conclusions as to the utility of various
methods and approaches which may be generalized to other combustion
emissions.
The soot from combustion emissions has been judged to be
carcinogenic in humans (IARC, 1985) and many soot8 have also been
found to be carcinogenic in animals. These emissions are also
mutagenic in bacteria, mammalian cells, and certain rodent assays.
Several of the available animal models for carcinogenesis studie8,
however, do not always produce results which agree with the human
data. In vitro methods are therefore extremely useful for: (1)
quantitative comparisons of emissions, (2) mechanistic studies, (3)
bioavailability studies, and (4) identification of the active
components In mixtures.
BIOASSAY DIRECTED FRACTIONATION AND CHARACTERIZATION
The objective of this strategy is to identify the biologically
active (e.g., toxic, mutagenic) components in complex mixtures.
This and other approaches to identifying genotoxic compounds have
been reviewed by laxton 1982. The two principle considerations in
applying this approach are the bioassay to be employed and the
fractionatlonatlon method, both of which are discussed below.
The bioassay(s) chosen to be employed will depend on the complex
mixture and what is known about either the chemical composition or
toxicological effects of the mixture. Bloassays which have been
used include assays for: mutagenicity (e.g., Ames test), tumor
Initiation, cytotoxiclty, etc.
The methods which have been used to fractionate complex mixtures
prior to bioassay Include nearly all the methods which have been
used for separations prior to chemical analysis. Often several of
these methods are combined sequentially as shown in Figure 1. One
of the moat widely used approaches to separate mixtures is to use
2-20
-------
solvent partitioning to separate a mixture into the organics and
inorganics. These will each be bioassayed. If most of the activity
of interest is in the organic fraction, it will be further solvent
partitioned into organic acids, organic bases and neutral
compounds. If bioassay of these three fractions shows that the
activity is largely isolated into the organic neutral fraction, it
may be further fractionated by silica gel chromatography or high
pressure liquid chromatography (HPLC) into the aiphatics, aromatics,
moderately polar and highly polar neutral compounds. This
reiterative approach, diagrammed in Figure 1, is repeatedly used
until the active compounds have been sufficiently separated from the
nonactive components to allow chemical characterization of the
biologically active compounds.
At each point in the fractionation scheme where activity is
measured, recovery of mass and activity should also be determined.
A reconstituted mixture is bioassayed and compared to the
unfractionated (neat) mixture to determine the recovery of
bloactivity. If the sum of the activity of the mass weighted
fractions equals the unfractionated (neat) mixture then the
mutagenicity is considered to be additive. One should be cautious
to consider that simultaneous chemical changes as a result of the
fractionation procedure could both increase the activity of some
components and decrease the activity of others resulting in the
appearance of no change and therefore that biological activity is
recovered.
In several cases this approach has resulted in the identification of
biologically active components of mixtures which were very minor
components by mass but were extremely potent. Examples of such
components are the potent mutagenic dinitropyrenes found in
Xerographic toners (Rosenkranz et al., 1980) and diesel exhaust
(Schuetzle, 1983); the potent toxin, aflotoxin, found in peanuts;
and the potent polychiorinated biphenyls (e.g., TCDD) found in
emissions from burning transformers.
Once a highly active fraction is obtained and chemical analysis
provides a list of the confirmed or suspected components, one
critical requirement must be met before biologically active
compounds are identif led. Either bioassay data must be available
for each of the compounds identified, and/or pure standards need to
be obtained to bioassay the identif led compounds in order to
determine their contribution to the total activity observed.
CHARACTERIZATION AND IDENTIFICATION OF THE MUTAGENS AND CARCINOGENS
EMITTED FROM COMBUSTION SOURCES
Bioassay—directed fractionation closely coupled to chemical
characterization has been shown to be the most efficient and
effective approach to identifying the mutagenic and turorigenic
2-21
-------
compounds in combustion emissions. This approach has been used to
identify tumor initiators and tumor promoters in cigarette—smoke
condensates (Hoffman and Wynder, 1976), automotive exhaust emissions
(Grier et al., 1982), and urban—air particles (Hueper et al.,
1982). More recently, this approach has been coupled with
short—term genetic bioassays, including both microbial and
lian—cell mutation assays, to identify mutagens and potential
carcinogens in complex mixtures (Epler et al., 1978). We first
employed this method to identify the chemical classes and specific
components associated with diesel particulate emissions that were
mutagenic in the Ames Salmonella typhimurium mutagenesis assay
(Huisiugh et al., 1979).
Diesel particles collected by the dilution—tunnel method (Bradow,
1982) were Soxhiet extracted with dichioromethane and
solvent—partitioned into organic acids, bases, and neutral
components. The neutral components were further fractionated Into
paraf fins, aromatics, transitional moderately polar (TRiO compounds,
and oxygenated highly polar (OXY) compounds. The mutagenic activity
of each fraction was determined using the Ames Salmonella
typhimurium/microsome assay in TA1535, TA1537, TA98, and TA100
(Huisingh et al. 1979). The distributions of the mass of each
fraction and of its mutagenic activity In TA98 are shown in Table 1
for a four—stroke V—8 Caterpillar 3208 engine used in urban service
vehicles. The moderately and highly polar neutral compounds In the
TRN and OXY fractions account for 89—942 of the mutagenic activity
of the extractable organics and only 322 of the mass. Conventional
gas chromatography/mass spectroacopy identified many fluorenones and
aethylated fluorenones as major constituents of these fractions
which we found to be non—mutagenlc. None of these or other
identified constituents accounted for the direct—acting frameshift
mutagenic activity observed. Studle8 with nitroreductase—deficient
strains of Salmonella typhimurium showed in a reduction in the
mutagenicity of these organics, suggesting that nitrated compounds
contributed to this direct—acting mutagenicity ( laxton and
Huisingh, 1980). Nitrated polycyclic aromatic hydrocarbons
(N0 2 —PAHs) are potent direct—acting frameshift mutagens detected in
xerographic toners (Rosenkranz et al., 1980). A series of N0 2 —PAHs
In diesel extracts were later identified and quantitated in order to
estimate their contribution to the mutagenic activity of diesel
particulate emissions (Nishioka et al., 1982) as described In the
next section. Similar studies have also been conducted by
Schuetzle, 1983. These studies show that N0 2 —PAHs, di—N0 2 —PAHs, and
hydroxy—N0 —PAHs together account for much of the mutagenicity
observed In Salmonella typhimurium . Particulate emIssions from
catalyst—equipped gasoline—engine vehicles using unleaded fuel
contain significantly less of these N0 2 —PABs (Nishioka et al.,
1982). The mutagenic activity of both leaded— and unleaded-gasoline
emissions is substantially Increased with the addition of an
exogenous metabolic activation (MA) system, suggesting that the
2-22
-------
unsubstituted PARs may play a more important role than do N0 2 —PAHs
in the mutagenicity and carcinogenicity of gasoline emissions
(Lewtas, 1982, 1983).
Characterization of the mutagenicity of emissions from both
vegetative and fossil fuel show general similarities. Comparison of
the distribution of mutagenic activity in different organic
fractions extracted from a light—duty diesel vehicle and urban air
particle extracts in (Lewtas, 1985) Table 2. In both wood and
diesel combustion, 82—99% of the mutagenicity was in the neutral
fraction. Very little mass or mutagenic activity was observed in
the organic bases. Differences In these sources were observed in
the distribution of mass and mutagenicity in the neutral
subfractlons.
Bioassay—directed fractionation and chemical characterization has
also been used to characterize and compare the complex organic
emissions from roofing tar pots, coke ovens, and cigarette smoke.
To obtain a gross characterization of the chemical classes present
In the samples, the chemical class distribution was determined by
solvent partitioning the organics Into acidic, basic, neutral, and
cyclohexane Insoluble (a highly polar fraction) fractions. The
neutral fractions were further separated into the nonpolar neutrals,
aromatics (nitromethane soluble) and polar neutrals.
Characterization of the distribution of mutagenic activity (Ames
Salmonella typhimurluni bioassay) in each of these fractions (Austin
et al., 1985; Williams et al., 1986) showed significant differences
between the diesel, coke oven main, roofing tar, and cigarette smoke
condensate samples as shown in Fig. 2. In the diesel samples, over
90% of the mutagenic activity is located in the aromatic and
polar—neutral fractions, and a significant portion of this activity
can be accounted for by N0 2 —PAHs. The cigarette smoke condensate,
coke oven main, and roofing tar samples did not contain detectable
amounts of N0 2 —PAHs (Williams et al., 1986). Most of the
mutagenicty of coke oven main sample was found in the basic fraction
(37%) and polar neutral fraction (39%). The cigarette smoke
condensate sample also bad significant activity in the basic
fraction (66%), but chemical analysis indicated that the components
differed significantly from those of the coke oven main sample. The
roofing tar sample contained aromatic (14%) and polar (75%)
mutagenic constituents that were not N0 2 —PAHs. The PAR subfractIon
of each of these samples accounted for only a small portion of the
mutagenicty (e.g., diesel (0.2%), cigarette smoke condensate (0.1%),
roofing tar (5%) and coke oven main (8%).
Although the specific mutagens in these different sources are not
identical, they all cause frameshift mutations and appear to be
compounds that could be classified as polycyclic organic matter
(POM). Chemical characterization suggests that in addition to
nitrated N0 2 —PAJ-!s found in the slightly and moderately polar
2-23
-------
neutrals, hydroxylated and carboxylated polycyclic organics are
found In the organic acid fraction, aromatic amines and nitrogen
heterocycles are found in the organic bases and highly oxygenated
qulnones, diones, and nitro—oxygenated compounds are found in the
polar neutral fractions.
CONTRIBUTION OF NO -PABs TO ThE MUTAGENIC ACTIVITY OF AUTOMOTIVE
EMISSIONS AND URBAN AIR PARTICLES
Although N0 2 —PAHs were qualitatively identified In extract of diesel
particles and urban air particulate matter as described earlier, the
detection and quantification of these compounds at low levels has
posed problems for analytical chemists because the conventional
analytical techniques for quantifying PARs [ e.g., high performance
liquid chromatography (HPLC) with fluorescence detection and gas
chromatography (GC)Imass spectrometry (MS) with electron Impact
ionization (El)] are relatively insensitive to nitro—substituted
PAils. The 1, 3—; 1,6—; and 1, 8—dinitropyrene isomers are so highly
iautagenic in the Ames (TA98 —S9) assay that trace concentrations of
these compounds, if present, could account for a major proportion of
the observed mutagenic activity. A capillary column GC/MS
analytical technique using on column Injection and negative chemical
ionization (Nd) detection was therefore developed to detect these
and other nitro—PARs at very low concentrations (Nishioka et al.,
1982). This technique was applied to extracts of soot particles
from diesel and gasoline vehicles and urban air particles.
Twenty—three different N0 2 —PARa were identified in the diesel engine
extracts; five N0 2 —PAHs in the urban air particle extracts; and only
one, l—nitropyrene was detected in the unleaded gasoline engine
extract. 1—Nitropyrene was the N0 2 —PAH detected In greatest
quantity in the diesel extracts (107—1590 ppm, relative to the
weight of the extract), followed by the nitrophenanthrene/anthracene
isomers. The only dinitro—PARa for which analytical standards were
available, the dinitropyrene isomers, were detected in one diesel
extract sample at sub—ppm concentrations (0.4—0.6 ppm).
Quantification of the concentration of these N0 2 —PARs and
determination of their contribution to the direct—acting
autagenicity in Salmonella typhimurium TA 98 (Table 3) shows that
the aono—N0 2 —PABs make relatively minor contributions to the
mutagenicity of the urban air particles ( 2%). 1—Nitropyrene was
present at the highest concentration (107—1590 ppm) in the diesel
particle extracts, and yet It accounted for only 3—13% of the
mutagenicity. The concentration of several of the N0 2 —PARs in a
series of extracts of diesel and gasoline exhaust particles were
however highly correlated with their mutagenicity both in S.
typhimurium and L5178Y mouse lymphoma cells and skin tumor
initiating activity in Sencar mice. 3—Nitrofluorauthene present at
1 ppm to 7 ppm In the diesel samples accounted for 0.8% to 1.4% of
the autagenicity. By using the mutagenicity values determined in
2-24
-------
separate experiments for 2—nitrofluorene and those reported in the
literature for l—nitroaphthalene, these compounds were estimated to
account for less than 0.01% of the mutagenic activity. Although the
dinitropyrene isomers: 1,3; 1,6; and 1,8 were detected only in
diesel sample C at 0.4—0.6 ppm (sum of 1.6 ppm), their mutagenic
activity (496,000; 629,000; and 870,000 rev/ug, respectively) was
high enough to account for 26% of the mutagerticity of this sample.
The total “direct—acting” mutagenic activity in S. typhimurium TA98
that can be accounted for by the 23 nltro—PAHs quantified in Diesel
C Is 40%. This estimation is supported by the loss of 50% the
mutagenic activity of this extract it was assayed in TA98FRI
(renamed TA98NRD), a classical nitroreductase—deficlent S.
typhimurium tester strain obtained form H. Rosenkranz (Rosenkranz et
al., 1981).
The fact that dinitropyrenes at concentrations below the ppm level
can account for nearly one—third of the mutagenic activity (—S9)
suggests that the presence of other highly potent dinitro—PAHs may
account for even more of the mutagenic activity (—S9) in the
moderately polar neutral fraction. A recent application of S.
typhlmurIum tester strains developed to exhibit resistance to
dinitropyrenes (e.g., TA98/l,8DNP 6 ) has led to even larger
estimations of the contributions of dinitropyrenes to the
mutagenicty of diesel particle extracts. The lack of quantitative
data on the concentrations of the dinitropyrene isomers in such
samples has previously made it impossible to confirm whether the
concentrations of the dinitropyrenes are indeed sufficient to
account for the contribution predicted by the resistant tester
strains. Pederson, 1983, has reported that in several light—duty
diesel particle extracts, 50% to 90% of the TA98 (—S9) mutagenicty
is lost when the extracts are tested in TA98/l,8DNP 6 . Since these
strains may show a resistance to other dinltro—PAHs, it is possible
that highly potent dinitro—substituted isomers of other parent PAils
present in the particle extracts may also contribute to the
mutagenicty of the moderately polar neutral fraction.
CHARACTERIZATION AND IDENTIFICATION OF MUTAGENIC METABOLITES
When toxicological studies are conducted on an Individual chemical
(e.g., benzo [ a]pyrene) which is metabolized to form a number of
metabolites, we don’t generally consider the resulting mixture of
metabolites to be a “complex mixture.” The strategy which is
described above for identifying mutagens In environmental mixtures
can also be applied to the identification of mutagens in a mixture
of metabolites.
The Identification of the bacterial mutagen, 1—nitropyrene (NP), In
diesel emissions and ambient air as described above, led us to
conduct a series of studies on the mammalian metabolism of NP. Ball
et al., 1984 reported that rats treated with NP excreted a mixture
2—25
-------
of mutagenic aetabolites. This mixture was separated by HPLC into
fractions which were bioassayed in the Ames S. typhimuriuni plate
Incorporation assay and characterized using high resolution CC/MS.
Among the principal metabolite fractions identified were 3’-hydroxy—
1—nitropyrene, 8—hydroxy—1—ni tropyrene, 6—hydroxy—N—acetyl —1 —amino -
pyrene, and 8-hydroxy—N—acetyl—l—aainopyrene. We used a similar
approach to Identify the autagenic metaboll tea of NP after metabo—
lis a by lung S9 (Xing et a]., 1984) and lung and tracheal cells
(King et al., 1986). These studies led to the Identification of a
whole series of autagenic hydroxylated nitropyrenes where the OH
group was In the 3, 6, 8 and 10 position. An example of the use of
the Ames assay to identify the autagenic metabolites of
1—nitropyrene formed by lung S9 Is shown In Table 4. This approach
led to further efforts to characterize the highly mutagenic
fractions (e.g., ) and the various Isomers of O}F-nltropyrene.
IDENTIFICATION OF HYI 0XYLATED NITRO-PABa IN URBAN AIR
We have recently Identif led and quantified hydroxy—N0 2 --PAHs in both
a polar neutral fraction and acidic fraction of an ambient air
particulate extract using bioassay-directed fractionation (Nishioka
et al., In press). Figure shows the overall fractionation and
bioassay data. The three fractions of greatest interest were the
acid, .ethylene chloride and methanol fractions which contained 38%,
232, and 29%, respectively, of the recovered mutagenlelty. Since
the aethylene chloride fraction was more amenable to
aubfractionation by existing HPLC methods, this fraction was chosen
for further fractionation. Normal phase HPLC was used to
fractionate the aethylene chloride fraction (24). The first HPLC
sub—fractionation resulted in one fraction, methylene chloride—C,
which contained nearly 50% of the iautagenlcity in 25% of the mass.
Analysis of this fraction by both El and NCI HRGC/MS techniques
indicated that the complexity of the sample was not reduced
sufficiently to Identify Individual components. Sub—fraction C was
further fractionated by normal—phase HPLC into 12 fractions for
bioassay. As we proceeded to higher levels of sub—fractionation,
less and less mass was available for bioassay. In this study, the
first level fractions obtained from the preparative techniques were
bloassayed In triplicate at seven doses with and without activation
In order to be able to quantitate the distribution and recovery of
autagens. Aa we proceeded to the first and second level of
fractionation, fewer doses and plates were used for the bioassay
until nearly the entire fraction was used for a single plate.
Therefore, our ability to quantitate the recovery of both mass and
autagenicty decreases as we subfractlonate. The bioassay
chromatogram showed peaks of relatively higher mutagenicty in
fractions 4 and 11 with 29 and 37 rev/ug, respectively. Mass
spectral analysis of peak 4 demonstrated that the goal of sequential
fractionation had been achieved in that Individual components could
be resolved and quantified using IIRGCIMS. The major mutagenic
2-26
-------
compounds identified in the nethy1eue chloride—C—4 fraction were
hydroxy—nitro—substituted fluoranthenes (OH—N0 2 —PAH) and work is
continuing to synthesize and identify other isomers.
MICRO—BIOASSAY METHODS COUPLED TO ANAL’YTICAL SEPARATION TECHNIQUES
The quantity of material (mixture or fraction) needed for many of
the established bioassay methods ranges from milligram quantities
(e.g., Ames assay) to gram quantities (e.g., tumor initiation).
Recently a number of micro—bioassay methods have been reported for
measuring forward mutation (Thilly et al., l983 Goto et al.,
Submitted), reverse mutation (Kado et al., 1983), in Salmonella
typhlmurium and phage induction in E. coil (Rossman et al., 1985).
Such micro—bioassay methods will facilitate the coupling of bioassay
methods to analytical scale separation techniques such as
reverse—phase HPLC. Thilly demonstrated the possibilities of this
approach by applying the forward mutation micro—assay to particle
emission extracts from HPLC fractions of a residential oil burner.
Figure 4 demonstrates the utility of this approach when this assay
is used to bioassay the fractions from a mixture of metabolites of
1—nitropyrene incubated with lung S9. Such a bioassay chromatogram
facilitates the immediate visualization of the most mutagenic
fractions for further characterization. This approach should
eliminate the need for large—scale fractionation techniques which
are often fraught with problems of the spillover of chemicals from
one fraction to another and with poor resolution. High resolution
fractionation techniques used sequentially may make it possible to
result iii fractions which contain either one chemical or a very
small number of chemicals which will greatly facilitate
characterization of the fractions.
COMPARATIVE POTENCY EVALUATION OF COMPLEX MIXTURES
Comparative potency studies of a series of substances in one or more
bioassay is a particularly useful approach to the toxicological
evaluation of complex mixtures. Many complex mixture problems
involve the assessment of a series of different mixtures, such as,
fuels derived from various sources, soots emitted from different
combustion engines, or air particles from various exposure
environments. Since all soots from Incomplete combustion are
expected to contain carcinogenic polycyclic organics, the question
often posed Is whether a new combustion source will produce soot
which is more carcinogenic than the existing sources.
The comparative potency approach has been used in comparing the
acute toxicity of a series of petroleum hydrocarbons using as range
of products form light oils and gasoline to heavy fuel oils (Beck et
al., 1982. The toxicity of shale derived fuels and other syn—fuels
have been determined primarily in comparative studies In general
toxicity tests, target organ studies, behavioral studies, muta—
2—27
-------
genesis, carcinogenesis, teratology and neurotoxicity tests. Many
of these studies are reviewed by MacFarland et al., 1982.
The design considerations for comparative potency studies include a
number of factors which may not always be considered when using
other toxicology methods. These include: 1) simultaneous
evaluation of all comparisons in one experiment, where possible; 2)
exposure doses (or concentrations) needed for statistical analysis
and potency measure; and 3) fractional factorial design for mixtures
strategy for evaluating the toxicity of mixtures of gasoline blends
using fractional factorial designs, multiple bioassays and a
standardized reference fuel as a center point.
DEVELOPMENT OP A COMPARATIVE POTENCY METHOD FOR CANCER RISK
ASSESSMENT OF COMBUSTION F14ISSIONS
A comparative potency method for cancer risk assessment has been
developed based upon a constant relative potency hypothesis. This
method was developed and tested using data from a batter of
short—term mutagenesis bioassays, animal tumorigenicity data and
human lung cancer risk estimations (Lewtas, 1981; Nesnow et al.,
1982a,b; Albert et al., 1983; Lewtas et al., 1983). This data base
was developed from a series of complex mixtures including emissions
from coke ovens, roofing tar pots, cigarette smoke, and automotive
engines.
The comparative potency method for cancer risk assessment Is based
on the hypothesis that there is a constant relative potency between
two different carcinogens (C 1 and C2) across different bioassay
systems (B 1 and B2). The mathematical expression for the constant
relative potency model Is the following;
Cl Cl
Relative Potency — in Bioasaay(i) (k) Relative Potency In Bioassay( 2 )
C 2 C 2
This assumption is implicit in any comparison which utilizes the
relative toxicity of two substances in animals to determine which
substance would moat likely be more or less toxic to man. This
constant relative potency assumption Is a testable hypothesis, if
the relative potency of two mixtures or components In one bioassay
(e.g., humans) can be determined and compared to the relative
potency in a second bioassay. The test of this model Is whether
there is a constant relationship (k) between the relative potencies
in the two bioaaaaya being compared such that:
Relative Potency (C14 C2) in Bioasaay(l ) — constant (k)
Relative Potency (Cl C2) in Bioassay(2)
2-28
-------
This hypothesis was tested for three complex organic emissions from
a coke oven, roofing tar pot, and cigarettes by using human lung
cancer data from epidemiological studies of humans exposed to these
emissions and testing these emissions in a series of short—term
mutagenesis bloassays and animal tumorigenesis assays (Lewtas, 1985).
COMPARATIVE ASSESSMENT OF THE MUTAGENICITY, TIJMORIGENICITY AND
NITRO—PAB CONTENT OF DIESEL AND GASOLINE EMISSIONS
The mutagenicity of the extractable organic material from diesel and
gasoline emissions has been reported for a battery of in vitro
bioassays detecting gene mutations, DNA damage, chromosomal effects,
and oncogenic transformations (Lestas, 1983). Skin tumor initiation
and carcinogenic activity of these samples has also been reported in
SENCAR mice (Nesnow, 1982). Very good correlations (r 2 0.90) were
observed when the slope of the dose—response for the mutagenic
activity in S. typhimurium strain TA98(—MA) was plotted versus the
mutagenic activity in the mammalian cell assays and the skin tumor
initiating activity.
The correlation of mutagenic and skin tumor initiating activity with
the concentration of selected nltro—PAHS was examined for these
diesel and gasoline samples. Table 5 shows the high correlations
observed between concentrations of l-nitropyrene and
3 —nitrofluoranthene in these samples, and their mutagenic activity
in S. typhimurium (—s9), L5178Y mouse lymphoma cells (—S9), and skin
tumor initiating activity in SENCAR mice. The r 2 correlation
coefficient in the presence of S9 (not shown) was somewhat lower.
The inutagenicity and tumor initiation activity (—S9) also correlated
well (r 2 )0.9) with the concentrations of the nitro—252 Isomer, the
nitro—228 isomers, nitromethylpyrenes and nitrofluorenes while no
correlation was observed with several of the other less mutagenic,
lower molecular weight nitro—PAHs (nitrophenanthrenes and
nitronaphthalenes).
The quantified mono—N02—PAHS account for less than 20% of the
direct—acting bacterial mutagenicity of these samples. The high
correlations observed between the concentration of these compounds
and the mutagenic activity of the total extract, therefore, suggest
that the concentrations of the unidentified mutagens responsible for
the remainder of the mutagenic activity and possibly the mutagenic
and carcinogenic activity in other bioassays is directly related to
the relative concentrations of these mono—N02—PAH8. Because the
remaining unidentified inutagens appear to be primarily located In
the chemical fractions that are more polar than the fraction which
contains the unsubstituted and mono—NO 2 —PAHs, it Is possible that
other di—N0 2 —PAHs (e.g., dinitrofluoranthenes) or other oxygenated
N02—PAH’s (e.g., hydroxy—nltro—PAHs) species are primarily
responsible for the major portion of the unidentified inutagenic
activity.
2-29
-------
SUMMARY
Toxicology studies of complex mixtures nearly always employ the same
general methods, models and bioassays that have been developed and
validated using single agents. The strategies for combining and
applying these methods however may differ due to the complexity of
the mixture and the questions being addressed. In initial studies
to determine the toxicological effects of an unknown mixture, the
approach is generally to administer the mixture as you would a
single agent in standard in vivo and in vitro bloassays. Attempts
to approach the toxicology of a complex mixture from an analysis of
its components are generally successful only if the number of
components is limited (2—5 components) and well defined. In
approaching a very complex mixture where both the toxicological
effects, Components and active agents are unknown, the two
strategies described above can be very useful. Prior to undertaking
either of these two strategies, however, the toxicological endpoint
of concern must be identified.
The two strategies described here have been successfully applied to
the evaluation of different complex mixtures which cause a variety
of effects. Wastewaters which are acutely toxic to fish have been
comparatively evaluated to determine which wastewaters, industrial
effluents, or sites are the most toxic using acute toxicity assays
in fish. Bioassay—directed fractionation techniques have also been
used to identify the class of compounds (e.g., organic vs. inorganic
etc.) responsible for this toxicity.
Recent inhalation studies of unleaded gasoline demonstrated renal
toxicity and carcinomas in rodents. Toxicology studies of the renal
toxicity of different fractions of the gasoline led to the
Identification of Isoparaff Ins as the active agents.
It Is Important to keep in mind that the two strategies described
here are limited by the limitations of the bioassay methods
employed. If, for example, you wish to identify carcinogens in a
complex mixture which you know contains chlorinated hydrocarbons,
then the Ames Salmonella typhimurium bioassay would not be a wise
choice since It is well documented that this bioassay does not
detect many carcinogens of this class. In some cases several
bioassay methods may have to be applied in using these strategies to
comparatively evaluate mixtures and elucidate the active
toxicological agents.
DISCLAIM
The research described in this paper has been reviewed by the Health
Effects Research Laboratory, U.S. Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency
2-30
-------
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
REFERENCES
Albert, R. E., Lewtas, J., Nesnow, S., Thorsiund, T. W., and
Anderson E. (1983). Comparative potency method for cancer risk
assessment: Application to diesel particulate emissions. Risk
Anal . 3, 101—117.
Austin, A. C., Claxton, L. D. and Lewtas, J. (1985). Mutagenicity
of the fractionated organic emissions from diesel, cigarette
smoke condensate, coke oven, and roofing tar in the Ames
Assay. Environ. Mutagen . 7, 471—487.
Bail, L. N., Kohan, M. J., Inmon, 3. P., Claxton, L.D., and Lewtas,
J. (1984) Metabolism of 1_nitro [ 1 4 C]pyrene in vivo in the rat
and mutagenicity of urinary metabolites, Carcinogenesis 5,
1557—1564.
Beck, L. S., Hepler, D. I., and Hansen, K. L. (1982). The acute
toxicology of selected hydrocarbons. In The Toxicology of
Petroleum Hydrocarbons (H. N. MacFarland, C. E. Holdsworth, J.
A. MacGregor, R. W. Call, and N. L. Kane, Eds.), pp. 1—12,
American Petroleum Institute, Washington, D.C.
Bradow, R. L. (1982). Diesel particle and organic emissions:
engine simulation, sampling, and artifacts. In Toxicological
Effects of Emissions from Diesel Engines (J. Lewtas, Ed.),
p. 33—47. Elsevier, New York.
Claxton, L. D. (1982). Review of fractionation and bioassay
characterization techniques for the evaluation of organics
associated with ambient air particles. In Genotoxic Effects of
Airborne Agents (a. R. Tice, D. L. Costa, K. M., Schaich,
Eds.), pp. 19—33. Plenum Press, New York.
Claxton, L. D., and Huisingh, 3. L. (1980). Comparative niutagenic
activity of organics from combustion sources. In Pulmonary
Toxicology of Respirable Particles (C. L. Sanders, F. T. Cross,
G. E. Dagle, 3. A. Nahaf fey, Eds.), CONF—791002, DOE Symposium
Series 53, pp. 453—465. Department of Energy, U.S. Government
Printing Office, Washington, D. C.
Epler, 3. L., Clark, B. R., Ho, C.—h., Guerin, M. R., and Rao, T. K.
(1978). Short—term bioassay of complex organic mixtures: part
II, mutagenicty testing. In Application of Short—Term
Bioassays in the Fractionation and Analysis of Complex
Environmental Mixtures (N. D. Waters, S. Nesnow, J. L.
Huisingh, S. S. Sandhu, and L., Claxton, Eds.) pp. 269—289.
Plenum Press, New York.
2-31
-------
Feder, P. I., Margosches, E., and Bailar, 7. (1984). A strategy for
evaluating the toxicity of chemical mixtures. Draft report.
EPA Contract No. 68—01—6721.
Coto, S., Wiillama, K., Claxton, L. D.., and Lewtas, J. (Submitted).
A forward mutation micro—assay in Salmonella typhimurium :
Modifications and applications to complex environmental
emissions, Mutat. Res .
Grier, G., Naujack, K.—W., Dettbarn, G., Brune, H.,
Deutsch-Wenzel, B., and Misfeld, J. (1982). Analysis of
balance of carcinogenic impact from emission condensates of
automobile exhaust, coal heating, and used engine oil by
mouse—skin—painting as a carcinogen—specific detector. In
Polynuclear Aromatic Hydrocarbons (N. Cooke, A. J. Dennis, and
G. L. Fisher, Ma.) pp. 335—345. Battelle—Columbus Press,
Columbus, Ohio.
Hoffman, D., and Wynder, E. L. (1976). Environmental respiratory
carcinogenesta. In Qiemical Carcinogens, ACS Monograph 173 ,
(C. E. Searle, Ed.) Cbap. 7, American ( iemical Society,
Washington, D. C.
Hueper, W. C., Kotin, P., Tabor, H. C., Payne, W. V., Falk, H. and
Sawicke, H. (1962). Carcinogenic bloassays on air pollutants,
Arch. Pathol . 74, 89—116.
Huisingh, J. Bradow, B., Jungera, B., Claxton, L., Zweidln.ger, B.,
Tejada, S., Buagarner, S., Duffield, F., Waters, H., Simmon, V.
F., Hare, C., Rodriguez, C., and Snow, L. (1979). Application
of bioassay to the characterization of diesel particle
emissions. In pplication of Short—Term Bloassays In the
Fractionation and Analysi8 of Complex Environmental Mixtures
(H. D. Waters, S. Nesnow, S. L. Huisingh, S. S. Sandhu, and L.,
Claxton, Eds.), pp. 381-418. Plenum Press, New York.
International Agency for Research on Cancer (LARC) Monograph
programme on the evaluation of carcinogenic risk of chemicals
to humans, Polynuclear Aromatic Compounds, Part 4, Bitumens,
Coal—tar and Derived Products, Shale Oils and Soots, 1985, Vol.
35, 271 pp.
Kado, N. Y., Langley, D., and Elsenstadt, E. (1983). A simple
modification of the Salmonella liquid—incubation assay;
Increased sensitivity for detecting mutagens in human urine.
Mutat. Rea . 121, 25—32.
King, L. C., Kohan, H. S., Ball, L. H., and Lewta8, S. (1984).
Mutagenicity of l—nltropyrene metabolites from lung S9. Cancer
Lett . 22, 255—262.
2-32
-------
King, L. C., Ball, L. M., Jackson, M., Inmon, J. P., and Lewtas, J.
(1986). Metabolism of 1-nitropyrene by cultured rabbit
alveolar macrophages and respiratory tract tissues. Toxicol.
Appi. Pharniacol . 82, 292—300.
Lewtas, J. (1982). Mutagenic activity of diesel emissions. In
Toxicological Effects of Emissions form Diesel Engines , (J.
Lewtas, Ed.), pp. 243—264. Elsevier Science Publishing Co.,
Inc., New York.
Lewtas, J. (1983). Evaluations of the mutagenicity and
carcinogenicity of motor vehicle emissions in short—term
bloassays. Environ. Health Persp . 47, 141—152.
Lewtas, J. (1985). Combustion emissions: tharacterization and
comparison of their mutagenic and carcinogenic activity. In
Carcinogens and Mutagens in the Environment, Volume V, The
Workplace , (H. F. Stich, ed.), pp. 59—74. CRC Press, Boca
Raton, Florida.
Lewtas, J. (1986). A quantitative cancer risk assessment
methodology using short—term genetic bloassays: the
comparative potency method. In Risk and Reasons: Risk
Assessment in Relation to Environmental Mutagens and
Carcinogens , (P. Of tedal, Ed.), pp. 107—120, Alan R. Liss,
Inc •, New York, New York.
Lewtas, J., Bradow, R. L., Jungers, R. H., Harris, B. D.,
Zweidlnger, R. B., Cushing, K. N., Gill, B. E., and Albert, R.
E. (1981). Mutagenic and carcinogenic potency of extracts of
diesel and related environmental emissions: study design,
sample generation, collection and preparation. Environ. mt .
5, 383—387.
I4acFarland, H. N., Holdsworth, C. E., MacGregor, J. A. Call, R. W.,
and Kane, N. L., Eds. (1982) The Toxicology of Petroleum
Hydrocarbons , 384 pp. American Petroleum Institute, Washington,
D. C.
Nesnow, S., Triplett, L., and Slaga, T. J. (1982). Comparative
tumor—initiating activity of complex mixtures from
environmental particulate emissions on SENCAR mouse skin. J.
Nati. Cancer Inst. , 68, 829—834.
Nesnow, S., Evans, C., Stead, A., Creason, J., Slaga, T. J., and
Triplett, L. L. (1982). Skin carcinogenesis studies of
emission extracts. In Toxicological Effects of F nissions from
Diesel Engines (J. Lewtas, Ed.), pp. 295—320. Elsevier Science
Publishing Co., Inc., New York.
2—33
-------
Nishioka, M. C., Petersen, B. A., and Lewtas, J. (1982). ComparIson
of the nitro—aromatic content and direct—acting mutagenicity of
diesel emissions. In Polynuclear Aromatic Hydrocarbons (M.
Cooke, A. J. Dennis, and C. L. Fisher, Fda.), pp. 603—613.
Battefle —Columbus Press, Columbus, Ohio.
Nighioka, N. C., Howard, C. C, and Lewtas, J. (In Press).
Detection of hydroxylated nitro-polynuclear aromatic
hydrocarbons in an ambient air particulate extract using
bioassay—directed fractionation. Environ. Sd. Technol .
Nishioka, N. C., Peterson, B., and Lewtas, 1. (1983). Comparison of
nitro—aromatic content and direct-acting mutagenicity of
passenger car engine emissions. In Mobile Source Enissions
Including Polycyclic Organic Species (D. Rondia, N. Cooke, and
K. K. Haroz, Eds.), pp. 197—210, Reidel Press, 1 rdrecht,
Holland.
Pedereon, T. C. (1983). Biologically active nitro—PAH compounds in
extracts of diesel exhaust particulate, ibid pp. 227—246.
Rosenkranz, H. S., McCoy, E. C., Mernelatein, R., and Speck, W. T.
(1981). A cautionary note on the use of nitroreductase—
deficient strains of Salmonella typhlmurlua for the detection
of nitroarenes as mutagens in complex mixtures including diesel
exhausts. Mutat. Rae . 91, 103—150.
Roaenkranz, H. S., McCoy, E. C., Sanders, D. K., Butler, N.,
Klriazides, D. K., and !4ermelstein, R. (1980). Nitropyrenes:
isolation, identification, and reduction of mutagenic
impurities in carbon black and toners. Science , 209, 1039—1043.
Roasman, T. C., Meyer, L. W., Butler, J. P. and Daisey, J. N.
(1985). Use of the microscreen assay for airborne particulate
organic matter. In Short—Term Bioassays in the Analysis of
Complex Environmental Mixtures IV (M. D. Waters, S. S. Sandhu,
J. Lew-tas, L. Claxton, C. Strauss, and S. Nesnow, Eds.), pp.
9—24, Plenum Press, New York.
Schuetzle, D. (1983). Sampling of vehicle emissions for chemical
analysis and biological testing. Environ. Health Persp. , 47,
65—80.
Thilly, W. C., Longwell, J., and Andon, B H. (1983). Ceneral
approach to the biological analysis of complex mixtures.
Environ. Health Perspect . 48, 129—136.
Williams, K., Sparacino, C., Petersen, B., Buagarner, J., Jungers,
K. H., and Lewtas, J. (In Press. Comparative characterization
of organic emissions from diesel particles, coke oven mains,
roofing tar vapors, and cigarette smoke condensate. mt. J. of
Eaviron. Anal. Qiem .
2-34
-------
TABLE I
Distribution of the Mass and Mutagenic Activity of Fractionated Diesel Particle Organtcsa
aDiesel particles obtained by the dilution tunnel method from a four—stroke V—8 Caterpillar 3208 engine operated
on the standard 13 mode series of steady states.
bDichloromethane extractable organics were solvent—partitioned into acids, bases, and neutrals as previously
described. The neutral fraction was further fractionated on silica gel into four fractions by the following
solvents: hexane (paraf fins), 1% ether in hexane (aromatics and transitionals), and 50% acetone in methanol
(oxygenates).
cSalmonella typhimurium mutagenesis assay performed with and without metabolic activation (S9). Slope determined
from linear regression analysis of the initial portion of the dose—response curve.
dweighted mutagenic activity of each fraction relative to the total extract is the product of distribution of
mass in each fraction (% mass) and the specific iuutagenic activity (rev/mg). The sum of these weighted muta—
genie activities is used to determine the distribution of mutagenicity (%) as described below.
eDistribution of mutagenic activity is the percentage of weighted mutagenic activity in each fraction compared to
the sum of the weighted mutagenic activities. This distribution assumes that any mutagenicity not recovered
through the fractionation is equally distributed across all fractions.
Fractionb
Mass
Specific
Mutagenic Activi tyC
(rev/mg)
Weighted
Mutagenic Act!. vity 1
(rev/mg)
Distribution of
Mutageni c Act!. vi tye
(%)
(%)
—S9 +S9
—S9
+S9
—S9
+S9
Organic acids (ACID)
14.9
193 248
28.8
37.0
4.9
9.5
Organic bases (BASE)
0.3
43.8 132
0.13
0.40
0.02
0.10
Ether insolubles (INS)
3.9
53.9 80.9
2.1
3.2
0.36
0.80
Paraffins (PRF)
36.7
Neg. Neg.
0.0
0.0
0.0
0.0
Aromatics (ARM)
6.9
49.5 30.1
3.42
2.1
0.60
0.54
Transitionals (TRN)
5.0
7520 2620
376
131
64.9
33.5
Oxygenates (OXY)
26.9
629 798
169
215
29.2
55.4
-------
TABLE 2
Characterization of Mutaganicity from Combustion Sources and Urban Air
Fractiona Diesel Autob Wood Stovec Urban Aird
Ames TA98 +S9 Ames TA98 +S9 Ames TA98 +S9
mass mUtagenicitye mass mutagenLcity mass mutagenicity
1. organic acids 7.6 17.9 44.0 0.0 7.3 21.0
2. organic bases 0.2 <0,1 2.2 4.1 0.8 1.0
3. neutrals
A 4 aliphatics (hexane) 56.1 0.0 14.2 27.9 21.5 10.9
B, aromatics (hexane/benzene) 12.4 37.0 18.1 23.3 13.6 12.1
C. moderately polar (di — 3.1 24.7 29.0 11.6 8.6 45.0
chioromethane)
D. highly polar (methanol) 20.6 20.4 38.7 32.6 34.0 21.0
ame acids and bases were first separated by liquid—liquid partitioning to separate the acids and bases.
The neutrals are partioned into four fractions by open column chromatography on 5% water deactivated silica gel
utilizing the solvents designated in parenthesis.
bDjesel particles collected by the dilution tunnel technique from a 1973 Datsun Nissan 220C diesel automobile
operated on a chassis dynamometer using the highway fuel economy test cycle (L4WFET). The total exhaust was
diluted (10:1) with filtered air prior to collection on 20 x 20 inch Teflon coated Pailfiex T60—A20 filters.
The particles were Soxhiet extracted with dichioromethane (Dot) for 48 hrs prior to fractionation.
CWood combustion particles collected by the dilution tunnel technique from an airtight woodstove burning oak
(Johnson Energy Converter) during the constant burn phase. The emissions were diluted, collected, and
Soxhlet extracted by the same techniques employed for the diesel particles.
dUrban air particles (<1.7 tim) collected by a massive air volume sampler in an industrial/residential area of
Philadelphia, PA. The particles were Soxhiet extracted with DGI for 24 hrs prior to fractionation.
eData from the Salmonella typhimurium plate incorporation assay in TA98 with metabolic activation was analyzed by
the model slope method to determine the specific mutagenic activity in revertants/Mg of fraction tested,
Distribution of mutagenic activity between each fraction (% mutagenicity) is the percentage of weighted mutagenic
activity Imass (%) x specific mutagenic activity (rev/pg) of each fractionj of each fraction compared to
the sum of the weighted mutagentc activities.
-------
TABLE 3
Contribution of N02—PAHs to the Mutagenic Activity of
Particle Extracts in Salmonella typhimurium TA98 (—S9)
Extract Sample
Diesel Auto 1
Diesel Auto 2
Diesel Auto 3
Gasoline Auto 4
Urban Air A
Urban Air B
Mutagenic Activity
(revh’g) TA98 —S9
13.
3.9
3.5
1.6
1.8
4.4
Concentration and Contribution to Mutagenic Activity (%)
1—nitropyrene nitrofluoranthene dinitropyrene isomers
Cone. Contrib. Conc. Contrib. Cone. Contrib.
(ppm) (7.) (ppm) (7.) (ppm) (%)
1590 11. 7.0 1.4 —— — —
589 13. 1.2 0.8 1.6 26.
107 2.7 0.9 0.8 — — —
2.5 0.1 — —— — —
1.0 0.05 0.8 1.8 — —
0.2 0.004 0.4 0.4 — — —
2—37
-------
TABLE 4
Mutagenicity of Lung S9 MetaboLite Fractions Isolated by HPLCa
HPLC Sped ftc Mut agent city
Ketabolite Fraction (revertants/naol)b
—S9 +S9
Solvent Front 55 14
a 554 70
1250 113
6 168 44
K—NW 373 94
6 201 94
NAAP 62 347
10—OH—1NP 77 215
1—AMP 157 151
Phenols 84 168
3—OH—1NP 584 222
1—NP 315 227
126 142
aI4 _4.. p (2773 nmol) incubated with rabbit lung S9 (2.8 g protein) for
45 mm as described in King et al ., 1984.
bMutagenicity determined in the Ames Salmonella typhiinurium plate
incorporation assay in TA98.
2-38
-------
TABLE 5
Correlation Analysis of Nitropyrene and Nitrofluoranthene
Concentrations with Mutagenic and Tumorigenic Activity
Ames Mouse Tumor
Particle Extract 1 _ Npa 3 _ NFb TA98 —S9 Lymphoma —S9 Initiation
ppm ppm rev! gC MF/jig/m1’ pap/mouse/i ge
Diesel Auto 1 1590 7.0 13.0 4.2 590
Diesel Auto 2 589 2.9 3.9 0.98 240
Diesel Auto 3 107 1.2 3.5 1.2 310
Gasoline Auto 4 2.5 0.9 1.6 0.38 170
Correlation Coef.
0.91
0.98
0.82
r 2
with
1—NP
r 2
with
3—NF
>0.99
0.99
0.95
a l...Nitropyrene
b3_Nj trofluoranthene
Cp vertants per i g
dMutation frequency (mutants per 106 survivors) per iig per ml
epapillomas per mouse per g in SENCAR mice
2-39
-------
FIGURE LEGEND
1. General Schematic Diagram of Bioassay Directed Fractionation and
Characterization
2. DistributIon of Mutagenicity in Coke Oven Emissions (Coke),
Cigarette Smoke Condensate (csc), Diesel Emissions, and Roofing
Tar Emissions
3. Bioassay Directed Fractionation of Urban Air Particulate
4. HPLC Bioassay airomatogram of the 1—Nitropyrene Metabolites from
Lung S9
2-40
-------
BIOASSAY DIRECTED
FRACTIONATION AND
CHARACTERIZATION
Complex Mixture
1. Fractionate:
2. Bioassay:
3. Fractionate:
4. Bioassay:
5. Characterize:
Mass:
20%
80%
Activity:
5%
95%
I I
Bi B2 B3
Mass:
70%
5%
25%
Activity:
0%
25%
75%
4
Identify Components
6. Determine if identified components are
bio-active.
7. Chemically quantitate bio-adive components.
8. Determine the contribution of identified active
components to the total bioassay activity.
FIGURE 1
Fraction A
Fraction B
2-41
-------
C13
+
I-
I
70
60
50
40
30
20
10
0
Acid Base NPN PNA1 PNA2 PNA3 PNA4 PN C l
Fractions
FIGURE 2
2-42
-------
Particulate Extract
Acids
MASS (%) 7
MUTAGENICITY (%) 38
1° Hexane
MASS (%) 21
MUTAGENICITY (%) 0
2°
MASS (%)
MUTAGENICITY (%)
3°
I
Acid/Base Partitioning
Neutral Compounds
92
Sili a Gel Column
Chromatography
I
Methanol Acidic
Hexane/ Methylene
Benzene Chloride
14 9
8 23 HPLC
I — I 1
A B C D E
5 65 25 5 1
5 30 50 20 1
HPLC
a
C 4
0 )
C)
.0
1 .4
0
CA
.0
I
1....
MOC I2-C-4
Analysis
FIGURE 3
I
12
Bases
1
1
I
Methanol
34 14
29 1
4 . .
I
r4
>
U
C OCA
‘I
. 1 - s CO
4-I
W I - .
bow
Jw
2-43
-------
‘I
a
I-
a
a
is
25
RETENTION TiME. mi.
31
45
FIG R 4 4
-------
MUTAGENICITY IN SALMONELLA OF HAZARDOUS WASTES
AND URINE FROM RATS FED THESE WASTES
David M. DeMarini, Research Genetic Toxicologist, Jefferson P.
Ininon, Genetic Toxicology Division, Jane Ellen Simmons, Ezra Berman,
Experimental Biology Division, Health Effects Research Laboratory,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina; Todd C. Pasley, Sarah H. Warren, and Ronald W. Williams,
Environmental Health Research and Testing, Inc., Research Triangle
Park, North Carolina
ABSTRACT
Fifteen hazardous industrial waste samples were evaluated for
mutagenicity in the Salmonella plate—incorporation assay using
strains TA98 and TA100 in the presence and absence of Aroclor
1254—induced rat liver S9. DicKloromethane/methanol extracts of the
crude wastes were also evaluated. Seven of the crude wastes were
mutagenic, but only 2 of the extracts of these 7 wastes were
mutagenic; extracts of 2 additional wastes also were mutagenic. In
addition, 10 of the crude wastes were administered by gavage to
F—344 rats, and 24—h urine samples were collected. Of the 10 raw
urines evaluated, 3 were mutagenic in strain TA98 in the presence of
S9 and —g1ucuronidase. The 3 crude wastes that produced these 3
mutagenic urines were, themselves, mutagenic. Adequate volumes of 6
of the 10 raw urines were available for extraction/concentration.
These 6 urines were incubated with —glucuron.idase and eluted
through Sep—Pak C 18 /columns; the methanol eluates of 3 of the urines
were mutagenic, and these were the same 3 whose raw urines also were
mutagenic. In general, the C 1 8/methanol extraction procedure
reduced the cytotoxicity and increased the mutagenic potency of the
urines. To our knowledge, this is the first report of the
mutagenicity of urine from rodents exposed to hazardous wastes.
Based on the present results, the use of only TA98 in the presence
of S9 might be adequate for general screening of hazardous wastes or
waste extracts for genotoxicity. The urinary mutagenesis assay does
not appear to be a useful adjunct tot he Salmonella assay for
screening hazardous wastes. The problems associated with chemically
fractionating diverse types of hazardous wastes for bioassay are
also discussed.
INTRODUCTION
The potential health effects of exposure to hazardous wastes may
include chromosomal damage and cancer (Maugh, 1979; Vianna and
Polan, 1984). Thus, knowledge of the mutagenic potential of
hazardous wastes would aid greatly in evaluating the hazardous
nature of and/or potential health effects from exposure to certain
wastes. Most studies on the genotoxicity of hazardous wastes have
used the Salmonella assay (Houk and Claxton, 1986; Nestmann et al.,
2-45
-------
1980); however, eukaryotic and/or mammalian cell assays been used to
a limited extent (Deflarini et al. 1984; Donnelly et al., 1985;
Bopke et al., 1982). Combinations of short—term genotoxicity assays
have been proposed as possible screens for hazardous wastes
(Barfknecht and Nalsmtth, 1984).
We have begun a series of studies to identify short—term assays and
chemical fractionation schemes that might be useful for screening
hazardous wastes for genotoxic activity. Our initial studies have
indicated that the use of mR.lIalian cell assays might not Improve
significantly the ability to detect the genotoxicity of hazardous
waBtes beyond that afforded by the Salmonella assay alone (DeNarini
et al., 1987a). because in vivo marmnalian metabolism may be a
critical factor in the generation of mutagenic metabolites from
complex hazardous wastes, we have investigated the utility of a
rodent urinary autagenicity assay as a possible adjunct to testing
the wastes themselves in the Salmonella assay.
Moat previous studies with rodent urinary mutagenicity a8says have
involved the use of pure compounds——frequently potent mutagens known
to require metabolic activation (Legator et al., 1982).
Consequently, the raw urine from such studies was mutagenic and did
not require concentration. However, B—glucuronidase was usually
required to observe rodent urinary autagenicity in these studies.
This is In contrast to studies of mutagenic activity of human urine
in which the urines must be concentrated (usually with XAD resins),
and p—glucuroaidase is usually not required (Legator et al., 1982).
We are aware of only 3 studies on the mutagenicity of urine from
rodents exposed to complex mixtures. Belisario et al. (1984, 1985)
demonstrated the mutagenicity of urine from rats exposed by gavage
to diesel emission particles (DEP), whereas Ong et al. (1985) did
not find mutagenic urine from rats exposed by inhalation to DEP
and/or coal dust. Concentration of the urine by XAD—2/acetone did
not reveal any mutagenicity in one study Cong et al., 1985), and It
reduced the mutagenic potency compared to raw urine in the other
8tudies (Bellsarlo et al., 1984, 1985). SolId—phase extraction
using Sephadex LB—20 resin or liquid extractions using chloroform,
toluene, or dichloromethane (DGI) also reduced the mutagenic potency
compared to raw urine (Belisario et al., 1985).
However, Bail et al. (1984) found that the mutagenic activity of
urine from rats administered 1—nitropyrene waj enhanced by
extracting and concentrating the urine on Sep—Pa1 ’ C 18 cartridges
eluted with methanol. Consequently, we have studied the
mutagenicity of urine from rats exposed by gavage to crude hazardous
wastes by testing both the raw urine as well as C18/methanol
concentrate of the urine in strain TA98 in the presence of Aroclor
1254—induced rat liver S9 and B—glucuronidase.
2-46
-------
Previous studies have shown that D M may extract mutagenic organics
from hazardous wastes (Andon et al., 1985; DeMarini et al., 1987a).
Thus, we have explored the applicability and utility of preparing
and bloassaying D M (as well as methanol) extracts of the 15
hazardous waste samples used in the present study. We selected
samples that were representative of wastes from diverse
manufacturing plants, including petrochemical, pharmaceutical, and
plastics manufacturing plants. We also tested combinations of
wastes from a variety of sources. Some of the samples were aqueous
wastes, some were organic wastes, and others were a mixture of both
types. Some wastes were from a single manufacturing process; others
were composite wastes from a variety of different manufacturing
processes. All of the wastes were liquid wastes, although many
contained solids. The fact that these 15 waste samples had been
partially chemically characterized (U.S. EPA, 1984), which is
information not normally available for hazardous wastes or other
complex mixtures, provided an additional reason for using these
wastes to explore bioassay procedures appropriate for hazardous
wastes.
MATERIALS ANT) METHODS
Wastes and Waste Extracts
Fifteen samples of hazardous industrial wastes were obtained from
Edward L. Katz, Hazardous Waste Engineering Research Laboratory,
U.S. Environmental Protection Agency, Cincinnati, OH (Table 1).
Three of the waste samples (A, B, and C) were from 3 different
manufacturing plants (petrochemical, pharmaceutical, and plastics).
The remaining samples were from 4 commercIal hazardous waste
Incineration facilities that burn a mixture of hazardous wastes from
a variety of industrial sources.
Table 1 shows the results of a partial chemical characterization
that was performed on these samples (U.S. EPA, 1984). For the
purposes of a previous study on the performance of hazardous was
Incinerators (U.S. EPA, 1984), 8 of the waste samples (B, E, F, G,
J, K, L, and M) were spiked with carbon tetrachloride and
trichioroethylene, which are not mutagenic in Salmonella (ICier et
al., 1986). Consequently, the concentrations of these 2 chemicals
In these 8 waste samples reflect this addition (Table 1). The
wastes were analyzed for the presence of a limited number of
priority organics and/or metals identified in the EPA Appendix VIII
list of priority pollutants (U.S. EPA, 1984). Thus, the chemical
characterizations shown in Table 1 should not be viewed as
Indicative of the overall chemical composition of the wastes. The
procedures used to extract the wastes and to evaluate the wastes and
waste extracts in the Salmonella plate—incorporation assay have been
described previously (DeMarini et al., 1987b).
2-47
-------
Urine and Urine Extracts
Ten of the 15 waste samples were evaluated for their ability to
produce mutagenic urine based on the chemical diversity that they
represented. We had planned to dose animals with 4 different doses
for 10 days before collecting 24—h urine from 3 rats per dose.
However, the available amounts of waste samples permitted the use of
this protocol for only waste C. For the other 9 waste samples, a
single dose of the crude waste samples was administered by gavage to
70—day—old male F—.344 rats. Doses were selected based on lethality
and the maximum volume that could be administered to the animals.
After dosing, rats were placed in metabolic cages (Nalgene), and
urines were collected on dry ice for 24 h from 2—3 rats receiving
the same dose of a particular waste. All of the urines were
centrifuged (10,000 rpm, 10 mm.), filter sterilized (0.45—urn
0
filter), and frozen at —20 C.
Portions of some of the urines were processed further as follows.
One ml of B—giucuronidase (Sigma Type VII from Escherichla coli ) at
concentration of 1000 units/mi of potassium phosphate buffer (0.15
H, pH 7.4) was added to 2—10 ml of thawed urine, and the mixtures
were incubated with shaking for 1 h at 370 C. Two serially
connected Sep—Pak C 18 cartridges (Waters Associates, Milford, MA)
containing 1 g of adsorbent were prepared for use by eluting 40 ml
of methanol followed by 50 ml of double—distilled water. Then, each
9 lucuronidase—urine incubate was poured through the columns,
followed by 40 ml of double—distilled water. Water was aspirated
from the columns until dry. Concentrates were then eiuted with 2,
4-mi. aliquots of methanol. The methanol concentrates were
evaporated under a stream of nitrogen and solvent exchanged into a
volume of DM 50 to produce 5—X concentrates. These were stored at
—20° C until they were bioassayed. Thus, there were two types of
urine for bioassay: (1) raw urine and (2) urine that had been
incubated with B—glucuronidase and extracted and concentrated 5X by
means of C18/methanol elution. The Salmonella plate—incorporation
assay was performed as described (Maron and Ames, 1983; DeMarini et
al., 1987b).
RESULTS
Of the 15 crude wastes evaluated for mutagenic activity in
Salmonella, 7 were mutagenic (Table 3). The mutagenic potencies
(rev/ug) ranged from 1 for waste G and 0, to over 1000 for waste A.
Four of the wastes (L, H, C, and 0) were positive only in strain
TA98, and all but waste G required S9. Three wastes (C, A, and F)
were positive only in strain TA100, and they were positive in the
presence and absence of S9. Dose—response curves for the mutagenic
wastes are shown in Fig. 1 (circles).
2-48
-------
Extract (Table 2) were prepared for all of the wastes except for
waste B, which reacted violently with methanol——leading to the loss
of the sample. Of the 14 waste extracts evaluated, 4 were
mutagenic, and their mutagenic potencies ranged from 2 for extract
H, to over 5000 for extract C (Table 3). Two extracts (H and J)
were mutagenic only in strain TA98; extract A was mutagenic only in
strain TA100; and extract C was mutagenic in both strains. Two
extracts (C and J) required S9, and 2 (H and A) were positive in the
presence and absence of S9. Dose—response curves for the mutagenic
extracts are shown in Fig. 1 (squares).
Raw urine from untreated control rats (up to 1 ml per plate) was not
mutagenic in strain TA98 plus S9 plus B—glucuronidase (Fig. 2). Of
the 10 wastes evaluated for urinary mutagenic 3 (C, L, and M) were
mutagenic (Table 4). Sufficient volumes of urine were available for
C18/methano l concentration for only 6 of the 10 urines studied. The
concentration procedure did not reveal mutagenic activity that was
not already evident from raw urine for these 6 urines. Thus, the 3
mutagenic raw urines, (C, L, and M) were also the only mutagenic
urine extracts (Table 4).
The C, 9 /methanol concentration procedure appeared to eliminate
cytotoxic components from the urines and to enhance the mutagenic
potencles of the mutagenic urines (Table 4, Fig. 2). In addition to
dose responses based on the amount of urine per plate, we used waste
C, which had been administered for 10 consecutive days, to generate
dose responses based on the amount of waste/rat (Fig. 2). A
qualitative summary of all of the data is shown in Table 5.
DISCUSSION
Mutagenicity Of Urine And Urine Extracts
Based on the wastes used here, the urinary mutagenesis assay does
not appear to provide any additional useful information beyond that
obtained from testing the wastes or waste extracts directly in the
Salmonella assay. Only 3 of the 10 wastes tested produced mutagenic
raw urine (Table 4), and all 3 of these wastes were mutagenic when
tested directly in Salmonella (Table 3). Combined with the
additional expense and time required to perform the urinary
mutagenesis assay, this assay does not appear to be a useful adjunct
to testing the wastes or waste extracts directly for mutagenicity.
Although the urinary mutagenesis assay appears to be of limited
usefulness in screening hazardous wastes, the data presented here
are, to our knowledge the first report of the mutagenicity of urine
from rodents fed hazardous wastes. The results raise the issue of
whether the mutagenic component(s) of the hazardous wastes were
excreted unmetabolized or were converted to mutagenic metabolites.
Table 3 shows that, in general, the 3 wastes that produced mutagenic
2-49
-------
urine (wastes C, L, and M or their extracts) required S9 in order to
be mutagenic in strain TA98. Our preliminary studies (data not
shown) indicate that —glucuronidase was required to detect
mutagenic activity in the urine or urine extracts. This is
consistent with findings for many pure compounds studied in rodent
urinary mutagenesis assays (Legator et al, 1982) and suggests that
certain components of the wastes were metabolized and conjugated to
glucuronide.
Fractionating and concentrating the urines by means of the
C 18 /methanol procedure did not identify a urine to be mutagenic that
was not identified as mutagenic from the studies with raw urine
(Table 4). Thus, this additional procedure did not enhance overall
detection capabilities. However, the Cl8/methanol concentration
procedure did appear to enhance the mutagenic potency of the three
mutagenic urines (Table 4, Fig. 2). The procedure either
concentrated mutagenic components of the urine and/or reduced the
concentration of cytotozic components in the urines, permitting
better detection of the mutagenic activity. The procedure did not
increase to the same extent the niutagenic potency of the three
urines. As shown in Fig. 2, it enhanced considerably the detection
of the mutagenicity of urine from rats fed waste L, but it did not
increase the potency of the urine from rats fed waste C (at least up
to 0.5 ml or ml equivalents/plate).
Mutagenicity of Wastes and Waste Extracts
The additional time and expense required to prepare the extracts did
not produce extracts that yielded much additional information that
was not obtainable from the crude wastes. Table 3 shows that 7
crude wastes were mutagenic, but for only 2 of theses (A and C) were
the extracts mutagenic. However, 2 wastes (H and J) were identified
as niutageaic based solely on the mutagenlcity of their extracts.
Thus, out of a total of 9 niutagenlc wastes, only 2 (22%) would have
been missed if only the crude wastes had been tested. It is
important to note that all wastes can be tested directly due to
microbial contamination or physical state, e.g., viscosity, pH,
etc. Thus, extraction/fractionation procedures will be necessary
for some, if not most, hazardous wastes in order to examine their
biological activity.
Design of Testing Schemes
Consideration also must be given to how complex a testing matrix
needs to be in order to have an acceptable detection capability and
still be cost—effective. For the 15 wastes tested, the matrix shown
in Table 3 requires the production of 120 dose—response curves, each
with at least 3 or 4 points, each in duplicate, and each repeated.
(inclusion of the urine assay for all 15 wastes would add an
additional 30 dose—response curves.) Although only 2 strains and 2
2-50
-------
activation conditions were used, would a reduced test matrix have
produced an acceptable level of detection for these waste samples?
Judicious selection of a test matrix (and test battery) are required
to screen hazardous wastes for biological activity in a
cost—effective manner without an unacceptable loss of detection
capability. Considering the effort required to produce a test
matrix as large as the one in Table 3, the use of only strain TA98 ÷
S9 would have reduced the time and expense considerably, with the
concomitant loss of only ‘•“ 20% of the detection capability provided
by the test matrix in Table 3. Consistent with this proposal is the
finding by the National Toxicology Program that the use of strain
TA98 or TA100 In the presence of S9 detects approximately 80% of the
mutagenic pure compounds In their data base that were identified
using 4 strains with and without S9 (Zeiger et al., 1985).
chemical Analysis and Biological Activity
Although extremely limited, the partial chemical characterization of
the waste sample (Table 1) is more extensive than would be available
for most waste samples——the vast majority of which would be of
completely unknown chemical composition. Because of the limitations
of the chemical analysis, a comprehensive picture of the composition
of each waste sample is lacking. Consequently, the chemistry does
not provide a ready explanation of the biological effects. For
example, a comparison of the mutagenicity of the aqueous to the
organic phase of the waste samples shows that the organic phase Is
more mutagenic than the aqueous phase for 3 out of 4 sets of waste
samples (Table 3). The wastes were analyzed for the presence of
only a few chemicals, representing a few chemical classes (organic
solvents, chlorinated organics, and metals). However, few of these
agents are mutagenic in Salmonella (Kier et al., 1986); thus, the
compounds and their concentrations shown in Table 1 do not predict
or even suggest the mutagenic activity exhibited by some of the
wastes.
This raises the Issue of the relevance of this type of chemical
analysis, which is expensive and time—consuming to obtain, for
predicting the potential mutagenicity (or other biological
activities) of hazardous wastes. In order to explore this problem
further, Simons et al., (1987) have examined the lethality and
hepatic toxicity of 10 of these waste samples in the rat. The
presence of known hepatotoxins in these wastes suggested that some
of the wastes might be hepatotoxic. However, the chemistry did not
predict the lethality of the waste samples.
In our search of a suitable short—term test to either complement or
replace the Salmonella assay for screening hazardous wastes, we also
have examined the ability of these wastes to Induce bacteriophage
lambda (Houk and DeMarini, 1987). The results indicated that this
2-51
-------
assay was more sensitive than the Salmonella assay, and it was
comparable to the Salmonella assay in terms of time and expense.
Current schemes used by the U.S. EPA and other agencies to identify
wastes as hazardous rely primarily on physical characteristics and
chemical composition of the wastes (Friedman, 1985). Greer (1984)
has discussed the limitations of the current definition of hazardous
waste, and the U.S. EPA has published possible guidelines to correct
this situation by adding health effects, including mutagenicity
data, to the evaluation of hazardous wastes (Federal Register, 1983;
1984a, b).
Both government (Federal Register, 1983; 1984 a, b) and industry
(Barfknecht and Naismith, 1984; Guiney, 1985) have recognized the
important role that short—term tests could play in the toxicological
assessment of hazardous wastes. This concensus has prompted us to
investigate the feasibility of such a proposal. Based on our
present results, short—term bioassays may offer an important insight
Into the potential health effects of wastes and, thus, may aid in
the Identifications of some wastes as hazardous.
REF ENCES
Andon, B., N. Jackson, V. Houk and L. Claxton (1985) The evaluation
of chemical and biological methods for the identification of
mutagenic an cytotoxic waste samples, in: J. K. Petros, W. J.
Lacy and R.A. Conway (Eds.) Hazardous Industrial Solid Waste
Testing, ASTMP Pubi. code No. 04—886000—16, Philadelphia, pp
204—215.
Ball, L. N., M.J. Kohan, J.P. Inmon, L.D. Claxton and J. Lewtas
(1984) Metabolism of l—nitrol 1 - 4 capyrene in vivo In the rat and
inutagenicity of urinary metabolites, Carcinogenesis, e, 1557—1564.
Barfknecht, T.R., and R.W. Naisinith (1984) Methodology for evaluating
the genotoxicity of hazardous environmental samples, Hazard.
Waste, 1, 93—109.
Belisarto, M.A., V. Buonocore, E. DeMarinis and F. De Lorenzo (1984)
Biological availability of mutagenic compounds adsorbed onto
diesel exhaust particulate, Mutation Res, 135, 1—9.
Belisarlo, M.A., C. Farina and V. Buonocore (1985) Evaluation of
concentration procedures of mutagenic metabolites from urine of
diesel particulate—treated rats, Toxicol. Lett., 25, 81—88.
DeMarini, D.M., P.A. Brimer and A. W. Hsie (1984) Cytotoxicity and
mutagenicity of coal oils in the CHO/HGPRT assay, Environ.
Mutagen., 6, 517—527.
2-52
-------
DeMarini, D.M., D. J. Brusick and J. Lewtas (1987a) Use of limited
protocols to evaluate the genotoxicity of hazardous wastes in
mammalian cell assays: Comparison to Salmonella, J. Toxicol.
Environ. Health (in press).
DeMarini, D.M., J.P. Inmon, J.E. Simmons, E. Berman, T.C. Pasley,
S.H. Warren and R.W. Williams (l987b) Nutagenicity in Salmonella
of hazardous wastes and urine from rats fed these wastes, Mutat.
Res. (in press).
Donnelly, K. A., J.W. Brown, J.C. Thomas and P. Davol (1985)
Evaluation of the hazardous characteristics of two petroleum
wastes, Hazard. Waste Hazard. Mater., 2, 191—208.
Federal Register (1983) Notification requirements; reportable
quantity adjustments, Vol. 48, No. 102, May 25, pp. 23552—23602.
Federal Register (194a) Hazardous waste management systems, Vol. 49,
No. 32, February 15, pp. 5854—5859.
Federal Register (1984b) Proposed deadlines for exposure assessment,
Vol. 49, No. 227, November 23, PP. 46304—46312.
Friedman, D. 1985) An overview of selected EPA RCRA test method
development and evaluation activities, in: J. K. Petros, W. J.
Lacy and R.A. Conway (Eds.) Hazardous Industrial Solid Waste
Testing, ASTMP Pubi. Code No. 04—886000—16, Philadelphia,
pp. 77—84.
Greer, LE. (1984) Definition of hazardous waste, Hazard. Waste, 1,
309—322.
Guiney, P.D. (1985) Use of predictive toxicology methods to estimate
relative risk of complex chemical waste mixtures, Hazard. Waste
Hazard. Mater., 2, 177—189.
Hopke, P.K., M. J. Plewa, J. B. Johnston, D. Weaver, S.G. Wood, R.A.
Larson and T. Hinesly (1982) Multitechnique screening of Chicago
municipal sewage sludge for niutagenic activity, Environ. Sd.
Technol., 16, 140—147.
Houk, V.S. and L. D. Claxton (1986) Screening complex hazardous
wastes for mutagenic activity using a modified version of the
TLC/Salmonella assay, Mutation Res., 169, 81—92.
Houk, V.S.and D. M. DeMarini (1987) Use of the Microscreen
phage—induction assay to assess the genotoxicity of 14 hazardous
industrial wastes, (submitted).
ICier, L. D., D.J. Brusick, A.E. Auletta, E.S. Von Halle, MM. Brown,
V.F. Simmon, V. Dunkel, J. McCann, K. Mortelmans, M. Prival, T.L.
2—53
-------
Rao, and V. Ray (1986) The Salmonella typhimurium/maminallan
microsomal assay. A report of the U.S. Environmental Protection
Agency Cene—Tox Program, Mutation Res., 168, 69—240.
Legator, M.S., E. Bueding, R. Batzinger, T.H. Connor, E. Eisenstadt,
M.C. Farrow, C. Ficsor, A. lisle, J. Seed and R.S. Stafford (1982)
An evaluation of the host-mediated assay and body fluid analysis,
Mutation R.es., 98, 319—374.
Maron, D.M. and B.N. Ames (1983) Revised methods for the Salmonella
mutagenicity test, Mutation Res., 113, 173—215.
Maugh, T.H. (1979) ToxIc waste disposal a growing problem, Science,
204, 819—823.
Nestinann, E.R., E.G.H. Lee, T.I. Matula, G.R. Douglas and J.C.
Mueller (1980) Mutagenlcity of constituents identified in pulp
and paper mill effluents using the Salmonella/niammalian-inlerosome
assay, Mutation Res., 79, 203—212.
Ong, T., W.—Z Whong, J. Xu, B. Burchell, F.H.Y. Green and T. Lewis
(1985) Cenotoxlcity studies of rodents exposed to coal dust and
diesel emission particulates, Environ. Res., 37, 399—409.
Simmons, J.E., D.M. DeMarini and E. Berman (1987) Hepatotoxic and
lethal effects of hazardous industrial wastes (submitted).
U.S. EPA (1984) Performance Evaluation of Full—Scale Incinerators,
National Technical Information Center Pubi. No. PB85—129500.
Vianna, N.J. and A. K. Polan (1984) Incidence of low birth weight
among Love Canal residents, Science, 226, 1217—1219.
Zeiger, E., K.J. Risko and B.H. Margolin (1985) Strategies to reduce
the cost of mutagenicity screening with the Salmonella assay,
Environ. Mutagen., 7, 901—911.
2—54
-------
YIIBLE 1
Concentration of Chemicals and Metals Identified in Hazardous Wastes ( g/g)
Chenh,caI/Meta a
Petro-
Chemical
Pharma-
ceutical
Plastics
Hazardous waste Incineration Facilities
1
2
3
4
Aqueous
Organic
Organic
Aqueous
Organic
Organic
Aqueous
A
B
C
D
E
F
C
H
I
J
K
M
N
0
Aniline 14000 550000
Benzyl Chloride 3000
Bis-(2-ethy l- 500 <100 3800 200 230 <10 <10
hexyll-pflthalate
Butylbenzyl-phthalate <100 320 120 450 <5 8 160 140
Chloraane <60 <60 19000 19000
Chlorophenvlisocyanate 21000
Cresol(s) 2000 2500
Diet hvlphtha late 620 1300 240 240
m-D ichlorobenzene 23000
o-Dichlorobenzefle 46000
p-Dichlorobenzene 59000
m-DinitrObenzene <100
Oipflenylamune 6200
2.4-D imethy lphenol 500 2000
Hexachlorobutaduene <10 <10 <10 <10
Hexacflloroettuane 560
HexachiOrocyclo- <10 <10 230 260
pentadiene
Isophorone <100 110
Mononitrobenzene <100
Naphthalene <100 <100 350 250 450 44 49 450 490 38 33 <10 <10
Phenol 34000 1000 1500 1700 1800 2900
Phenyleneduamine 2300
PhenylisOcyanate 160000
Trans-1,4 -dichloro- 59000
2-butene
1.2.4-Trichlorobenzene 290
Benzene <3 260 46000 58000 <10 <10
Carbon tetrachloride 68000 44000 <2 6000 3700 4400 <10 <10 5900 5700 9100 11300 <10 <10
Chlorobenzene 4100 390 500 <10 <10
Chlorometruane 1200
Chloroform 2900 170 270 21 22 110 60 <10 <10
CiS-1,4 -diChloro- 18000
2-butene
Methylene Bromide <10 <10 3100 4700
Methylene Chloride 21000 100 340 44 60
Methyl Ethyl Ketone 27 9700 18000 38000
Tetrachloroethylene 11000 <1 28 7100 9800 1600 1300 7900 8100 87 150 <10 <10
Toluene 240000 110 2400 32000 45300 2900 2700 56000 48000 160000 390000 22 20
1,1,1-Trich loroethane <100 24000 16000 330 230 <10 <10
Trichloroethylene 4000 40000 <1 5500 3700 4400 90 85 8100 7800 8300 10300 <10 <10
Ag <3 <1 <1 <1 <1 <1 <1
AS <24 <20 <20 <20 <20 <14 <14 <23 <23
Ba <7 110 140 6 7 1160 1150 990 1100
Be <2 <1 <1 <1 <1 <1 <1 <1 <1
Cd <5 6 6 <1 <1 153 15 49 55
Cr <5 50 57 3 3 431 425 250 290
Hg <22 <10 <10 <10 <10 <4 <4 <50 <50
Ni <68 7 7 2 2 26 27 <4 <4
Pb <19 140 150 <10 11 1830 1800 1200 1300
Sb <12 58 61 <10 <10 437 373 <24 <24
Se <470 340 <100 <100 <100 <21 <21 <160 <160
TI <23 <20 <20 <20 <20 <9 <9 <22 <22
Si <1 <1
2 8 7 7 6 7 7 7 7
Water l ) 2 S 94 3 38 48 95 67 21 23 5 5 94
aData from EPA l1984. Waste samples 8. 8. F, C, i, K. L and M were spiked With carbon tetrachloride and trichloroethylene; therefore, the concentrations of these 2
chemicals in these wastes reflect this addition.
2-55
-------
TABLE 2
SOLVENT EXTRACTIONS OF HAZARDOUS WASTES
Phase
Waste DCM
separationa
Final
(m
volumeb
I)
Methanol
A No No 4
B No C
C No No 5
D Yes 2
E No No 2
F No Yes 2.5
G Yes 2.5
H Nod Yes 2
Yes 2
J No l’4o 5
K No Yes 2
L No No 7
M No No 5
N Yes 2
0 Yes 2
aTen ml of each waste was subjected to
extraction as described in Materials
and Methods.
bFinal volumes vary because variable amounts
of DMSO were required to reconstitute each sample
due to the different sofubilitles of the samples.
cE, ,, 1ve reaction occurred upon the
addition of methanol; no extract was
available for bioassay.
dphase separation occurred when both
solvents were added; DCM phase was
used for bioassay.
2—56
-------
TABLE 3
MUTAGENIC POTENCIES OF WASTES AND WASTE EXTRACTS IN SALMONELLA
Revertants
per 11 ga
Crude
Wastes
Waste
Extracts
TA98
TA100
1A98
TA100
Waste
+S9
—39
+S9
—S9
÷S9
—S9
+S9
—S9
A
1870
1326
79
37
B
NTb
NT
NT
NT
C
18
329
5581
168
F
3
4
.
G
1
1
H
6
2
J
3177
L
58
M
31
0
1
aValues are the model slopes calculated from the dose—response curves as
described in the Materials and Methods. Blank areas represent nonmutagenic
responses. Wastes or extracts of samples D, E, I, K, and N were not mutagenic.
The average DMSO control values (rev/plate ± S.D.) were: 52 ± B (TA9B + S9),
33 ± 5 (TA98 — S9), 156 ± 19 (TA100 + S9), 151 ± 12 (TA100 — S9).
bNT, not tested.
2-57
-------
TABLE 4
MUTAGEN1C POTENCIES OF URINES AND
URINE EXTRACTS IN STRAIN TA98
Amount
of
Revertants per ml
waste per
Waste rat (glkg)
or
ml equivalents 8
Raw
extracts
C 0.2 395 455
L 2.5 259 1586
M 5.0 63 205
G 5.0
0 5.0
E 5.0
H 5.0 NT
J 2.5 NT
K 2.5 NT
B 5.0 NT
°Values were calculated from the linear
portion of the dose—response curves as
described in the Materials and Methods.
Urines were tested in the presence of
S9 end —glucuronidase. Blank areas
represent nonniutagenic responses. The
average control (no urine) value ± S.D.
was 41 ± 6.
2-58
-------
TABLE 5
SUMMARY OF RESPONSES
Wastesa
Waste Crude extracts
Urines
Raw extracts
C + ÷ + +
L + + +
M ÷ + +
G ÷
0 +
E
H + NT
J + NT
K NT
B NTb NT
A + + NT NT
F + NT NT
D NT NT
I NT NT
N NT NT
8 Four responses (2 strains +1— S9) were merged
into one response for both the crude wastes and
the waste extracts (a total of 8 responses)
using the following criterion: If a positive
response was obtained under at least 1 of the 4
test conditions, the summary response was
positive.
bNT, not tested.
2-59
-------
LEGENDS TO FIGURES
Fig. 1. Dose—response curves of crude wastes (circles) and waste extracts
(squares) in the presence (open markers) or absence (solid markers)
of Aroclor 1254—induced rat liver S9. Experiments were performed
as described in the Materials and Methods. Results are from
representative experiments.
Fig. 2. Dose—response curves from strain TA98 of raw urines or urine extracts
from F—344 rats administered crude wastes. Dosing of animals, preparation
of urines and extracts and bioassay protocols are described in Materials
and Methods. Waste C urines were collected after 10 days of dosing;
waste L and M urines were collected after a single dose performed
as described in Materials and Methods. Results are from representative
experiments.
2-60
-------
Di
CD
I
LI)
4 )
wO 2 4 6 8 10
150
J TA98
100
50
250
100
‘0 0.01 0.02 0.03 0.04 0.05
F ,)
TA IOO
0.1 0:2 0:3 0:4 0:5
50
2 4 6 8 10
Dose
(pg/plate)
800
00
400
200
150
5C
400
300
200
100
p.
C TAIOO
F TAIOO
CD
4 )
C-
c ii
>
Di
1 1
150
100
H TA98
2—61
-------
Urine/C
U-)
-p
(0
0
C D
-p
N) C
(0
-p
U-)
>
U)
150
400
Dose (ml
or ml equivale
nts/p late)
250
200
0.2 g/kg
0.1 g/kg
100
0.6
0.8
I
Ur in /L
2.5 g/kg
600
C—18/ur me
Urine/H
5
g/kg
200
200
I
Raw Urine
C—lB/urine
150
0.2 0.4
0 0.2 0.4 0.6
—F
-------
APPLICATION OF A SIMPLE SHORT—TERN BIOASSAY FOR THE
IDENTIFICATION OF GENOTOXINS FROM HAZARDOUS WASTES
Shahbeg S. Sandhu, Research Biologist, Genetic Toxicology Division,
U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina
ABSTRACT
The proper disposal of hazardous wastes currently generated and the
clean up of the waste disposal sites of the past is a challenge
facing regulatory agencies in the industrialized nations. The
estimation of the levels of toxicity is an essential step in
prioritizing the industrial effluents and solid wastes for treatment
and disposal. A number of short—term bioassays has been developed
to supplement information from chemical analysis for evaluating the
potential of chemical complex mixtures to induce adverse human
health effects and environmental contamination. Among these
bioasaays, plant test systems provide simple, inexpensive, and rapid
means oO evaluating the toxic effects of industrial wastes based on
multimedia exposure. Two such assays, tradescantia thaliana and
Zea maya , have been used for monitoring the genotoxic effects of
ambient air, municipal wastes, industrial effluents, solid wastes,
water sediments, and pesticides. We applied the Arabidopsis embryo
assay to evaluate the mutagenicity of complex environmental mixtures
including coal and wood combustion extracts (from the People’s
Republic of Qilna), industrial effluents, and sludges. The coal and
wood smoke condensate samples were extracted with methylene chloride
and solvent exchanged to DMSO. The industrial waste samples were
tested either unextracted or as dichioromethane extracts. The
comparative analysis of the data on these samples in Arabidopsis and
in other commonly used bloassays will be presented. The
significance of short—term plant bloassays for use in environmental
assessment will be discussed.*
INTRODUCTION
Approximately 200 short—term in vitro and in vivo test systems (STT)
are currently available to evaluate the geziotoxic potential of
environmental chemicals (Waters et al., 1984). These bloassays were
developed in the early 70’s after a discovery by Dr. Bruce Ames
(1973) of the University of California, Berkeley, CA, that most of
the carcinogens can be identif led by using simple microbial
mutagenicity bioassays. Most of these STT employ either in vitro
procaryotic or eucaryotic cells or in vivo animal test systems.
Substantial progress has been made in utilizing these assays for
*Thjs is an abstract of a proposed presentation and does not
necessarily reflect EPA policy.
2-63
-------
identifying mutagens and carcinogens in a variety of products such
as agricultural chemicals, drugs, consumer products, industrial
chemicals, etc. Since humans are rarely exposed to individual
chemicals, but exposure occurs mostly to a mixture of chemicals, the
STT technology has been applied for evaluating the genotoxicity of
complex chemical mixtures such as industrial emissions, effluents,
and hazardous waste samples, and for evaluating the level of
mutagenicity in the ambient environment.
One of the STT that appears very useful for identifying mutagens and
presumptive carcinogens in the environment is the Arabidopsis embryo
assay (AEA). Originally developed by Andreas Muller (1963, 1965),
this assay has been used to assess the mutagenicity of over 120
chemicals (Gichner and Veleminsky, 1967; Redei et al., 1980;
Velealnaky and Gichner, 1968; Gichner et al., 1982), and complex
mixtures (Sandhu and A.cedo, 1986; Acedo et a?., 1987). One of the
easily detectable genetic endpoints in the AEA is the loss of
chlorophyll pigmentation in the developing embryos of the mutant
population. As an in vivo bioassay, Arabidopsis offers certain
unique advantages which include: (1) short study period (4 weeks),
(2) economy of space (200 plants can be grown in a petri dish), (3)
high fecundity, (4) easy- to grow and score for mutation, (5)
multiple genetic endpoints, and (6) induced mutations are observed
in the progeny of the exposed population. Although data are
limited, it appears that this small crucifer has the capability to
biotransform promutagens into ultimate mutagens (Redel et al., 1984).
Since the usefulness of AEA for evaluating the mutagenicity of
individual compounds has been established, we have attempted to
extend its utility for assessing the genotoxic hazard of complex
environmental mixtures. In the study reported here we have applied
AEA for analysis of the genotoxicity of five hazardous waste samples
obtained from a phenyl mercuric acetate plant.
MAT UALS AND METHODS
Hazardous Waste Samples
Two liquid effluent and three solid waste samples from a mercuric
acetate plant were received through Versar, Inc., New Jersey. The
liquid samples were Cosan Scrubber water and Cosan Separated water.
The Cosan Scrubber water sample was very alkaline (pH 11.0), whereas
the Cosan Separated water sample was extremely acidic. The three
solid samples were identif led as Nouder Preascake No. 2 (NP2),
Noudex Presacake No. 3 (NP3), and Noudez—Hg (N—I g) recovery sludge.
These samples were evaluated for their genotoxic effects either as
crude samples or after extraction with methylene chloride or water.
The pH of the liquid samples was adjusted to 7.0 for Cosan Scrubber
and to 5.0 for Cosan Separated to allow the growth of Arabidopsis.
Concentrations of the sample ranging for 0.1—1,000 ag/mi were made
2-64
-------
with buffered mineral medium (Redei, 1965). A total volume of 1 ml
of the test solution contained in scintillation vials was used for
treatment.
Sample Preparation
The solid samples were tested in their crude form, as received, and
as aqueous or methylene chloride extracts. For testing the crude
samples, seeds were exposed by planting them directly in mixtures
containing various proportions of the test material and ProinixlM
medium (Premier Brands, New Rocheile, NY 10804). The mixtures were
prepared by kneading the desired amount of PromixTM and sample
together In a plastic bag and putting the mixture in petri dishes
where the Arabidopsis is grown.
Aqueous extracts of the solid samples were obtained by preparing 1:1
(v/v) mixtures of sample and deionized water. These mixtures were
allowed to stand for 1 h, centrifuged at 4,500 rpm for 10 win, and
the supernatants were collected. Dilutions of these extracts were
tested for evaluating the the genotoxicity of these samples.
Methylene chloride extracts were solvent exchanged to DMSO to a
concentration equivalent to 1 mg/mi, and further dilutions ranging
from 0.01—0.50 mg/mi were made and evaluated for their genotoxicity.
Arabidopsis Assay Protocol
For testing liquid samples and sample extracts, approximately 300
seeds of Arabidopsis thaliana (Columbia wild type), contained in a
small cloth bag, were immersed in each vial of test solution for
about 12—15 h at room temperature. At the end of the treatment
period, the seed bags were rinsed under running tap water for 1 h.
The washed seeds were dispersed in water and planted In glass Petri.
dishes (85x75 mm) containing pasteurized Promix medium. The
dishes were subsequently placed In a growth chamber set to provide a
16—h photoperiod and a temperature of 24 C.
Crude samples (nonextracted solid samples) were tested by growing
the Arabidopsis seeds in Petri dishes containing sampie/PromiP
mixtures, then placing the dishes in the growth chamber.
After 3 weeks of growth (Fig. 1), when the seeds were mature but not
yet brown, three fruits from each of the 100 plants selected at
random for each concentration were opened with sharp forceps. The
embryos within these fruits were scored for white (mutant) embryos
among the normal green embryos (Fig. 2). The extent of sterility
Induced by each concentration was also determined. Sterility is
manifested as empty spaces within a fruit or In severe cases no
fruits are formed at all along the stem.
2-65
-------
Each experiment contained a concurrent positive control (0.2Z EMS)
and a solvent negative control. The data presented are an average
of the two experiments except for the methylene chloride extracts
for which the results of a single experiment are given. Confidence
intervals (CI) were determined using the statistical methods of
Ehrenberg (1977).
In evaluating the experimental data, a sample was considered
mutagenic to Arabidopais if it induced a twofold or greater increase
in the nt sber of chlorophyll mutants at two or more doses as
compared to solvent control or if it induced dose—related increases
at three or more doses.
RESULTS
The data on the mutagenicity of the liquid samples, Cosan Scrubber
water and Cosan Separated water, presented in Table 1 indicate that
both samples are autagenic for Arabidopsis. A slight dose response
in mutagenicity is indicated by Cosan Scrubber water, and although
Cosan Separated water shows a similar trend, there was no further
increase in mutagenicity beyond the concentration of 1 mg/mi
(Pig. 3). In addition to autagenicity, a high degree of sterility
was induced by 10 ag/mi of Cosan Scrubber water and the undiluted
(1,000 ag/at) Cosan Separated water (Table 1). Germination of the
treated seeds was considerably reduced at higher concentrations.
These results must, however, be viewed with caution since the pH of
both samples had to be drastically adjusted prior to testing to
allow growth of Arabidopsis. The alkaline Cosan Scrubber sample was
adjusted using whole pellets of ROll. Although these pH changes were
necessary to allow testing, it is quite probably that some
components of these complex mixtures were substantially altered in
the process. The nature of these alterations and their effect on
the overall autagenicity of the samples is not known.
The solid waste samples, Noudex NP2, NP3, and N—Hg recovery sludge
were tested as unextracted crude samples as well as after extraction
with methylene chloride and water. The toxicity range—finding
experiments showed that the unextracted crude samples were very
toxic for Arabidopais. Therefore, these materials were tested for
mutagenicity after various dilutions with Promix medium. The
mutagenicity of these samples for Arabidopsis is shown in Table 2.
NP2 was very toxic to Arabidopsis. Even after mixing 1 part of the
sample with 9,000 parts of ProaixIthe survival of the treated plants
was only 50X. At this low concentration, NP2, however, induced
about four times more chlorophyll mutations than the solvent control
(Table 2). It was interesting to note the relatively low degree of
sterility (7.fl) induced in the surviving plants, indicating that
the observed toxicity may be primarily due to inhibition of growth
rather than interference with genetic mechanisms. Proportions of
sample lower than 1:9,000 were not tested due to difficulties
2-66
-------
encountered in accurately measuring and mixing with Promix such a
small amount of the sample.
NP3 and N—Hg recovery sludge were relatively less toxic to
Arabidopsis as compared to NP2. Nevertheless, both samples were
mutagenic for Arabidopsis (Table 2).
The results of the treatment of Arabidopsis with the samples
extracted with inethylene chloride are shown in Table 3. All three
samples were positive for mutagenicity; however, dose—related
response was not observed for these extracts. For lack of sample
availability, testing at doses lower than 0.1 mg/mi and a repeat
experiment could not be performed.
Table 4 shows the mutagenicity for Arabidopsis after treatment with
the aqueous extracts from each of the three solid samples. NP3 and
N—Hg recovery sludge samples showed dose—related mutagenic response
although the mutagenicity values are much lower than that of NP2. A
high degree of sterility was induced by N—Hg especially at 100Z
concentration of the extract.
DISCUSSION
The identification and proper disposal of hazardous wastes are
challenges facing regulatory agencies In the Industrialized
nations. At present, most of the regulatory agencies, Including the
U.S. EPA, use chemical components of the Industrial wastes as a
basis for hazard evaluation (Friedman, 1985). However, chemical
analysis is currently limited to the detection of only a few known
toxic compounds, whereas numerous other potentially harmful
genotoxins may be undetected. Even If all the chemicals present in
hazardous wastes were Identified (which is an improbable task),
chemical analysis alone will not provide Information on the
synergistic and antagonistic effect that may reBult from the
interaction of these complex chemical mixtures. Therefore a
biological evaluation along with their chemical analysis has been
repeatedly emphasized for assessing the health hazards of complex
chemical mixtures (Waters et al., 1978; Sandhu et al., 1987).
A large assortment of STT may be utilized for the evaluation of
hazardous industrial wastes, and plant bioassays are uniquely
suitable for this purpose. Certain bloassays such as Zea mays
(Plewa, 1984), Tradescantia micronucleus assay (Ma et al., 1983),
and Tradescantla gene mutation assay (Schairer et al., 1983) have
been very useful for evaluating the levels of genotoxins in the
terrestrial and aquatic environment (see review by Ma and Harris,
1985). The primary advantages of the plant systems are their
simplicity, cost effectiveness, and their utility in assessment of
simultaneous multimedia exposure. In contrast to cell culture
bioassay, one does not have to worry about microbial contamination.
2—67
-------
The purpose of analyzing the hazardous waste is twofold: (1) to
assess its overall health hazards for its proper disposal and (2) to
identify the specific components of the waste responsible for the
particular effects so that control technology can be applied to
minimize the production of toxic chemicals. In the study reported
here, we have applied a simple bioassay in an attempt to accomplish
the first objective. Our choice of test samples, obtained from a
phenyl mercuric acetate plant, was based on their availability. The
data presented in this study are based only on results from crude
samples and on methylene chloride and aqueous extracts. Although
extraction reduced cellular toxicity, it did not alter the
qualitative mutagenicity results of the samples. Therefore, if
samples are not very toxic, testing complex mixtures directly
without elaborate sample preparations may be a convenient way to
make a preliminary assessment of their genotoxic potential.
REFERINCES
Acedo, C. N., S. S. Sandhu, D. M. DeMarini, and J. L. Mumford (1987)
Utility of Arabidopsis embryo assay for testing complex mixtures,
Environ. Mutagen., 9 (Suppi. 8), 2.
Ehrenberg, L. (1977) Aspects of statistical inference in testing for
genetic toxicity, in: B. J. Kilbey, M. Legator, W. Nichols, and
C. Ramel (Ms.), Handbook of Mutagenicity Testing, Elsevier, New
York, pp. 420—459.
Ames, B. N., W. E. Durston, E. Yamasaki, and F. D. Lee (1973)
Carcinogens are mutagens: A simple test system combining liver
homogenates for activation and bacteria for detection, Proc.
Nati. Acad. Sd. USA 70:2281—2285.
Friedman, D. (1985) An overview of selected EPA RCRA test method
development and evaluation activitieB, in: J. K. Petros, W. J.
Lacy, and R.. A. Conway (Eds), Hazardous and Industrial Solid
Waste Testing: Fourth Symposium, American Society for Testing
and Materials, Philadelphia, PA pp. 77—84.
Gichner, T., and J. Velemlnsky (1967) The mutagenic activity of
1—alkyl—1—ni t rosoureas and alkyl —3—nit ro —1—nit ro soguanidines,
Mutat. Res., 4, 207-212.
Gichner, T., J. Veleminaky, and K. Pankova (1982) Differential
response to three alkylating nitrosocompounds and three
agricultural chemicals in the Salmonella (Ames) and in the
Tradescantia, Arabidopsis and barley mutagenicity assays, Biol.
Zbl. 101, 375—383.
Ma, T.—L, and M. M. Harris (1985) In situ monitoring of
environmental mutagens, Hazard Assess. them., 4, 77—105.
2-68
-------
Ma, T.—H., W. R. Lower, F. D. Harris, J. Poker, V. A. Anderson, N.
M. Harris, and J. L. Bare (1983) Evaluation by the
Tradescantia —micronucleus test of the outagenicity of internal
combustion engine exhaust fumes from diesel and diesel—soybean
oil mixed fuels, in: N. Waters, S. Sandhu, .3. Lewtas, L.
Claxton, N. Chernoff, and S. Nesnow (Eds.), Short—Term Bloassays
in the Analysis of Complex Environmental Mixtures III, Plenum
Press, New York, pp. 89—99.
Muller, A. J. (1963) nbryonentest zum Nachweis rezessiver
Sefalfaktoren bel Arabidopsis thaliana , Biol. Zbl., 83, 133—163.
Muller, A. .3. (1965) A survey of agents tested with regard to their
ability to induce recessive lethals in Arabidopsis, Arabidopsis
Inf. Serv., 2, 22—24.
Plewa, M. J. (1984) Plant genetic assays to evaluate complex
environmental mixtures In: M. D. Waters, S. S. Sandhu,
.3. Lewtas, L. Claxton, G. Strauss, and S. Nesnow (Eds.),
Short—Term Bloassays in the Analysis of Complex Environmental
Mixtures IV, Plenum Press, New York, pp. 45—64.
Redei, C. P. (1965) Genetic blocks in the thiamine synthesis of the
angiosperm Arabidopsis, Amer. .3. Bot., 52, 834—841.
Redel, C. P., C. N. Acedo, and S. S. Sandhu (1984) Sensitivity,
specificity, and accuracy of the Arabidopsis assay in the
identification of carcinogens, in: E. H. Y. thu, and
W. M. Ceneroso (Ms.), Mutation, Cancer and Malformation, Plenum
Press, New York, pp. 689—708.
Redei, C. P., M. N. Redei, W. R. Lower, and S. S. Sandhu (1980)
Identification of carcinogens by mutagenicity for Arabidopsis
testing of complex mixtures in Arabidopsis, Environ. Mutagen.,
8(Suppl. 6), 72.
Sandhu, S. S., D. M. DeMarini, M. Mass, M. M. Moore, and
.3. L. Mumford (Ms.) (1987) Short—Term Bioassays in the Analysis
of Complex Environmental Mixtures V, Plenum Press, New York, in
press.
Schairer, L. A., R. C. Sautkulis, and N. R. Tempel (1983)
Mutagenicity of smog and diesel emissions implies that UV and/or
visible light are activating agents, Environ. Mutagen., 5, 466.
Veleminsky, .3., and T. Gichner (1986) The mutagenic activity of
nitrosamines in Arabidopsis thaliana , Mutat. Res., 5, 429—431.
Waters, M. D., S. Nesnow, J. L. Huising, S. S. Sandbu, and
L. Claxton (Ms.) (1978) ApplicatIon of Short—Term Bioassays in
the Fractionation and Analysis of Complex Environmental Mixtures,
Plenum Press, New York, 588 pp.
2-69
-------
Waters, N. D., S. S. Sandhu, J. Lewtas, L. Claxton, C. Strauss, and
S. Nesnow (Ms.) (1984) Short—Ters Bloassays in the Analysis of
Co.plez Envirozental )IixtUres IV, Plenum Press, New York, 384 pp.
Figure 1.—A dish containing 200—250 Arabidopsis thaliana plants 3
weeks after planting.
Figure 2.—A Arabidopsia thaliana plant with segregating white and
green embryos.
Figure 3. —Mutagenictty induced in Arabidopsis by the Cosan liquid
samples.
2-70
-------
Figure
1
2-71
-------
- 1 -
: 4 !
Ptw i Pr
‘7
;‘- $ . ‘$J 4 ‘ - ? -
14. , -‘ p . ____
____
- - —
Figure 2
2-72
-------
0.01 0. 1 1.0 10 100 1000
CONCENTRATION. mg/mi
Figure 3. Mutagenicity induced in Arabidopsis by the Cosan liquid samples.
0.4 —
0.3 —
( I
0 COSAN SCRUBBER WATER
COSAN SEPARATED WATER
>-
2
L i i
a
w
U-
I-
2
I-
0.1 —
I I
2-73
-------
Table 1. Mutagenicity of Cosan Liquid Samples in the Arabidopsis Embryo
As say
Sample
Concentration
(mg/mi)
Germination
(Z)
% of Fruits
Containing
Mutants ± CI
Sterility
(2)
Cosan Scrubber
0.1
95.5
15.5 ± 4.94
1.5
Water
1.0
10.0
100.0
1000.0
97.5
99.0
73.0
50.0
19.5 ± 8.90
20.8 ± 13.77
23.6 ± 7.75
32.0 ± 4.76
8.5
14.0
2.0
3.5
Cosan Separated
0.1
97.5
18.0 ± 1.96
4.5
Water
1.0
10.0
100.0
1000.0
89.0
96.0
55.0
52.0
22.6 ± 8.64
24.4 ± 13.95
28.9 ± .16.60
28.6 ± 9.29
7.8
4.6
2.0
40.0
Solvent Control
0
100.0
4.0 ± 2.78
Ô
(mineral medium)
Positive Control
0.20
97.0
80.0 ± 1.96
64.0
(E? )
2-74
-------
Table 2. Mutagenicity of Three Solid Waste Samples in Arabidopsis
Proportion
% of Fruits
Sample
(Sample:
T I !
Promix
v/v)
Germination
(/ 0)
Containing
Chlorophyll
Mutants ± CI
Degree of
Sterility
(%)
Noudex
Presscake
1:9000
50
19.2 ± 5.50
7.7
No. 2
Noudex
Presscake
1:4000
85
15.0 ± 5.90
8.3
No. 3
1:2000
1:1000
95
94
19.3 ± 2.70
20.5 ± 8.44
13.8
17.9
Noudex
Hg
1:4000
1:2000
80
83
19.0 ± 7.86
16.5 ± 4.91
6.5
8.5
Prom1x
only
98
5.0 ± 1.96
3.0
2—75
-------
Table 3. Mutagenicity of Methylene Chloride Extracts in Arabidopsis (data
from one experiment)
Sample
Concentration
(mg/mi)
Germination
(%)
% of Fruits
Containing
Chlorophyll
Mutants
Degree of
Sterility
(Z)
Noudex Presscake
No. 2
0.01
0.10
0.25
0.50
90.0
91.0
95.0
88.0
25.6
22.0
22.6
22.0
15.0
4.0
1.0
13.0
Noudex Presscake
No. 3
0.01
0.10
0.25
0.50
98.0
86.4
85.0
82.7
12.0
13.0
18.0
13.0
10.0
12.0
0
12.0
Noudex Hg
0.01
0.10
0.25
0.50
100.0
100.0
89.0
87.3
8.6
14.0
18.0
14.0
1.0
10.0
3.0
13.0
Solvent Control
(DMSO)
100.00
86.5
7.0
7.0
2-76
-------
Table 4. Mutagenicity of Aqueous Extracts in Arabidopsis
% of Fruits
Sample
Concentration
Germination
Containing
Degree of
(%)
(%)
Chlorophyll
Sterility
Mutants ± CI
(%)
Noudex Presscake
1.0
85.0
18.9 ± 5.70
0
No. 2
10.0
100.0
67.0
50.0
26.5 ± 0.98
21.5 ± 2.95
1.0
9.0
Noudex Presscake
1.0
100.0
5.0 ± 3.93
0
No. 3
10.0
100.0
100.0
65.0
8.0 ± 5.90
13.7 ± 6.55
0
2.8
Noudex Hg
1.0
10.0
100.0
100.0
70.0
63.0
6.0 ± 3.93
8.3 ± 7.27
11.5 ± 6.88
9.0
6.0
18.0
Solvent Control
0
100.0
5.0 ± 1.96
0
(water)
Mutagen Control
0.20
100.0
81.5 ± 0.88
80.0
(EMS)
2—77
-------
APPLIC1 TION OF BATU RY OF AQUATIC ‘IOXICITY TESTS TO SOLID WASTE
LE1 CHA.TE CHARACrERIZATION I ND ENVIRONMENTAL EFFECTS PREDICTION
Donald I. Mount, Senior Scientist, Environmental Research Laboratory
Duluth, U.S. EPA, Duluth, Minnesota
ABSTRACT
Most regulatory concerns about chemicals in effluents and leachates
are based on toxicity to some kind of organisms. Only organisms can
sense toxicity; analytical instruments cannot. Minimization and
control of toxicity, as a characteristic of complex water is becoming
commonplace. Tests to measure toxicity of effluents are well
established and are performed by many laboratories. The ability of
these tests to predict impact in receiving water has been extensively
tested. Methods are now becoming available to identify, fractionate,
and reduce cause of toxicity to permit source control, if practicable.
The use of toxicity tests is a cost—effective method to identify where
toxicants need to be controlled and what type and degree of treatment
may be needed to limit emissions to acceptable levels.
2-79
-------
SCREENING OF COMPLEX SOLID WASTES FOR CHEMICALS WHICH BIOACCtJMULATE
ND CAUSE ENVIRONMENTAL HAZARDS
Gilman D. Veith, Director, Environmental Research Laboratory—Duluth,
Duluth, Minnesota
ABSTRACT
The complex waste research program of this laboratory is developing
rational hazard ranking protocols to assist in option selection. If
the waste is uncharacterized chemically, the low—cost toxicity
assessment and fractionation protocol presented by Donald Mount in
this conference will enable toxic waste streams to be minimized. If
the waste has been characterized chemically, the relative hazards of
the chemical as well as the toxicity of the mixture of chemicals can
be adequately estimated from the list of chemicals using the QSAR
system.
The QSAR system is an integrated chemical information and modeling
system based on structure—activity relationship. The system
integrates the environmental toxicity database, AQUIRE, the chemical
properties database, CHEMPROF, with a variety of other toxicity
databases. Unavailable data for 14 key chemical properties,
bioaccumulation potential, acute and chronic toxicity, persistence,
and carcinogenicity are estimated for each chemical from the
structure. Finally, the toxicity data are incorporated into the
ERL—Duluth joint toxic action models to estimate the toxicity of the
complex mixture. We have found these estimates limited only by the
accuracy of the chemical analysis.
2-81
-------
BIOACTIVITY DIFFERENCES OF WATER AND SODIUM ACETATE
ELUATE FROM MUNICIPAL AND INDUSTRIAL WASTES
Spencer A. Peterson, Joseph C. Greene and William E. Miller,
Corvallis Environmental Research Laboratory, Hazardous Wastes and
Water Branch, Hazardous Waste Assessment Team; and David C. Wilborn,
Northrop Services, Inc., Corvallis, Oregon
ABSTRACT
Aqueous and sodium acetate eluates prepared from municipal and
industrial waste products, and the sodium acetate extraction fluid
(eluent) recommended in the EPA Toxicity Characteristic Extraction
Procedure (TCLP) were assayed for their toxicity potential. The
bioassays included algae, Selenastrum capricornutum ;
macroinvertebrates, Daphnia magna ; lettuce root elongation, Lactuca
sativa L.; and Microtox, Photobacterium phosphoreum . The pH 5.0
TCLP sodium acetate eluent was highly toxic to each of the
organisms. Adjustment of its pH to 7.0 decreased toxicity
approximately 2.5—fold for algae and lettuce, and 6—fold for D.
magna . This reduction, while statistically significant, did not
change the toxic classification of the TCLP eluent. Photobacterium
phosphoreum was unaffected by pH 7 sodium acetate after 30—minutes
exposure. Toxicity of the industrial waste TCLP eluates to S.
capricornutum and D. magna was similar to that obtained with water
extraction. The response of B. magna to the TCLP eluates from
sewage sludge (POTW #2), municipal ash, paint sludge, Midco volatile
soil, and First Chemical indicated that these samples contained
acetate soluble contaminants which were more toxic than the acetate
eluent itself. Bioassays detected toxicity in either TCLP or water
eluted samples. Toxicity of water eluted samples can be attributed
only to materials leached from the waste. Toxicity of the TCLP
eluted samples is complicated by the uncertainty of how much
toxicity is attributable to the leachate itself vs. the materials
leached from the samples.
INTRODUCTION
The Environmental Protection Agency (EPA) has developed a Toxicity
Characteristic Leaching Procedure (Friedman, 1985), designed to
simulate waste leachate from a sanitary landfill. The Toxicity
Characteristic Leaching Procedure (TCLP) replaces the Extraction
Procedure (EP) toxicity test recommended in the 1976 Resource
Conservation and Recovery Act. The EP toxicity test is designed to
extract and screen for heavy metals, whereas the TCLP is designed to
extract and screen for heavy metals and organic contaminants.
The TCLP involves mixing site samples with sodium acetate in a
specially designed extractor vessel capable of handling volatile and
2-83
-------
nonvolatile organic compounds. The extracted contaminants are
chemically analyzed to determine the composition of the eluates from
samples collected at hazardous waste sites and sanitary landf 111g.
Sodium acetate was chosen after comparing the chemical analysis of
waste mixture eluates extracted with carbonic acid, deionized water,
or pH 2.9 and 5.0 buffered solutions of sodium acetate. No attempt
was made to determine toxicity of the various eluents. It is
assumed that toxicity of a sodium acetate TCLP waste eluate is
represented by the sum of the toxicities of its individual,
chemically analyzed constituents. This assumption is flawed because
the authors of chemical criteria did not Intend for them to be used
additively (they specifically advised against It). Furthermore,
chemical criteria values were developed in aqueous solutions or with
water soluble organic solvents known as toxicity. The toxicity of
an eluent must be established if it Is used to prepare eluates in
which living organisms will be used to assess toxicity.
DanIels (1981) cited acetic acid as an example of an organic solvent
that has toxicity characteristics in and of itself. The concern for
this problem led us to evaluate the feasibility of using TCLP acetic
acid derived eluatea during biological assessment of the toxicity of
solid waste leachates. The evaluation was performed by testing both
aqueous and TCLP eluatea and comparing the resultant EC 50 or LC 50
concentrations.
Jackson et al. (1984) and Porcefla (1983) recommend deionized water
as the eluent for determining the potential fate and environmental
hazard of chemicals at hazardous waste sites and landfills. The
advantages of deionized water are that: (1) it is nontoxic to test
organisms; (2) it is a clean, readily partitioned solvent for
Inorganic and organic analysis; (3) it is a realistic solvent, given
that soil leacbate mobility is driven primarily by precipitation
events; and (4) its use permits prediction of the amount of water
soluble contaminants that might be mobilized from a specific waste
site soil, sludge, or landfill sample.
The U.S. EPA Office of Solid Waste (OSW) wanted to determine the
feasibility of using TCLP eluates for biological toxicity
characterization of wastes in addition to the chemical
characterization. The objectives of this study were to: (1) define
toxicity of the recommended TCLP sodium acetate eluent; and (2)
determine the suitability of aqueous vs TCLP eluates for assessing
toxicity of leached substances.
METHODS
Selected Industrial wastes, sewage sludges, municipal ash, a
dimethylphenol Ful].era earth positive toxicant control, and the TCLP
eluent solutions (pH 5 and pH 5 adjusted to 7) were bloassayed.
TCLP eluates from the wastes were prepared by ENSECO, 1 Inc.,
2-84
-------
Cambridge, Massachusetts, in accordance with the conditions set
forth in the proposed TCLP and sent to CERL under the direction of
Gail Han8on, EPA, OSW, Washington, DC.
Bioassays performed on the eluates were a 96—hr algal test
( Selenastrum capricornutum ) and a 48—hr macroinvertebrate test
( Daphnia magna ) according to the methods and data quality assurance
requirements prescribed by Porcella (1983). They were bioassayed
with the Microtox ( Photobacteriurn phosphoreum ) test conducted
according to Beckman (1982) and the lettuce ( Lactuca sativa L.) root
elongation (RE) test.
The RE tests were conducted in 100—turn diameter glass petri dishes.
Each dish contained 4.0 ml of test solution dispensed onto a sheet
of 90—mm diameter #3 Whatman filter paper. Five seeds were placed
in each of three replicate control and treated petri dishes and
incubated at 24 ± 2°C iii the dark for 120—hours. At the end of the
incubation period, percent germination was recorded and individual
root length was measured.
The pH tolerance limits for each bioassay test organism has been
established (Miller et al., 1978; Greene, 1984) as follows: S.
capricornutum , 5.5—10.0; D. magna , 6.5—10.05; L. sativa , 3.5—10.0;
P. phoaphoreurn , 6.0—9.0. Therefore, the pH of TCLP eluates (3.8 to
5.8) were adjusted, prior to assay, to ensure against pH shock. The
pH of all water eluates ranged from 6.9 to 8.9 and were not adjusted.
Water eluates were prepared by mixing 1.0 kg of solid waste sample
with 4.0 liters of deionized water. The mixtures, each contained in
a 10—liter cubitainer, were shaken at 100 rpm iii a constant
temperature room held at 24 ± 2°C. The waste/water mixture was
centrifuged in 300 ml polycarbonate bottles for 10 minutes at 10,000
riup. The centrifugate was filtered through a 0.45 urn membrane
filter prior to assay.
The pH of each TCLP eluate was adjusted to fall within the range of
6.5 to 7.5 with 0.1 N NaOH prior to assay with S. capricornutum , D.
magna , and L. sativa . The neutral pH (7.0 + 0.5) of the 2Z NaCl
Microtox osmotic adjustment solution was unaffected by the small
volume (usually 1.5 ml) of the TCLP eluates added to the Microtox
test medium, thus no pH adjustments were necessary.
EC 50 eluate concentrations for S. capricornutum and the results of
the root elongation (RE) test were calculated by linear regression
analysis of the control and treatment measurements. The trimmed
Spearman—Karber method of probit analysis (Hamilton et al., 1977)
was used to calculate D. magna LC 50 values.
2-85
-------
RESULTS AND DISCUSSION
Sodium Acetate Toxicity
The first objective of this study was to define the toxicity of the
TCLP sodium acetate eluent. A volume to volume dilution series of
the pH 5 and pH 5 adjusted to pH 7 eluents was prepared and assayed
with P. phosphoreum , D. magna , S capricornutum , and L. sativa . The
mean LC 50 or EC 50 responses are reported in TiTle 1. Tests
conducted at pH 5 demonstrated very high “toxicity” responses, i.e.,
very toxic at low concentrations of eluent. It took only 0.5%
eluent to produce an EC 50 response with S capcricornutum while 7.6%
eluent was required to produce an LC 50 response from D. magna .
Assays with P. p iosphoreua , D. magna , S. capricornutum , and L.
8ativa are normally run within the tolerance limits of pH 6 to pH
10. Therefore, the p11 5.0 sodium acetate bioassay results are
unreliable relative to toxicity .2 ! se vs pH shock. Because of
this, the assays were also conducted on the eluent with the pH
adjusted to 7 (well within the tolerance limits of all the test
organisms). The same order of sensitivity was demonstrated with
TCLP eluent adjusted to pH 7 as was seen in tests conducted at PH
5.0. Adjustment of p11 from 5 to 7 eliminated p11 shock. Therefore,
the pH 7 column reflects toxicity of the TCLP sodium acetate eluent
in and of itself (Table 1). These data demonstrate that the eluent
is: highly toxic to S. capricornutum and L. sativa ; moderately
toxic to D. magna ; and nontoxic to P. phosphoreum , according to the
rankings of Porcella (1983).
Water Eluate Toxicity
BC 50 values, and their respective 95% confidence limits, for the
water eluates of waste samples, are presented in Table 2. The
municipal ash samples, including the algal response observed in the
12 sample, were nontoxic. Chemical analysis of the 12 municipal ash
eluate measured 1.5 mg/i aluminum. Aluminum is a phosphorous
removal agent (Cooke at al. 1986). Addition of phosphorus to 12
municipal ash eluate, in excess of its reactive aluminum content,
restored the yield of S. capricornutum to that obtained in the
control. The highly toxic algal response to the 12 ash eluate was
due to inactivation of phosphorus and not to toxicity. D. magna was
the only test organism to identify the toxicity of the water
extracted Fuliers soil 11 and 12 (positive toxicant control) samples
spiked with 50 and 150 mg/i dimethyiphenol, respectively. Each of
the aquatic tests identified the First Chemical and paint sludge
industrial waste eluates as being highly toxic (Table 3).
Approximately 78% f the water eluates were toxic to D. magna , 67% to
S. capricornutum . and 22% to P. phosphoreum .
2-86
-------
Table 1. Response of test organisms to TCLP sodium acetate extraction eluerits
at pH 5 and at pH 5 adjusted to pH 7. Results are reported as v:v
(%) EC 5 or LC 53 ( Daphnia ) eluent concentrations.
Bioassay
pH7
pH5
Mean
95%
CI
Mean
95% CI
S. capricornutum
1.2
NO
0.9
EFFECT
— 1.7
0.5
2.1
0.2- 0.8
0.3- 3.7
Microtox 30-mm
Lettuce Root Elong.
15.0
6.4
—26.8
5.7
0.0-18.3
D. magna
45.8
43.7
—48.0
7.6
7.1- 8.0
2-87
-------
Table 2. IC 50 1 or [ C 50 response (as % of eluates) of test organisms to water
eluates from selected sewage sludge, municipal ash, and industrial
waste samples.
Sample ID
Algae
B. inagna
Microtox
Lettuce RE 2
P01W SLUDGE #1
13.840
34.280
lower 95% CI
8.350
29.02
NE 3
NE
upper 95% CI
18.340
40.490
P01W SLUDGE *2
21.760
43.43
lower 95% CI
5.200
42.430
NE
NE
upper 95% CI
48.900
43.43
MUNICIPAL ASH #1
lower 95%C1
NE
NE
NE
NE
upper 95% CI
MUNICIPAL ASH #2
14.560
lower 95% CI
9.140
NE
NE
NE
upper 95% CI
21 .400
PAINT SLUDGE
0.210
3.930
0.940
1.4
lower 95% CI
0.038
3.490
0.830
0.6
upper 95% CI
0.580
4.420
1.050
2.9
MIDCO VOLATILE SOIL
26.560
86.260
83 .5
lower 95% CI
11.330
77.050
NE
32.4
upper 95% C I
43.740
95.850
100.0
DMP FULLERS SOIL #1
84.510
lower 95% CI
NE
80.430
NE
NE
upper 95% CI
100.000
DMP FIJILERS SOIL #2
62.300
lower 95% CI
NE
48.500
NE
NE
upper 95% CI
80 .010
FIRST CHEMICAL
0.005
0.001
0.054
3.1
lower 95% CI
0.001
0.001
0.046
2.5
upper 95% CI
0.009
0.002
0.062
3.9
1 Ic 50 is for D. inagna .
2 RE = root elongation.
3 NE = no observable toxic effect.
4 Results were caused by inactivation of phosphorus by aluminum. This sample
did not contain toxic constituents.
2-88
-------
Table 3. Toxicity classification of OSW water and TCLP eluates. 1
Waste Sample
Algae
Daphnia
Water/TCLP
30 Mm
Microtox
Water/TCLP
Water/TCLP
First Chemical
H / H
H / H
H / H
Paint Sludge
H / H
H / H
H / H
P01W #1
H /NSD
M /NSD
NE I H
Municipal Ash #2
H / H
NE / H
NE / H
P01W #2
M /NSD
M /NSD
NE / NE
MidcoSoil
DMP Fullers Soil
#1
M /H
NE 3 /NSD
L /H
L /NSD
NE /H
NE I H
Municipal Ash #1
NE / H
NE / H
NE / H
DMP Fullers Soil
#2
NE /NSD
M /NSD
NE / NE
67 / 56
78 / 56
22 / 0
1 Toxicity of the eluates ranked as high (H = < 20%), moderate (M = 20-75%),
and low (L = > 75%) according to Porcella (1983).
2 NSD = Not significantly different from that obtained for pH 7.0 TCLP eluent
(Table 1).
3 NE = no observable toxic effect in 100% water eluate.
4 Percent of OSW water and TCLP eluates, respectively, which exhibited toxicity.
2-89
-------
TCLP Eluate ‘vs. Water Eluate Toxicity
The second objective of this study was to evaluate TCLP and water
eluate toxicity of the same samples (Table 2 and 4). Interpretation
of these results is confounded by the fact that the TCLP eluent may
have introduced a toxic effect into the overall TCLP toxicity
value. If the toxic effect of the eluent was predictable, one could
theoretically account for it by calculation of the leached substance
toxicity value, i.e., total toxicity — eluent toxicity = eluate
toxicity. However, as pointed out by Daniels (1981), there are too
many uncertainties associated with chemical mixture toxicity
(synergism, antagonism, degradation compound effects, etc.) to make
this simple assumption. For example, toxicity of four TCLP eluates
for S. capricornutum and D. Magna (Table 4) fall within the
confidence limits developed for the pH 7 TCLP eluent (Table 1).
However, interpretation of these results is impossible, relative to
eluent vs. eluate toxicity given the test organisms sensitivities
and the range of response exemplified by the confidence limits for
the test, i.e., all of the toxicity response could be due either to
the eluent or to a combination of eluent and eluate components.
S. capricornutum and D. Magna response to the municipal ash 11 and
72, paint sludge, Mldco volatile soil, and First chemical TCLP
eluatea (Table 4) all produced EC 50 s or LC5Os, lower than, and
outside of, the confidence intervals shown for the pH 7 TCLP eluent
in Table 1. The highly toxic response for these samples would
suggest that something in these eluates produced a toxic effect in
addition to the toxic effects of the eluent.
D. Magna responses for POTW sludge 11, and DMP Fullers soil 11 and
72 eluates are within the levels of toxicity associated with the
TCLP eluent. The F capricornutum response to the POTW sludge 12
eluate was the same as that measured for the TCLP eluent. These
observations are of particular interest when compared with D. magna
and S. capricornutum water eluted toxicities for the same samples.
Water eluate toxicity for POTW 11 was moderate for D. magna to high
for S. capricornutum ; that for DMP Fullers soil 11 and 12 were low
to moderate for D. Magna and nontoxic to S. capricornutum . The TCLP
eluate D. responses for the POTW sludge and Fullers soil
samples fall within the range of confidence limits expected for the
pH 7 TCLP eluent.
First chemical and paint sludge TCLP and water eluates was highly
toxic to all four of the test organisms. Comparison of the toxicity
classification of the water vs TCLP eluates In Table 3 shows that
the First Chemical, paint sludge, Mldco soil and municipal ash TCLP
eluates are highly toxic to S. capricornutum , D. magna and P.
phos phoreum . A shift in toxicity from low or moderate to high
(Midco soil) and no observable effect highly toxic (Municipal Ash
11) was obtained by TCLP extraction using D. magna and S.
2-90
-------
Table 4. LC 0 1 or ECç 0 response (as % of eluates) of test organisms to pH
adjusted TCLP eluates from selected sewage sludge, municipal ash, and
industrial waste samples.
Sample ID
Algae
D. magna
Microtox
Lettuce RE 2
POTW SLUDGE #1
0.630
50.5b0
12.23
4.28
lower 95% CI
0.520
44.400
7.30
2.82
upper 95% CI
0.740
57.400
17.12
6.09
P01W SLUDGE #2
lower 95% CI
1.250
0.990
39.360
30.940
NE 3
56.77
45.84
upper 95% CI
1.730
50 .070
70.29
MUNICIPAL ASH #1
0.390
8.300
11.38
41.29
lower 95% CI
0.280
6.970
10.38
34.41
upper 95% CI
0.520
9.880
12.37
48.44
MUNICIPAL ASH #2
0.240
5.140
5.10
19.2
lower 95% CI
0.040
3.960
3.69
13.21
upper 95% CI
0.450
6.670
5.61
26.51
PAINT SLUDGE
0.360
1.210
0.55
1.61
lower 95% CI
0.180
1.040
0.42
1.10
upper 95% CI
0.570
1.400
0.60
2.28
MIDCO VOLATILE SOIL
0.200
0.270
7.15
14.18
lower 95% CI
0.190
0.210
5.08
6.24
upper 95% CI
0.200
0.350
10.62
22.54
DMP FULLERS SOIL #1
0.520
45.440
16.80
24.00
lower 95% CI
0.150
41.380
11.43
12.34
upper 95% CI
1.400
49.890
21.83
38.56
DMP FULLERS SOIL #2
0.590
54.330
23.00
lower 95% CI
0.210
47.520
NE
9.35
upper 95% CI
1.060
62.100
38.34
FIRST CHEMICAL
0.005
0.004
0.13
•
0.85
lower 95% CI
0.002
0.003
0.12
0.53
upper 95% CI
0.008
0.006
0.15
1.32
1 LC 50 is for 0. magna .
2 RE = root el ngation.
3 NE = no observable toxic effect.
2-91
-------
capricornutum , respectively. We are, however, confident that the
toxicity classifications of the water eluates were due to water
extractable contaminants since no artificial toxicants were added.
Highly toxic samples, such as paint sludge and First Chemical, will
be detected by either the TCLP or water elution process. However,
the use of water eluates will greatly reduce the potential for false
positive responses while using an environmentally realistic
extrac taut.
SUMMARY AND CONCLUSION
Toxicity of the Office of Solid Waste proposed TCLP (sodium acetate)
extraction fluid was characterized using four bioassay tests (algae,
S. capricornutum ; nacroinvertebrates, D. magna ; Nicrotox, P.
phosphoreum , and lettuce L. sativa L.). The same bioassay tests
were used to compare toxicity responses for TCLP and water eluted
sample. fro. seven wastes and two Fullera soil positive controls.
Conclusions from the study are as follows:
1. The TCLP (sodium acetate) eluent was highly toxic to S.
capricornutum , and L. sati’va , and moderately toxic to D. mag ia
despite adjustments of its pH to 7 prevent pH shock to the test
organisms. It was nontoxic to P. phoaphoreum exposed for 30
minutes.
2. Toxicity of the TCLP eluent may confound interpretation of
toxicity in eoe of the TCLP eluates, since it introduces an
unpredictable toxic factor that cannot be corrected for in any
simple *nner.
3. Water elution is suggested as an environmentally relevant and
meaningful procedure for measurements of leachable toxicity from
soil and solid waste.
FOOTNOTES
1 -Mention of trade names or commercial products does not
constitute recommendation or endorsement by the EPA.
ACKNOWLEDGEMENTS
Special thanks are extended to Cathy Bartels, Julius Nwosu, and
Sheila Smith for providing the bioassay support for this study.
Also, the typing of Nancy Lanpheare is acknowledged in preparing
this report. Their help is greatly appreciated.
2-92
-------
REFERENCES
Beckman. 1982. Microtox system operating manual. Beckman
Instruments, Inc., Microbics Operations, Carlsbad, California.
Cooke, D. C., E. B. Welch, S. A. Peterson, and P. R.. Newroth.
1986. Lake and Reservoir Restoration . Butterworth
Publishers, Massachusetts. 450 pp.
Daniels, S. L. 1981. Development of realistic tests for effects
and exposures of solid wastes. Hazardous Solid Waste
Testing: First Conference. ASTM STP 760, R. A. Conway and B.
C. Malloy, Eds., M er1can Society for Testing and Materials,
pp. 345—365.
Friedman, D. 1985. Proposed EPA toxicity characteristic leaching
procedure. Memo #9, October 1985. Office of Solid Waste,
U.S. Environmental Protection Agency, Washington, D.C.
Greene, J. C. 1984. Unpublished data. Environmental Research
Laboratory, U.S. Environmental Protection Agency, Corvallis,
Oregon.
Hamilton, M. A., R. C. Russo, and R. V. Thurston. 1977. Trimmed
Spearman—Karber method for estimating median lethal
concentrations in toxicity bloassays. Environ . Sci. Tech .
ll:714—719. Correction 12:417 (1978).
Inside EPA. 1985. Weekly report . Vol. 7, No. 51. P. 0. Box
7167, Ben Franklin Station, Washington, D.C. 20044.
Jackson, D. R., B. C. Garrett, and T. A. Bishop. 1984.
Comparison of batch and column methods for asse8sing
leachability of hazardous waste. Environ . Sd. Tech .
18:668—673.
Miller, W. E., J. C. Greene, and T. Shiroyama. 1978. The
Selenastrum capricornutum Printz Algal Assay Bottle Test:
Experimental Design, Application, and Data Interpretation
Protocol. EPA—600/9—78—018, U.S. Environmental Protection
Agency, Corvallis, Oregon.
Porcella, D. B. 1983. Protocol for Bloassesament of Hazardous
Waste Sites. EPA—600/2—83—043. U.S. Environmental Protection
Agency, Corvallis, Oregon.
2-93
-------
THE USE OF MOSSES AS INDICATORS OF AIR POLLUTION
Gloria W. Sage, Syracuse Research Corporation, Merrill Lane,
Syracuse, New York
ABSTRACT
Epiphytic moss possess certain properties that make them ideal
specimens for assessing air pollution; they obtain their nutrients
from the air and concentrate these pollutants. Among the pollutants
they are known to collect are heavy metals, polyaromatic
hydrocarbons, and chlorinated hydrocarbons. Recent studies have
shown that pollutant concentrations in moss correlate with the
atmospheric deposition of the pollutant over a time period,
demonstrating that mosses are integrators of air pollutants. The
measurement of pollutant concentrations in moss biomonitors is
therefore a very convenient, simple, and inexpensive means of
quantitating air pollution over a period of time. Some recent
studies in the use of mosses to quantify levels of pollutants in air
and map pollutant distributions will be reviewed with an eye to
their utility in assessing air pollution from waste incinerators.
INTRODUCTION
Coupled with the increased reliance on resource recovery plants or
incinerations to solve the solid waste crisis, is the need to insure
the safety of these facilities. This requires knowing the nature
and quantity of the pollutants they emit into the environment.
Investigators have reported on various classes of these chemicals
including polyaromatic hydrocarbons and their derivatives,
polyhalogenated aromatics, and metals which may impact on health and
the environment 1 -. If one is to insure the safety of waste
incinerators, it is prudent to perform a certain amount of
monitoring of both the ambient air and deposition in the vicinity of
the plants.
Many of the pollutants of greatest concern emanating from
waste—burning facilities are associated with particulate matter 10 .
These chemicals may be deposited on soil or leaves and taken up by
plants or may simply be adsorbed on the surface of the plant. The
contaminated plants may then be ingested by grazing animals.
Ingestion has been found to be the most important route of intake in
humans for many of these substances 2 ° and dioxins bioconcentrated
into mother’s milk was the cause of a recent moratorium on resource
recovery plants in Sweden. Traditional air monitoring is extremely
expensive and time consuming to perform on a meaningful scale. In
addition, air concentrations of pollutants such as trace metals, for
example, do not correlate well with amounts deposited on land and
taken up into the food chain 3 . The reason for this is apparent when
one notes the Importance of factors such as wind speed and surface
2-95
-------
characteristics of the receptor on the amount of deposition 3 . While
modeling stack emissions is a useful way of estimating environmental
concentrations, these calculations must be validated and the results
supplemented with actual measurements. Results of this modeling is
often highly inaccurate for simple gaseous emissions 2 and modeling
Is far more difficult when chemicals adsorbed or partially adsorbed
on particulates are included.
Mosses posses properties that make them ideal for monitoring
concentrations of air pollutants emitted from waste incinerators,
particularly heavy metals, polyaromatic hydrocarbons, and higher
molecular weight halogenated chemicals. Accumulations of pollutants
in mosa better reflect the amount of the chemical deposited on
vegetation and take up in the food chain than traditional air
monitoring 3 . The development of moss bags has produced an element
of flexibility and axi enhanced ability for standardization that make
the use of bryophites extremely attractive for site monitoring 4—6k
Other recent experiments which show that the btoaccumulation of
heavy metals and polyaromatic hydrocarbons in moss is highly
correlated with dry and wet deposition 7 ’ 8 , makes the use of these
biomonitors even more attractive. It 18 our intention to review
some of the key work and recent developments concerning the use of
mosses for assessing aerial burdens of pollutants.
Mosses have been used as biomonitora of airborne metal and organic
pollutants, primarily in Europe, for close to 20 years 9 ’ 14 . The
early studies primarily used indigenous mosses and showed that
mosses were efficient indicators and integrators of aerial metal
burdens. In one of these studies, Goodman and Roberts 9 demonstrated
that concentrations of the metals Zn, Pb, Cd, Ni, Cu, and Mg In the
epiphytic moss, Hypnum cupreasiforme, around Swansea in South Wales
were Indicative of local sources and wind direction and that the
moss was much more efficient at accumulating metals than grass or
soil. They also pioneered the use of transplants, moss on logs
taken fro. uncontaminated sites, and moss bags, moss hung out in
nylon mesh bags, as bloaccumulators. The results of the first study
employing moss bags showed that elevated metals levels appeared 10
km upwind of Swansea, rose to about 4—20 tImes background in the
city, and declined to background levels 25—30 km downwind of
Swansea. It was also observed that mosses kept accumulating metals
by a passive process after they died.
There are a number of properties that make moss so useful for
monitoring aerial burdens of pollutants. Many species obtain all
their nutrients from the air. Moss has an extremely high
adaorptivity due to its extraordinary surface area and roughness and
high cation exchange capacity. The byrophytes have no epidermis and
cutic1e 1 1 . This makes the cell walls easily penetrable. Transport
of nutrients between segments of new growth Is poor due to lack of
vascular tissue 11 . In addition mosses have wide geographic
2-96
-------
distribution and are extremely drought resistant. The strong
adsorptivity for metals combined with lack of connectivity between
new and old growth has enabled Lee and Talus to show the historical
development of the lead industry in England by plotting peat
profiles 12 .
As with ail bloinonitors, one has problems resulting from differences
in accumulation between different species, individual plants
depending on age and environmental factors, and parts of the
plant 1 - 1 ’-’- 7 ’ 21 -. Problems may arise with indigenous bryophites from
lack of knowledge of the period of accumulation, from having mixed
species, and from not having flexibility in placing samples.
Accumulation of chemicals will depend on how exposed the moss is to
the wind and rain whether it is beneath the leaf canopy where it
collects throughfall of rain. Therefore it is best to standardize
the procedure carefully and use composite samples of a single
species for regional studies employing indigenous samples 15 ” 6 .
This was done in Crodzinska’s survey of heavy metal pollution iii
Polish National Parks 15 . She even separated the moss into brown and
green parts, the green parts usually representing the latest 2
years, and the brown the previous three. Finally at very high
exposure levels to more than one metal, the more strongly adsorbed
metal may displace a less strongly bound one 1 - 7 . Fewer regional
surveys of polyaromatic hydrocarbons and chlorinated hydrocarbons in
mosses have been reported. Thomas reported on accumulations of some
of these chemicals in Hypnum cupressiforme in Bavaria 7 and Bacci et
al. measured accumulations of chlorinated hydrocarbon insecticides
and PCBs in moss in Antarctica 23 .
By taking moss covered logs from unpolluted areas, one can overcome
many of the drawbacks inherent in the use of indigenous moss. The
transplants can be placed at desired locations around a site of
interest and variability of the monitor due to environmental factors
is reduced because they can be taken from the same location or even
the same log. While mapping the PAM concentrations of mosses from 9
sites in Western Europe, Thomas 8 noted that the ratio of
fluoranthene to beazoperylene is 1.4 in an industrial area and rises
to 6—10 in remote areas since a higher proportion of the
fluoranthene is in the vapor phase and therefore transported
further. Analogous regional, differences in the ratios of .C —BHC
to ‘7—BHC were also noted and explained by the greater presence of
B—BHC in agricultural areas. In using transplants, rather than
indigenous moss, exposure time can be controlled. Pilegaard used
transplants of Dicranoweisia cirrata to study the relation between
the accumulation of nine heavy metals with exposure time and fallout
from the atmosphere - 9 . Accumulations in moss and fallout, both wet
deposition and dry fallout, were measured at 12 stations around a
steelworks and regression equations and correlation coefficients
determined after seven months of exposure. Good correlations were
obtained (correlation coefficients between 0.816 and 0.985) for all
2—97
-------
metals but Vanadium. Thomas was able to establish a good multIple
regression equations relating the concentrations of the metals Zn,
Cd, and Pb and the polyaromatic hydrocarbons, fluoranthrene,
benzo(a)pyrene, and beuzoperylene in Hypnum cupressiforme with their
respective concentrations in rainwater atmospheric particulate
matter and the amount of precipitation” 8 . On the other hand,
chlorinated hydrocarbons did not give good correlations and it was
suggested that a high exchange rate between the plant and air
prevented long—term accumulation by the moss. The results clearly
show that the accumulation of PARs is most strongly correlated with
the concentrations in atmospheric dust, while concentrations in dust
and bulk precipitation contribute equally to the accumulation of
metals.
While the use of transplants has enabled researchers to control
sample variability to a large extent and have flexibility in sample
placement, moss bags allow for even greater standardization and
improvement in technique. Sphagnum moss soon became the bryophite
of choice for use in moss bags. Sphagnum moss is abundant and its
high absorbancy and cation exchange capacity are well known and have
long been put to practical use. One can be selective and choose
similar specimens for an entire project since even the height above
the water table affects (increases) the cation exchange capacity of
the moss 22 . When used to monitor aerial metal pollution, the plants
can be soaked in acid to replace any bound metal and assure a low
background level 4 . An additional benefit is that there is no need
to separate the moss from a log or rock and remove attached debris
before analysis. Galley and Lloyd 4 performed a detailed study at
Armadale, a town in central Scotland containing a foundry.
Spherical nylon net bags containing sphagnum moss were hung at 47
locations around the town for 8 two month periods after which the
bags were analyzed for seven heavy metals. Results revealed two
main areas of metal deposition in the town, areas where clusters of
lung cancer cases had been found. One area was north of the foundry
at which the high concentrations of metal were due to local
topography and wind. Napping enabled gradients to be seen and
correlation between concentration gave testimony to a common
source. This type of study would enable good validation of modeling
studies. The co8t of materials for the 17 month survey was less
than $500.
REF ENCES
Oehme M. et al., Formation and presence of polyhalogenated and
polycyclic compounds in the emissions of small and large scale
municipal waste incinerators. themosphere 16:143—153 (1987).
U.S. General Accounting Office (GAO). Publication RCED—86—94.
Washington, DC: General Accounting Office (1986).
2-98
-------
Goodman, G.T.; Smith; Inskip, M.J.; and Parry, G.D.R. Trace Metals
as Pollutants: Monitoring Aerial Burdens. mt. Conf. Heavy
Metals Environ. [ Symp. Proc.] 1st, Vol 2, Issue 2 pp. 623—42.
Hutchinson, T.C. ed. Univ. Toronto Inst. Environ. Studies,
Toronto, Ont. (1975).
Gailey, F.A.Y. and Lloyd, O.L. Atmospheric Metal Pollution Monitored
by Spherical Moss Bags: A Case Study of Artnadale. Environ.
Health Perspectives 68: 187—96 (1986).
Galley, F.A.Y. and Lloyd, O.L. Methodological investigation into low
technology monitoring of atmospheric metal pollution. Part I.
The effect of sample size on metal concentrations. Environ.
Pollut. (Ser. B) 12: 41—59 (1986).
Galley, F.A.Y. and Lloyd, O.L. Methodological investigation into low
technology monitoring of atmospheric metal pollution. Part III.
The degree of replicability of metal concentrations. Environ.
Pollut. (Ser. B) 12: 85—109 (1986).
Thomas, W. Statistical Models for the Accumulation of PAH,
Q-ilorinated Hydrocarbons and Trace Metals in Hypnum
cupressiforme. Water, Air, and Soil Pollution 22, 351—71 (1984).
Thomas, W. Representativity of Mosses as Biomonitor Organisms for
the Accumulation of Environmental chemicals in Plants and Soils.
Ecotox. Environ. Safety 11, 339—46 (1986).
Goodman, G.T. and Roberts, T.M. Plants and soils as indicators of
metals in the air. Nature 231; 287—92 (1971).
Cass G.R. and McRae, G.J. Emissions and air quality relationships
for atmospheric trace metals. In: Toxic metals in the
Atmosphere. Nriagu, J.O. and Davidson, C.I. eds. pp. 145—171
(1986).
Steubing, L. Problems of Bioindication and the necessity of
standardization. In: Monitoring Air Pollution by Plants.
Methods and problems. Steubing, L. & Jager H.L. eds. pp. 19—24.
Dr. W. Junk, The Hague. (1982).
Lee, J.A. and Talus, J.H. Regional and historical aspects of lead
pollution in Britain. Nature 245: 21618. (1973).
Grodzinska, K. Monitoring of air pollution by mosses and tree
bark. In: Monitoring Air Pollution by Plants. Methods and
Problems. Steubing, L. & Jager, H.L. eds. pp. 33—42. Dr. W.
Junk, The Hague. (1982).
2—99
-------
Ruhling, A. and Tyle Ca. An ecological approach to the lead
problem. Bot. Notiser 121: 321 (1968).
Grodzlnaka, K. Mosses as bloindicators of heavy metal pollution In
Polish National Parks. Water, Air, and Soil Pollution 9: 83—97
(1978).
Yule, l.A. and Lloyd, 0. LL. Metal Content of an indigenous moss In
Armadale, Central Scotland. Water, Air, and Soil Pollution
21:261—70 (1984).
Brown, D.H. Mineral Nutrition. In: Bryophite Ecology pp.
383—444. SmIth A.J.E. ed Qiapsan & flail, London (1982).
Little, P. and Martin, M.H. Biological monitoring of heavy metal
pollution. Faviron. Pollut., Ser. A. 6: 1—19 (1974).
Pilegaard K. Heavy metals in bulk precipitation and transplanted
Hypogymnia physodes and Dicranoweisia cirrata in the vicinity of
a Danish steelworks. Water, Air and Soil Pollution 11: 77—91
(1975).
Bennet, B.C. posure assessment for metals involved In
carcenogenesia. pp. 115—127 in &ivironaental Carcinogens.
Selected Methods of An*lyais. Vol 8 — Some Metals: Aa, Be, Cd,
Cr, Ni, Pb, Se, Zn. O’Neill, I.X. et al. eds. International
Agency for Research on Cancer, Lyon (1986).
Folkeson, L. Interspecles calibration of heavy-metal concentration
In nine mosses and lichens: Applicability to deposition
measurements. Water, Air, and Soil Pollution 11:253—60 (1979).
Andrus, R.E. Some aspects of Sphagnum ecology. Can. 1. Hot. 64:
6116—26 (1986).
Mcci E. et al. Qilorinated hydrocarbons In lichen and moss samples
from the Antarctic Peninsula. ( temoaphere 15: 747—54. 1986.
2-100
-------
USE OF TRADESCANTIA FOR TOXICITY TESTING OF HAZARDOUS WASTE
William R. Lower, Group Leader Research, Environmental Trace
Substances Research Center, University of Missouri, Columbia,
Missouri
INTRODUCTION
The current output of solid waste increases the need for a variety
of rapid, cost—effective and accurate bloassays for toxicity.
Traditionally, animals have been used for most bioassay, although
bacteria, such as the Ames test, have also become important. By
contrast, plant bioassay, particularly higher plants, are lamentably
underdeveloped and underexploited. There are several plant systems,
e.g., seed germination and early seedling growth tests, and duck
weed and fresh water and marine algal tests, but compared to animal
test the number of systems is small. Several higher plant bioassays
show potential and one in particular, the spiderwort Tradescantia ,
is immediately applicable. Tradescantia can be used as a system
capable of the integrative monitoring one at a time or any
combination of air, water and soil contamination associated with
solid waste as well as the gaseous, liquid and solid phases of solid
waste. It is readily used both in the laboratory and in situ in the
field under the complexity of real world conditions and for acute
and chronic testing area for periods of exposure of minutes to
months. A variety of biological effects including genotoxicity are
measurable: five aspects of perturbations of electron transport in
photosynthesis; chromosome breakage in pollen mother cells and the
production of micronuclel; somatic stamen hair mutations; sister
chromated exchange in root tips; and growth as flower production.
Stomatal conductance and CO 2 production in photosynthesis and pollen
abortion are under investigation. Currently, the main assays are
the stamen hair test and micronucleus test.
Tradescantia may qualify as a sentinel species, that is, a species
which can be used as an indicator of ecosystem health. Even though
it is not a normal resident species its wide applicability to
assessing air, water and soil contamination, its applicability in
situ to a large variety of ecological situations as well as its use
in the laboratory and the variety of biological end points qualify
it for consideration.
The stamen hair mutation system of Tradescantia was initially used
in radiation studies (Sparrow et al., 1972; Underbrink et al., 1973;
Nauman et al., 1975) and exposure to conditions in Biosatellite II
(Sparrow et al., 1971). Subsequently, the stamen hair system was
used for the testing of chemicals (Sparrow et al., 1974; Underbrink
et al., 1973; Sparrow et al., 1976; Nauman et al., 1976). The
pollen mother cell micronucleus assay was developed by Na (1979) and
2-101
-------
then applied to the testing of over 140 agents (Ma et al., 1976; Ma
et al., 1978; Ma et al., 1984). Both systems have also been used
for in situ field testing for integration of the effects of
pollution variously ranging in time for hours to over five months to
assess ambient air pollution at truck stops around cities and
petrochemical complexes, soil and air contaminants of a lead smelter
and an oil refinery complex, military obscurant smokes, water
pollution from drinking water reservoirs, water from shallow and
deep water wells, sewage sludges, cooling tower water, bottom
sediments from a drinking water reservoir, ambient environment at a
nuclear facility, airports, coal—burning residential district in the
Peoples Republic of china and a foundry complex (Ma et al., 1983a;
Ma et al., 1983b; Lower et al., 1978; Lower et al., 1983a; Lower et
al., l983b; Lower et al., 1985; Schaeffer et al., 1986; Sandhu and
Lower, 1987).
DESCRIPTION OF THE TRADESCANTIA BIOASSAY SYSTEM
The Tradescantia tests differ in their sensitivity and one or more
of these tests may differentially recognized particular
physiologically hazardous or mutagenic agents that might go
unregistered in the others. This aspect is currently under
consideration. The use of the combined test may also permit
accurate monitoring over a wide range of concentrations and various
effects of toxic substances, particularly complex mixtures.
Tradescantia is particularly useful because of its versatility.
Either cuttings or entire rooted plants may be exposed to gaseous
agents in air or liquids, or grown in contaminated soils, and grown
hydroponically in or watered with aqueous solutions of single
chemicals or complex pollutants. Unrooted cuttings may be exposed
for periods of minutes and hours or up to 2 weeks and rooted plants
may be exposed for months If provision is made for suitable light
and temperature conditions.
STAMEN HAIR TEST
Tradescantia has been used for radiobiological cytogenetic studies
for over twenty—five years. However, In the mid 1960’s it was found
that its array of long filimentous chains of contiguous, single
cells growing from the stamens (stamen hairs) presented a plant
structure which could be used as a sensitive and cost efficient test
of chemical mutagens (Underbrlnk et al., 1973).
The average number of stamen hairs per flower, the denominator of
mutation frequency, is routinely determined at the beginning of the
scoring period. A total of 36 stamens from six randomly picked
flowers are counted for each experiment. The average number of
stamen hairs per flower ranges between 300—400. Each Inflorescence
(flower cluster) produces about 18—20 flowers; one blooms about
2—102
-------
every other day and the mutations appear as a phenotypic color
change from the dominant blue to the recessive pink in cells of the
stamen hairs of flowers of clones heterozygous for flower color.
The mutation frequency is expressed as the number of mutant events
per thousand stamen hairs.
In short term experiments scoring of stamen hairs for pink mutant
events usually begins 5 to 7 days after treatment and continues
daily until approximately 15 days after treatment, or until the
number of mutations declines from any evident peak. The time from
treatment to scoring constitutes time for development of buds from
sensitive stages at which they are exposed to maturity when they
exhibit the mutation events in the stamen hairs. If the plants are
allowed to remain in a contaminated area for long periods of time,
the scoring is done daily until the mutation frequency reaches a
stable plateau or as long as the experiment dictates.
The sample size can be pre—deterinined and depends on the anticipated
strength of the mutagen and percent standard error deemed acceptable
(Underbrink et al., 1973; Underbrink and Sparrow, 1974).
Mutagenicity studies using this test have been carried out routinely
in situ in the field and in the laboratory with agents either in the
gaseous or liquid state (Lower et al., 1978; 1983a; 1983b; 1984;
Nauinan et al., 1979; Schairer et al., 1979).
MICRONUCLEUS TEST
The micronucleus test of Tradescantia was developed in 1976 and is 6
to 8 times more sensitive than the stamen hair test (Ma. 1983). The
mutagenic end point is the induction of chromosome breaks or lagging
chromosomes which appear as micronuclei and are scored during the
tetrad stage of meiosis in pollen development. After acetocarmine
staining, the micronuclel appear as small dark red bodies in the
pink cytoplasm and are easily distinguished from the much larger,
darker red nucleus.
The preparation of iriflorescences for scoring of micronuclei
involves the following: After termination of the treatment period,
cuttings are placed in aerated 1/2 strength Hoagland’s nutrient
solution in a controlled—environment chamber under a light regimen
of 16 hrs light/8 hrs dark for 24—32 hours, which is the time
necessary for the chromosomes damaged during early prophase of
meiosis to progress to the tetrad stage of pollen formation (Ma.
1983). Flower buds are then fixed in ethanolacetic acid (3:1 ratio)
for 4 to 30 hours, and transferred to 70 percent ethanol for storage.
For scoring, individual flower buds are removed from the
inflorescences which have been stored in alcohol and the buds likely
to contain the tetrad stage cells are selected based on bud size and
position on the inflorescence. The anthers are dissected out of the
2-103
-------
bud, macerated, and stained with acetocarmine. Tetrads are scored
at 400 X magnification. Approximately 300 tetrads are scored from
each bud, and from 5 to 10 buds are scored for each experimental and
control point.
The micronucleus test has been used as a method of estimating
mutagenesis resulting from chromosome breakage in investigations of
the mutagenic effects of a wide variety of environmental pollutants
In water reservoirs, parking garages, truck stops, diesel exhaust,
sewage sludge, cooling tower water, etc. (Cleveu,ger, et al., 1983;
Ma. 1983; Ma et al., 1983; Lower et al., 1984; Schaeffer et al.,
1986).
PERTURBATIONS OF THE ELEC1RON TRANSPORT OF PHOTOSYNTHESIS
Room temperature chlorophyll fluorescence in vivo is a measure of
oxidation—reduction state of the photosystem II primary acceptor
(Q). Segments of leaves which are illuminated with a specific
wavelength of light will fluoresce a range of wavelengths of light
in response. Rapid changes in the yield of chlorophyll fluorescence
occur within the first 15—30 sec. of illumination in all
photosynthetic plants (Xautsky effect). This chlorophyll
fluorescence induction can be used as an indicator of photosynthetic
energy conversion and has been important in the characterization of
photosynthetic reaction mechanisms.
The fluorescence induction curve (a plot of fluorescence against
time for the period immediately following onset of illumination) is
modified by many factors which affect photosynthesis. Present
understanding of the various fluorescence transients allows the
fluorescence induction curve to be used as an immediate and reliable
test of photosynthetic activity. Variable fluorescence yield can be
increased if electron transport is stimulated, or decreased if
electron transport is inhibited depending upon where a given
chemical produces a lesion in the complex chain of biochemical
events prior to and including the photolysis of water in the Hill
reaction.
Perturbations in electron flow are measured with a plant
fluorometer, a portable AC/DC operated unit suitable for
measurements in the field as well as the laboratory. It is
connected to an AC/DC strip chart recorder. At the end of an
exposure or treatment, samples of leaf material are collected and
placed in plastic bags with moist cotton and dark—adapted for at
least 15 minutes. A 2 cm segment of leaf is placed in a leaf
holder. The fluorometer probe is Inserted on top of the leaf
segment, the dark adapted leaf segment is illuminated at 670 mm and
simultaneous fluorescence at wavelengths greater than 710 nm is
detected and recorded on the strip chart over 10 or more seconds
(Schreiber et al., 1978). For each data point at least 10 strip
2-104
-------
chart recordings are made from 10 samples of leaf material and
measurements are averaged to determine percent increase or decrease
in variable fluorescence.
Changes in the electron transport system have been observed in
studies of cooling tower water (Lower et al., 1985) and military
obscurant smokes used as smoke screen (Schaeffer et al., 1986).
GAS EXCHANGE IN PHOTOSYNTHESIS
The measurement of photosynthetic rate as CO 2 evolution and stomatal
conductance is commonly applied to obtain information on growth of
plants and plant gas exchange research. The same technique can be
applied to plants exposed to toxic materials in both the laboratory
and in situ in the field as an adjunct to the measurement of the
electron transport systems of photosynthesis. No known systematic
investigation has to date been done with this in reference to solid
waste.
Both net photosynthetic rate calculated from net CO 2 exchange
between a leaf and the atmosphere and stomatal resistance calculated
from measured transpiration rate and leaf and air temperature can be
determined in both the laboratory and field using a portable
instrument developed for that purpose.
GROWTH AS FLOWER PRODUCTION
An inflorescence of Tradescantia contains 18 or more flower buds
under continuous development which produce, on the average, a flower
every second day. This continuous production of flowers can be used
as a measurement of growth. Flower production has been used as an
indicator with exposure of Tradescantia to obscurant smoke
(Schaeffer et al., 1986).
Pollen Abortion
Pollen abortion normally occurs to some degree in plants, has both
physiological and genetic causes often difficult to differentiate
under many test conditions, but has potential as a measurement of
toxicity. Anthers are placed in a drop of lacto phenol cotton blue
stain and mascerated with forceps. The debris is removed, a cover
slip is placed over the stained pollen and the non—aborted blue
pollen and aborted empty pollen cells are counted under high dry
400X magnification with a compound microscope.
Examples of Use of Tradescantia
Tradescantia may be used in a variety of ways. With chemicals
assayed in the laboratory the cut flower stalks are placed in an
aqueous solution of the test material or exposed in a chamber if a
2-105
-------
gas or an aerosol is tested. Examples are presented on the use of
Tradescantia in a variety of situations involving complex mixtures.
FREQUENCY OF STMIEN HAIR MUTATIONS OF TRADES CANTIA GROWN IN A GREEN
HOUSE IN BOTTOM SEDIMENT OF A DRINKING WATER RESERVOIR (FREQUENCY +
S.E.) x iO —
Ratio of
Mutation
Frequency
Time Sampling of Sediment/
Period Period Control Sediment Control p Value*
1 09/11—11119 3.2 + 0.3 5.6 + 0.3 1.7 0.005
2 11/20—12/03 2.2 ± 0.1 4.2 + 0.2 1.9 0.005
3 12/04—12/17 2.8 ± 0.2 3.6 ± 0.2 1.3 0.05
4 12/18—01/07 2.1 ± 0.1 3•3 ± 0.1 1.6 0.005
5 01/08—02/05 2.5 ± 0.3 3.1 ± 0.2 1.2 0.05
*p Value of Two Tailed Approximation to Binomial
Bottom sediment from a water reservoir was removed and placed in
untreated wooden boxes in a greenhouse. Tradescantia were rooted in
the sediment and the stamen hair mutation frequencies were
determined beginning approximately one month after planting.
Controls were grown in soil. The mutation frequencies were
continuously determined over the time of 149 days. The data are
grouped for presentation.
The mutation frequency of Tradescantia grown In sediment is
consistently higher than the controls during the five month time
period. There is also a continual decrease in the mutation
frequency of the Tradescantia in the sediment with r 2 10.589 and
b0.59 as the mutagenicity of the sediment diminishes.
2-106
-------
FREQUENCY OF STAMEN MUTATIONS OF TRADESCANTIA FROM SEPTEMBER 30 TO
OCTOBER 10 AT LOCATIONS MEASURED FROM A LEAD SMELTER (FREQUENCY +
SE) X lO2 —
Location (kin) Mutation Frequency
0.3 •9 ± 0.4
1.7 4.8 ± 004
3.2 4•4 ± 0.3
11.4 (Local Control) 3.5 ± 0.3
Laboratory Control 2.8 ± 0.3
Regression with distance of 0.3, 1.7, 3.2 and 11.4 km
b = —0.125
r 1 = 0.974
Tradescantia were planted as whole rooted plants in situ in the
native soil at the various distances down wind from the lead
smelter. All old flower stalks were removed and the new flower
stalks and inflorescences developed. The plants were in place May
29 to October 10. These data are part of a larger study, but
illustrate the in situ use of Tradescantia with significant changes
in stamen hair mutations with distance from the smelter.
FREQUENCIES OF MICRONUCLEI OF TRADESCANTIA EXPOSED TO SEWAGE SLUDGES
FROM FIVE DIFFERENT COMMUNITIES (X 10—2)
Source 0% Sludge 25% Sludge 50% Sludge p Value of Regression
1 3.0 3.6 7.6 0.005
2 3.0 4.1 11.5 0.01
3 3.0 8.3 13.5 0.005
4 3.0 8.9 32.0 0.001
5 3.0 14.0 15.0 0.005
Sewage sludges from five different municipalities were diluted and
mixed with reverse osmosed—deionized water to concentrations of 25%
and 50% sludge. Control was reverse osmosis—deionized water.
Tradescantia flower stalks were cut and the flower stalks were
2—107
-------
placed upright in the control water and sewage sludges for 24 hours
in a growth chamber under 16h llght:8h dark. At the end of the
exposure the distal tips of the flower stalks were trimmed off and
the stalks were placed in reverse osmosis—deionized water for
another 24 hours under 16h llght:8h dark. The Inflorescences were
then removed, pickeled and processed for the determination of
micronuclei. The cut flower stalks negate the root barrier and may
allow any materials that can be translocated by the vascular system
of the plant to be transported to the target tissues or cells of
interest.
The sludges all show mutagenicty and exhibit regression coefficient
significantly different from zero.
TRADESCANTIA EXPOSED TO DIESEL OBSCURANT SMOKE
GENERATED BY ARMY TANK
Distance (N) M N SN VF EP PP
15 40.2 3.6 33.2 10.4 NS
25 29.8 3.6 33.2 9.6 126.9
50 MS 2.7 34.1 10.2 68.4
Control 30.7 0.9 26.8 6.8 100.0
— micronuclel X10 2
SN stamen hair pink enents XlO
VP — variable fluorescence
EP — electron pool
PP flower production
MS not statisticaily slgnficant at p > 0.10
cut flower stalks with attached leaves were exposed in the Southern
California desert to diesel oil obscurant awoke used in military
maneuvers. The diesel is aerosolized into screening—smoke by an
attachment on the combat tank’s engine manifold. The flower stalks
were exposed for 30 mitt. at distances of 15, 25, 50 meters — and
further distances not reported on here — and measurements taken of
five biological end points. All end points showed some response to
the obscurant smoke, but only the frequencies of inicronuclel (NcN)
and stamen hair mutations (SN) showed a gradient with distance.
CONCLUSION
Tradescantia has many appealing properties that qualify It as an
organism as part of a repertoire systems for wide use in the
toxicity assessment of solid waste. Other plants with a relatively
2-108
-------
modest amount of additional development and appropriate
standardization offer great potential for solid waste toxicity
assessment both in the laboratory and in the field. Examples are:
Selenastrum , which has been developed for water pollution, but is
being developed and tested for soil contamination; Arabidopsis , a
small mustard, with a data base of over 200 chemIcals tested; seed
germination/root elongation and early seedling growth of a number of
species of plants including tomato rye, oats, cucumber, etc.; corn,
used for both laboratory and field testing; barley, laboratory
tested for more than 60 agents. To date, however, Tradescantia
offers the most versatility for bioassay.
REFERENCES
Lower, W.R., P.S. Rose, and V.K. Drobney. 1978. In situ mutagenic
and other effects associated with lead smelting. Mutat. Res.
54: 83—95.
Lower, W.R., V.K. Drobney, B.J. Aholt, and R. Politte. 1983a.
Mutagenicity of the environments in the vicinity of an oil
refinery and a petrochemical complex. Terat., Carcinog.,
Mutagen. 3:65—73.
Lower, W.R., W.A. Thompson, V.K. Drobney, and A.F. Yanders. 1983b.
Mutagenicity in the ‘vicinity of a lead smelter. Terat.,
Carcinog., Mutagen. 3:231—253.
Lower, W.R., A.G. Underbrink, A.?. Yanders, K. Roberts, T.K. Ranney,
G.T. Lombard, D.D. Hemphill, and T. Clevenger. 1984. New
methodology for assessing mutagenicity of water and water
related sediments. In: Second International Conference on
Groundwater Quality Research. N.N. Durham, and A.E. Redelfs,
eds. Oklahoma State University Printing Services Publishers:
Stillwell, OK, pp. 194—196.
Ma, T.—H. 1979. Micronuclei induced by x—rays and chemical
mutagens in meiotic pollen mother cells of Tradescantia . A
promising mutagen test system. Mutat. Res. 64, 307—313.
Ma, T.—H. D. Isbandi, S.H. Rhan, and Y.S. Tseng. 1973. Low level
of SO 2 enhanced chromatid aberrations in Tradescantia pollen
tubes and seasonal variation of the aberration rates. Mutat.
Res. 21, 93—100.
Ma, T.—H., A.H. Sparrow, L.A. Schairer, and A. F. Nauman. 1978.
Effect of 1,2—dibroinoethane (DBE) on melotic chromosomes of
Tradescantia . Mutat. Res. 58, 251—258.
2—109
-------
Ma, T.—H, V.A. Anderson, and I. Ahined. 1980. In situ monitoring of
air pollutants and screening of chemical mutagens using
Tradescantla—inicronucleus bioassay. Environ. Mutagen. 2,287.
Ma, T.—H., V. Anderson, N.M. Harris, and J.L. Bare. 1983a.
Tradescantia—micronucleus (Trad—MCN) test on the genotoxicity
of malathion. Environ. Nutagen. 5:127—137.
Ma, T.—H., W.R. Lower, F.D. Harris, J. Poku, V.A. Anderson, M.M.
Harris, and J.L. Bare. 1983b. Evaluation by the
Tradescantia—micronucleus test of the mutagenicity of internal
combustion engine exhaust fumes form diesel and diesel—soybean
oil mixed fuels. In: Short—Term Bioassays in the Analysis of
Complex Environmental Mixtures III. N. Waters, S. Sandhu, 3.
Lewtas, L. Claxton, N. Chernoff, and S. Nesnow, eds. Plenum:
New York, pp. 89—99.
Ma, T.—H, and N.M. Harris. 1985. In situ monitoring of
environmental mutagens. Hazard Assess. Chem. 4:77—105.
Ma, T.—H, N.M. Harris, V.A. Anderson, I. Ahmed, K. Mohammad, J.L.
Bare, and G. Lin. 1984. Tradescantia—micronucleus (Trad—MCN)
tests on 140 health—related agents. Mutat. Res. 138:157—167.
Nauman, C.H., A.G. Underbrink, and A.H. Sparrow. 1975. Influence
of radiation dose rate on somatic mutation induction in
Tradescantia stamen hairs. Radiation Res. 62:79.
Nauman, C.H., P.3. Klotz, and A.H. Sparrow. 1976. Dosimetry of
trltiatedl,2—dibromoethane in floral tissues of Tradescantia .
Mutat. Res. 38:406.
Sandhu, S.S. and W.R. Lower. In situ monitoring of environmental
genotoxins In Short—Term Bloassays In the Analysis of Complex
Environmental Mixtures v. Editors S.S. Sandhu and M. Waters.
Plenum Press, 1987.
Schaeffer, D.J., W.R. Lower, S. Kapila, A.F. Yanders, B. Wang, E.W.
Novak. 1986. Preliminary study of effects of military
obscurant smokes on flora and fauna during field and laboratory
exposures. US—CERL Technical Report N—86/22 1—86.
Schreiber, U., W. Vidaver, V.S. Runeckles, and V. Rosen. 1978.
Chlorophyll fluorescence assay for ozone injury in intact
plants. Plant Physiol., 61, pp. 80 — 84 .
Sparrow, A.H., L.A. Schairer, and K.M. Marimuthu. Radioblologic
studies of Tradescantla plants orbited In Biosatelilte II.
In: The Experiments of Biosatellite II. J.F. Saunders, Ed.
NASA Special Publication 204. Scientific and Technical
information Office. NASA, Washington, D.C., 1971 pp. 99—122.
2—110
-------
Sparrow, A.H., A.G. Underbrink, and H.H. Rossi. 1972. Mutations
induced in Tradescantia by small doses of x—rays and neutrons:
analyses of dose—response curves. Science 176:916.
Sparrow, A.H., L.A. Schairer, and R. Villalobos—Pietrini. 1974.
Comparison of somatic mutation rates induced in Tradescantia by
chemical and physical mutagens. Nutat. Res. 26:265.
Sparrow, A.H. and L.A. Schairer. 1976. Response of somatic
mutation frequency in Tradescantia to exposure time and
concentration of gaseous mutagens. Mutat. Res. 38:405.
Underbrink, A.G., L.A. Schairer, and A.H. Sparrow. 1973. The
biophysical properties of 3.9 GeV nitrogen ions. V.
Determination of relative biological effectiveness for somatic
mutations in Tradescantia . Radiation Res. 55:437.
Underbrink, A.G., L.A. Schairer, and A.H. Sparrow. 1973.
Tradescantia stamen hairs: a radiobiological test system
applicable to chemical mutagenesis. In: Chemical Mutagens:
Principles and Methods for Their Detection. Vol. 3, A.
Hollaender, Ed., Plenum Press. New York. pp. 171—207.
2—111
-------
STATISTICAL APPROACHES TO SCREENING HAZARDOUS
WASTE SITES FOR TOXICITY
J. H. Thomas, L. A. Athey, and J. R. Skaiski, Environmental Sciences
Department, Battelle, Pacific Northwest Laboratories, Richiand,
Washington
ABSTRACT
The kinds of questions bioassays can answer, as well as their
advantages, are given. The ability to answer questions, such as
“where is the waste?” or “Is it toxic?” results from field studies
based on twelve procedural steps for conducting remedial action
field bioassay studies. Simple or stratified random, systematic,
and judgment sampling are presented as three possible paradigms for
field sampling.
Bioasay results from two field studies are used to illustrate how
maps of toxicity can be prepared based on systematic sampling and to
show how cleanup decisions can be made using bioassay results based
on few samples.
In the first study, logarithmically spaced soil samples (0—15 and
15—30 cm depths) were obtained along four parallel transects (90 m
long and 15 m apart) at the Rocky Mountain Arsenal. A total of 72
soil samples (36 at each of two depths) were subjected to phytoassay
using lettuce seeds; most samples were also subjected to Daphnia,
Microtox, algal, earthworm and lettuce root elongation bioassays.
These latter bioassay results (exception earthworms) were
inconclusive regarding toxicity, but allowed us to ignore several
classes of compounds (i.e, water—soluble heavy metals, herbicides,
and insecticides) since our prior results using pure chemicals
showed depressed algal growth in the presence of these contaminants.
In order to depict the spatial pattern of observed seed mortality at
each depth, we used kriging (a statistical technique developed for
use in the mining industry) to produce contour maps. The results
clearly showed that lettuce seed mortality was higher in the 15—30
cm fraction, that waste—trench soil was highly phytotoxic, and that
toxicity decreased as a function of distance from the trench. In
addition, we found that mortality contours produced by kriging could
be useful in site cleanup decisions.
The second study was conducted using a series of water and sediment
samples collected from a narrow stream adjacent to a wood treatment
plant in Canton, Mississippi. Both creosote and pentachiorophenol
were used for wood treatment. Sediment samples (15 cm) were
collected every 20 iii in the visibly contaminated zones. Based on
sample linear interpolation of bioassay results, we found that
different bloassays led to different conclusions regarding the
2—113
-------
toxicity of different areas, suggesting that contaminants other than
creosote may have caused the observed toxicity. Moreover, chemical
analysis was an inaccurate predictor of toxicity.
INTRODUCTION
In the context of hazardous chemical waste site management,
bioassays nay be defined as the exposure of biological indicators to
field—collected environmental samples in order to detect the
presence of toxicity and/or to identify potential for toxic effects
on resident species. Typically, a hazardous waste site bioassay
involves laboratory testing of soil, soil leachates, water, or
sediment samples, using a standard array of test organisms under
controlled laboratory conditions.
Bioassay studies are appropriate before, during, and after remedial
action as a cO8t—effective way to detect the presence of toxic
wastes and/or to determine their biological availability at known or
suspected hazardous chemical waste sites.
Bioassay studies may be used before remedial action to detect the
presence of hazardous materials and to determine If immediate
remedial action is needed. Because bioassays directly evaluate the
effect of chemical wastes on blota, they are powerful and efficient
tools for ranking sites requiring remedial attention. Bioassay data
are also useful for gauging the areal extent of needed remedial
action, evaluating remedial action alternatives, and assessing and
characterizing waste sites.
During remedial action, bioassay studies may be used both to monitor
the cleanup process and to evaluate cleanup impacts on the site.
Bioasaaya are also useful after remedial action has been taken to
evaluate the efficacy of the cleanup activities.
The questions that often need to be addressed in support of remedial
action decisions may be grouped Into four major areas: 1) Where are
the contamInants? 2) Are the contaminants toxic ? 3) What
quantities of the contaminants are present? and 4) What are the
contaminants?
Bioassay studies are a cost—effective method for evaluating “where”
questions because they can detect contamination, determine
contaminant distributions, and define contaminant migration beyond
site boundaries. Bloassays are also currently the only means of
evaluating the bloactivity of chemical wastes. However, bioassay
studies are not generally useful in answering “how much” and “what”
questions regarding hazardous chemical waste sites. Further,
bioassay tests usually cannot identify the specific chemical
contaminants present. However, they may give some indication of the
class of chemicals to which the contaminants belong (e.g., organics,
2—114
-------
metals). Some examples in which the identification of specific
chemicals have been attempted are given Thomas et al. (1986) and
Miller et al. (1985).
DESIGNING BIOASSAY STUDIES
Design of bioassay studies includes planning the collection of field
samples and the laboratory analyses of those samples. It is
imperative that the entire project, from objectives to expected
results, be thoroughly planned before the actual fieldwork is
started. Without proper planning, the study will waste both time
and resources. Further, all individuals who will contribute to the
project should be involved as early as possible in project
planning. Personnel who should be involved at the initial stages of
the project include the project manager, a statistician, a field
biologist, a chemist, the scientist who will oversee the laboratory
work, possibly a hydrologist, meteorologist, or modeler when
appropriate, and risk assessment and Quality Assurance/Quality
Control experts.
The steps to be used in designing and executing a bioassay study are
listed below.
1. Assemble information relevant to the problem.
2. Prepare a statement of the study objectives.
3. Define the evaluation criteria and reliability requirements
for the results.
4. Determine what is to be sampled in the field.
5. Choose test organisms for the bloassays.
6. Define the data analysis techniques.
7. Design the field sampling and laboratory studies.
8. Determine the sample collection methods.
9. Define the operational procedures.
10. Review the design.
11. Periodically evaluate progress in the field and laboratory.
12. Analyze and evaluate the results.
These steps are listed in the order in which they should be applied
by the planning team. However, study design is an iterative
2—115
-------
process, and decisions made at later steps in the process nay
require the review and revision of decisions made at previous steps.
These 12 steps should be considered whether the study Is an initial
site Investigation, a feasibility study, or a full remedial action.
The amount of effort and degree of professionalism required for each
step will vary depending on the objectives and cost of the study and
reliability required of the results. For example, in a full and
potentially expensive remedial action investigation, the entire team
should be involved from the beginning. During preliminary
Investigations, the objectives, uses of the results, and methods
should be carefully planned before sampling; however, the advice of
team members from each area of expertise may not be necessary.
Depending on the quality of information gathered during the
preliminary investigations, the results of the initial studies may
be used to design the remedial action program.
SAMPLING STRATEGIES FOR COLLECTING FIELD SAMPLES
The three basic design strategies for collecting field samples are
simple or stratified random, systematic, and judgment sampling.
These strategies are described in Ford and Turina (1985) and are
discussed here only as they apply to bioa8say studies at hazardous
chemical waste sites.
All remedial action decisions for the hazardous waste site will be
based on the results obtained from samples collected on the site.
Therefore, It is important that the samples obtained accurately
represent the conditions on the site. The traditional approach to
collecting a representative sample is to randomly select the
sampling locations over the entire site with the aid of a random
number table or similar device. This procedure Is called simple
random sampling and is an appropriate strategy when no prior
information Is available on the likely location or distribution of
the contaminant.
With stratif ted random sampling, sampling sites are randomly chosen
within several defined site areas or strata. Appropriate strata for
a hazardous chemical waste site might be any division of areas in
which It is anticipated that toxicity will differ; for example,
areas at increasing distance from the known or suspected source of
chemical contamination. Stratified random sampling is often a
useful technique even if there Is Insufficient Information available
to identify distinct strata a priori, and is likely to produce a
more widespread distribution of sampling locations on the site than
will a simple random sampling strategy. The construction of
arbitrary stata can allow the variability In toxicity between
different areas of the site to be estimated and may help Identify
actual strata. For these reasons, stratified random sampling is a
particularly useful approach for preliminary studies.
2—116
-------
Systematic sampling involves the collection of samples at regular
intervals over the site. This method is sometimes preferred to
random sampling strategies because it ensures even coverage of the
site. However, as Eberhardt and Thomas (1986) warn, a
systematically distributed contaminant may not be detected when
using this strategy. Systematic sampling (often in grids) is often
the preferred method to provide input data for kriging and other
mapping techniques.
Judgment sampling relies on the sampler’s judgment of what
constitutes a representative site sample. The purpose of these
samples might be to assess the presence or absence of contaminants
in obvious places (e.g., a streambed if transport is a concern), or
for use in special studies of a preliminary nature (e.g., are toxic
chemicals nearer the spill source?). However, judgment sampling is
biased and such results should not be used in a statistical
analysis. In addition, judgment samples can be collected along with
samples from the designed study in the event that unusual or
interesting circumstances arise or are discovered during sampling.
A statistical evaluation of the data from the judgment samples can
be used to suggest additional sampling or to make statements without
accompanying error statements or probablistic assertions. In cases
where prior information about a possible 8pill location becomes
available during sampling, that information should be used to
advantage in the survey design. In fact, design modifications can
be made onsite (e.g., an extra grid, transect, or stratum).
Locating field sampling sites in order to implement any of the
sampling designs discussed is not a trivial matter. At least
one—third of the field effort should be devoted to locating and
marking the sampling sites. Each sample location should be
accurately recorded to aid in the interpretation of the bioassay
results, to accurately define areas in which remedial action may be
necessary, and to permit return to any sampling site to collect
necessary additional material.
FIELD STUDIES: ILLUSTRATIONS OF BIOASSAY NETHODS AND DESIGN
DECISIONS
The two field studies in this paper were selected to illustrate the
steps in conducting a bioassay study, and the statistical principles
involved. The studies were designed to evaluate toxicity and define
areal extent, if toxicity was detected. At two New Jersey sites not
discussed in this paper, extra samples were collected and some
coinposited, [ see Skaiski and Thomas (1984) for a discussion of
compositing]. Field sampling and the use of bioassay results from a
systematically sampled grid at the Rocky Mountain Arsenal (Commerce
City, Colorado) are illustrated in site 1, while the more usual case
where inferences are based on a few bioassayed samples form the
2—117
-------
basis for a discussion of results from site 2 [ a wood treatment site
(Canton, Mississippi)].
SAMPLING AT THE ROCKY MOUNTAIN ARSENAL
The Rocky Mountain Arsenal was used principally by the U.S. Army and
the Sheil Chemical Company to manufacture, test, and dispose of
toxic chemicals. Certain areas of the 26—mile 2 (67.6-k it 2 ) arsenal
have been contaminated by various spills during waste disposal
operations. The site is surrounded by homes, farms, and businesses
(e.g., Stapleton Airport, Commerce City, and Denver). In 1974
diisopropylaethylphosphonate (DIMP) and dicyclopentadiene (DCPD)
were detected in the surface water draining from a manmade bog at
the northern boundary of the arsenal.
Study objective 1 at the arsenal was to assess the toxicity of a
trench site in Basin A (Figure 1). If one or more bloassays
identified to devise a contour map useful for cleanup decisions
(objectives 2). Finally, an assessment of contaminant mobility was
needed (objective 3). To meet those objectives, a toxic site at the
Rocky Mountain Arsenal had to be located and the most sensitive
bioassay selected.
SITE CONTAMINATION
Decontamination wastes and process waste streams containing toxic
materials consisting of salts, heavy metals, and pesticides were
deposited in defined areas on the Rocky Mountain Arsenal. Two waste
basins (A and F), were the major sites of waste material storage.
Certain portions of the arsenal were leased to private industry for
chemical manufacturing. A major chemical company leased a
considerable portion of the manufacturing facilities at the Rocky
Mountain Arsenal since 1952. Alterations and additions were made to
the facilities for the manufacture and disposal of waste residuals
of GB, a chemical warfare agent, TX, a biological anticrop agent,
and cyanogen chloride and phosgene. Since 1970, several major
chemical demilitarization actions have been conducted. These
actions include the incineration of both the anticrop agent TX and a
mustard agent.
Because of the foregoing manufacturing and disposal actions, various
analyses of chemicals in air, water, soil, and certain organisms
have been conducted over the years. Table 1 contains a partial list
of some of the chemicals identified.
DESIGN OF THE FIELD SAMPLING AND LABORATORY STUDIES
Previous results indicated that an area in Basin A was toxic
(objective 1) and would be useful for addressing the second study
2-118
-------
objective. Soils from the Basin A location caused a major reduction
in lettuce seed germination (insoluble compounds were likely
causative) and had a variable effect on other bioassays (little
likelihood of soluble compounds; objective 3). Because of these
results and, in part, because it appeared that plant growth
diminished with distance from the trench, a sampling location was
established near the trench In Basin A (to meet objective 2).
The results also suggested a possible gradient of contamination on
the west side of Section 36, extending north—south from a trench
that drains Basin A and runs to the west. This possibility of a
toxicity gradient offered good prospects for the kriging method of
preparing contour maps because it appeared that a required, physical
TABLE 1 . Examples of Some Chemicals Found In Soils, Air,
Water, Animals, and Plants at the Rocky
Mountain Arsenal
Aidrin Methyiphosphonic acid
Arsenic Compounds Isopropyl methyiphoaphonate
Benzene Diisopropyl methyiphosphonate
Chiordane Dicyclopentadiene
Chloroform p—Chlorophenyl methyl suif one
Dieldrin Hexachioronorboradiene
Endrin Tetrachioroethylene
Lewisite 1,4—Thiozane
Lewisite oxide Methylene chloride
Mercury salts Toluene
Mustard gas CUD) Xylene
Thiodiglycol
mechanism of toxicant dispersal (Journal 1984) for valid error
predictions was a reasonable assumption. Thus, four parallel
transects were established on the west side of Basin A, each
beginning on the north bank of the trench and running south for 90 in
(Figure 2). A logarithmic scale was used beyond the south trench
edge because it was assumed that contamination might have been moved
by some physical means (e.g., wind or water). The transects were 15
in apart and labeled L, N, N, and P. The first three sample points
of each transect fell within the trench and the fourth was on top of
the south bank. Sample numbers 5 through 9 were 15, 20, 30, 50, and
90 in, respectively, south of the north trench edge (Figure 2). Each
of the 36 sampling points was marked with a stake.
A drilling company was hired to do most of the soil sampling. At
each sampling point a split spoon was used to take two
7.6—cm—diameter soil cores. One core was taken from a 0— to 15—cm
depth, and the second was taken from a 15— to 30—cm depth.
Together, these cores weighed approximately 4 kg. Between sampling
points the split spoon and drill bit were decontaminated by washing
2—119
-------
U,
w
FIGURE 1 . Location of the Study Site in Basin A (which includes most of
Section 36) at the Rocky Mountain Arsenal. The areas in
Section 26 labeled C, D, E, and F are or were waste ponds.
C ’,
w
STAPLETON
INTERNATIONAL
AIRPORT
2—120
-------
-z
0
z
E
0
E
U
C
FIGURE 2 . Location of Logarithmic Sampling Points in Basin A (Section 36)
at the Rocky Mountain Arsenal
Transect Letter
2-12 1
-------
with methanol and rinsing with distilled water. All samples were
put in plastic bags, sealed, and labeled. The area being sampled
and any problems encountered (e.g., mud, accessibility) dictated
exactly how the cores were taken and the variations on the basic
sampling scheme (see below).
In Basin A, the first two points in each transect were in the
trench, which was very wet and soft. The samples from these points
were difficult to obtain. It was impossible to sample by depth, so
a hand trowel was used to take two surface samples (to approximately
15 cm deep) from these points. The split spoon was used to take
moat of the other samples in Basin A (points 4 through 9 in
transects L, N, N, and P). Sample points L, N, and P—3 were just
over the south bank of the trench and could not be reached from the
drill rig. At these points, the split spoon was hammered into the
ground and extracted by hand. Only a 0— to 15—cm sample was
obtained from each of these points; the soil from 15 to 30 cm deep
was to wet to stay in the split spoon. A surface sample of
undefined depth was taken from the 1 1—3 transect because the entire
profile was very wet.
Analysis and Evaluation of the Results
Lettuce seeds were used in the bloassays of the Rocky Mountain
Arsenal field study soils, because they are more sensitive than
other seeds (Thomas et al. 1986) and our previous work has shown
these aoil to be phytotoxic. The mortalities (mean from three
subsamplea from each core fraction; Figures 3 and 4) indicate that
four samples in transect N and three in transect P showed large
differences as a function of depth, suggesting that the contaminants
had either migrated below 15 cm or were purposely placed there. We
found no record to support the latter argument and concluded that
the toxic material had migrated.
The results for all other bloassaya (Porcella 1983) conducted using
Basin A field study soils were inconclusive. No mortality was
observed using the Daphaia or Microtox bloassays. Results from the
algal bioassay revealed that small quantities of elutriate from all
but four samples were stin*ulatory. The four elutriate samples that
Inhibited algae were obtained on or very near the waste trench on
transects L and N. According to the criteria outlined In Porcella
(1983), these sites would be classified as moderately toxic.
Interestingly, only a 1% to l4 elutriate from transect N samples
was needed to stimulate algal growth. Because there was little
effect on the algal bioassay, It appears that the toxic components
detected using lettuce seeds (see Figures 3 and 4) were not
water—soluble heavy metals, herbicides, or insecticides (except
perhaps at sites L—2, L —4, and N—2), since results using pure
chemicals (Miller et al. 1985) showed depressed algal growth in the
presence of these contaminants. This evidence suggests the presence
2—122
-------
UQ
Efl
10 - 15 — 11 —— — 10 — 4)
13 19 14—5
30—U 1 8 11 -7
I
40
6 7 17 14—8
w
z ’J
D
0
U60
w
U
z -
C l )
0
80 -
15m
Is
90 22 14i 9
p N M L
TRANSECT LETTER
FIGURE 3 . Observed Mean Lettuce Seed Mortality at Each Basin A
Plot, 0— to 15—cm Soil Fraction. Means enclosed with a box
are >75%, an enclosing circle indicates means from 30% to 75%,
and numbers with no symbol are means <30%.
2—123
-------
__ I J
J-i
E 1
r IJ fl....,3
10 - —-————- --— J— 22 -
( : _ 9 16 12—5
20 -E J
4O- 10 29 2O-8
5o.
0 z
I-
6O 0
w
U
2
I-
U)
0
80
15m
Ii l I
—- 15 9
p - N M L
TRANSECT LETIER
FIGURE 4 . Observed Mean Lettuce Seed Mortality at Each Basin A
Plot, 15— to 30-cm Soil Fraction. Means enclosed with a
box are >75%, an enclosing circle indicates means from
30% to 75%, and numbers with no symbol are means <30%.
2-124
-------
of water—insoluble contaminants that are not likely to migrate
(objective 3).
Except for plot P—3, earthworms were only affected by Basin A soil
when lettuce seed mortalities were greater than 70%. In contrast,
the five soil samples that caused 20% to 70% lettuce seed mortality
resulted in no earthworm deaths. Thus, for these Basin A samples,
it appears that lettuce seeds are more sensitive to lower levels of
an insoluble toxic component than are earthworms (objective 3).
Result8 from the lettuce root elongation (based on elutriates) and
the lettuce seed mortality (based on intact soil) bioassays show
only a slight correspondence for samples from transect L. It
appears that the phytotoxic component that Impairs lettuce seed
germination may not be water soluble (objective 3) or does not
affect root elongation.
Cleanup Decisions Based on Bioassays and Kriglng
One way to depict the lettuce seed mortality patterns observed at
each depth (see Figures 3 and 4) is to prepare a contour map based
on the observations (objective 2). We estimated contours for a map
of lettuce seed mortality using a relatively new 8tatistical
technique called kriging, which was developed for use In the mining
Industry and Is used principally in Europe and South Africa (Clarke
1979). KrIging provides a variance estimate that can be used to
construct a confidence interval for the true value. Results based
on block kriging are presented in Figures 5 and 6. The results
clearly show the lettuce seed mortality differences at the two
depths. Estimated contamination is greater from 15 to 30 cm deep
than from 0 to 15 cm deep. This contamination difference was also
indicated by the qualitative analyses of results (see Figures 3
and 4)
Contour maps are useful for making site cleanup decisions. As a
scenario, we selected 30% lettuce mortality as a criterion for
cleanup of the Basin A site (see Figures 5 and 6). In the absence
of any other guidance, the cleanup criterion was selected as two
8tandard deviations above the mean control mortality (i.e., 16.7
+14.0, ii 6). The shaded areas below the heavy solid black lines
would be targeted for cleanup. The cleanup decision would be
different for the 0— to 15—cm—deep (Figure 5) and the 15— to
30—cm—deep fractions (Figure 6). hi1e this difference complicates
decision making, the available data and the kriging maps show that
the field situation Is complex, and cleanup decisions based solely
on either soil fraction would not result in a “clean” site. Cleanup
using the 30% mortality contour of the 15— to 30—cm samples would
remove all known contamination, but additional samples taken below
this depth would be needed to ensure that the site meets the 30%
mortality cleanup criterion.
2—125
-------
90
DISTANCE (m) FROM NORTHEAST CORNER
FIGURE 5 .
Estimated Lettuce Seed Mortality (based on kriging) for
the 0- to 15-cm Soil Fraction from the Rocky Mountain
Arsenal
2-126
80
70
60
z
w
z
2
-7-
c i
0
z
0
E
U i
U
z
I-
(1,
C
40
30
20
o
0 15 30 45
DITCH
-------
FIGURE 6 .
Estimated Lettuce Seed Mortality (based on kriging)
for the 15— to 30-cm Soil Fraction from the Rocky
Mountain Arsenal
2-127
>.
-J
2
z
z
I
U
I —
a
-a-
0
z
0
LL.
E
w
U
2
U)
C
0 15 30 45
DISTANCE (m) FROM NORTHEAST CORNER
-------
It appears that bioassays of field samples and subsequent krlglng
analyses (objective 2) offer a practical method to aid In cleanup
decisions based on environmental toxicity, especially when
accompanied by error estimates for the mortality isopleths. We did
not present confidence limits here because of some questions about
the possibility that the observed toxicity was caused by pollutants
that were “dumped TM rather than spread from the trench by wind or
water. Journal (1984) argues that the confidence limits may be
invalid unless movement is caused by physical forces.
The limits in our study average between ±10% and 25%, and depended
on data density, whether block or point kriglng was used, and the
contour of concern.
SAMPLING AT A WOOD TREATMENT SITE
The wood treatment site in Mississippi ceased operation in 1979 and
was known to be contaminated with creosote and other wood—preserving
materials. The results discussed below are from an exploratory
study to determine whether bloassays can be used to detect creosote
contamination in stream sediments and water, and if feasible, to
define the boundaries of creosote—contaminated zones on the site.
Samples were also subjected to chemical analyses. The objectives of
the study were to 1) determine if standard bloassays could detect
creosote contamination in water and sediments, and, if so, 2) map
the distribution of creosote contamination in the creek.
Site Description
A preliminary site visit was made to obtain background information
on the history of creosote disposal, to determine the dimensions of
the site, and to define any special sampling problems. The site is
bounded on the south by a city street (Covlngton Avenue) and on the
east and north by a creek. The creek is approximately 2 m wide at
the widest point, with 2—rn—high banks on either side. The creek
flows northwest (on the site), enters an open concrete channel at
the western site boundary, and then flows Into a nearby city.
Records obtained from the Mississippi Department of Natural
Resources (Bureau of Pollution Control) indicate that creosote and
occasionally pentachiorophenol were used for wood treatment on the
site. From 1965 to 1979, the owner of the site permitted wastes
from the treatment process to flow overland to the creek, in
violation of state pollution laws. Little cleanup had been done
when the site closed in 1979. However, it was clear from the site
visit that the site had been covered with fill material. Creosote
was still being emitted from the bank into the water along some
parts of the stream. Piles of creosote—contaminated material, as
well as pools of black sludge, were located Immediately adjacent to
some old tanks on the south side of the site.
2-128
-------
The primary interest in this study was creosote contamination of
creek sediments. Therefore, stream sediment samples comprised the
bulk of the samples taken. A small number of water samples were
taken, as well as some samples from the bank where creosote appeared
to be entering the stream. In addition, samples were taken from an
upstream site (negative control) and from the pile of
creosote—contaminated sludge (positive control).
A description of the spatial distribution of the contamination was
one objective of the study. Kriging was the first choice as a data
analysis technique because it permits generation of confidence
intervals about estimates of areal distribution. However, kriging
generally requires a large number of data points, and in the absence
of a sufficient amount of data to perform krlging, a simple linear
interpolation could be substituted (principally because the area of
concern was a narrow stream channel).
Design of the Field Sampling and Laboratory Studies
The field sampling scheme is diagrammed in Figure 7. The girder
bridge was used as the staging area and the starting point from
which distances to each sampling location was measured. One
sediment sample was collected at the point just before the stream
entered the concrete channel at the western end of the site, 660 m
west of the initial sampling location at the girder bridge. The
next sample was collected at 420 m west of the initial location,
then every 40 m to the east until the visibly contaminated zone of
the stream was reached. Samples were collected every 20 m in the
visibly contaminated zone, and beyond until a location 220 m east of
the starting point was reached. Three tributaries (Eastern,
Western, and Northern on Figure 7) drain into the creek. The last
sample (220 m east of the Initial point) was collected upstream of
the three tributaries. In addition, one composite sample was
collected from each of the three tributaries. The negative control
sediment sample was taken from the creek south of Covlngton Avenue,
upstream from the site. The positive control sample was taken from
the piles of creosote—contaminated sludge near the storage tanks.
This sludge was believed to be the same material that was visibly
contaminating the stream.
Water samples were collected at 660 m west of the initial point
(farthest downstream location), 380 m west of the initial point,
220 m east of the initial point (farthest upstream location), and
south of Covlngton Avenue, where the negative control sediment
sample was taken (Figure 7). All samples were taken on the same day
to maximize the comparability of bioassay results and to minimize
the sampling costs [ samples were collected from west to east
(downstream to upstream) to minimize cross—contamination of
samples). The laboratory analyses were completed in two phases. In
phase 1, only sediment samples from 660, 380, and 20 m west of the
2-129
-------
Northern Girder Old
Tributary Bridge Railroad
Bridge
t1!I1iI11lI1l1It1I1III111i11I1 11l1i1ll!1III11 lMhIIl1I lI!I1l1!IUi1
160 80 40 0 40 80 160
, , , , + ,
660 420 340 260 180 100 20 20 100 18C 220
380 300 220 140 60 60 140 A
I A _________________
Visibly
Western Creek Flow Contaminated
TributarvJ Area
1 -‘
I Eastern
I Tributary
I Storage
I Tanks
0
0.
Covington Avenue
• Soil or Sediment Sample
A Water Sample
FIGURE 7 . Location of Samples Collected from the Wood Treatment Site in
Mississippi (distances are in meters). Soil or sediment
samples are indicated by solid circles and water samples with
solid triangles.
initial point, 120 and 220 m east of the initial point, the positive
and negative controls, the composite sample from the eastern
tributary, and the water samples were analyzed. The results of
these bioassay tests were used to bracket the contaminated zone. In
phase 2, samples from 300, 140, and 60 a west of the initial point
were analyzed to aid in defining the contaminant boundaries.
Because the clay sediments lining the creek were very bard, they
were sampled using a hand coring device. Where possible, surface
sediment samples were collected to a depth of 15 cm with hand
trowels.
Analysis and Evaluation of the Results
The results of bioassay analyses of phase 1 samples are shown in
Table 2. The only locations where appreciable toxicity occurred
were the positive control and locations 660 and 380 rn west of the
initial sample location. The water sample from the 380—rn—west
location was highly toxic, while the sediments from that location
were less toxic. At 660 in west, however, the sediments were highly
toxic to some organisms while the water was not toxic.
2—130
-------
Figure 8 contains toxicity maps of creek sediment elutriates from
660 m west to 220 in east of the initial point. These maps were
based on bioassay and chemical analyses from both phase 1 and phase
2 samples (distances to the west of the initial point on Figure 8
are indicated by negative numbers). The respective creosote
concentrations determined for each sample by IR (infrared
spectroscopy) are expressed as a percentage of the highest creosote
value measured (9500 and 25 ppm for sediments and sediment
elutriates, respectively). We note that creosote Is a complex
mixture of organic compounds. Since there were insufficient data to
use the kriging technique to devise maps, the maps were prepared
using simple linear interpolation of the results between the
sampling points. Therefore, the precision of the division locations
between different zones of contaminant concentrations cannot be
estimated (as in kriging).
The top four bar8 on Figure 8 illustrate EC5Os from the algal,
Daphnia , Microtox, and root elongation bioasaays in which the test
materials were sediment elutriates. The result from different
elutriate bioassays led to different conclusions regarding the
relative toxicities of different areas of stream sediments. Such
variable biological responses could result from different organic
compounds in creosote, which may bind differentially in each area of
stream sediments, or from In—stream seeps from the waste site. IR
analyses for creosote indicated that the most severe contamination
occurred in the extreme downstream portion of the creek study area.
In contrast, the algae, Daphnia , and Nicrotox bioassays indicated
that the most extreme toxicities actually occur about 300 m west of
the initial point. The Microtox bioassay was most sensitive to the
chemical contaminants in the downstream sediment elutriates. The
results from root elongation tests show a complete absence of a
detectable phytotoxic component. A comparison between relative
creosote amounts and algae, Daphnla , and l4icrotox response to
sediment elutriates collected between —140 and —400 in (Figure 8)
suggests that contaminants other than creosote caused the toxicity.
CONCLUSIONS
Several conclusions are possible based on results from this study.
First, standard bioassay organisms are sensitive to contaminants
resulting from wood treatment operations, and different bioassay
organisms have different sensitivities to the mixture of creosote
resulting from wood treatment operations. Second, infrared
measurements of sediment contaminants resulting from wood treatment
operations are inaccurate predictors of biotoxicity. Finally,
bioassay results can be mapped using kriging and these maps offer
good prospects for aiding in regulatory decisions.
2-131
-------
TABLE 2 . Bioassay Results from Phase 1 Samples Collected from the Wood Treatment
Plant in Mississippi
EC50
C) Root
Sample Location Sample Type ALgae pap i .i Microt . E1ongapn Neubauer Earthworm
660 m west Sediment 637 73.6 k.0 100 70.3 27.9
(c)
Water NE NE NE NE NE NE
380 m west Sediment 73.1 NE 29.9 100 100 27.9
Water 6.6 0.2 9.6 100 NR NR
20 rn west Sediment NE NE NE NE NE
120 m east Sediment NE NE NE NE NE NE
220 m east Sediment NE NE NE NE NE NE
Water NE NE NE NE NE NE
Negative Control Sediment NE NE NE NE NE NE
Water NE NE NE NE NE NE
Positive Control Sediment 0.6 69 8.5 8.1 0.9 3.9
Eastern Tributary Sediment NE NE NE NE NE NE
(a) Tests conducted with sediment elutriates.
(b) Tests conducted with sediment samples.
(c) NE No effect.
(d) NR = Bioassay not required.
-------
Bioassy %
(least toxic) IR%
75-NE 0-25
N 50-75 L 3 25-50
25-50 50-75
0-25 75-100
Selanastrum (EC5O) (most toxic)
Daphriia (EC5O)
Ii I
Microtox (EC5O)
Root Elongation (EC5O)
Creosote
FlU
i i’/ ///r//
f/ I//I/I
/1/1/71/
/11/
I/i//I
////////
/71
L
I
I
I
I
!
I!
—660 —380 —300 —140 —60 —20 120 140 220
Meters
FIGURE 8 . Bioassay Results from Sediment Elutriates at the Wood
Treatment Site in Mississippi. Negative numbers
represent samples collected downstream from the site.
2-133
-------
REFERENCES
clark, I. 1979. Practical Ceostatistics . Applied Science, London.
Eberhardt, L. L., and J. N. Thomas. 1986. Survey of Statistical
and Sampling Needs for Environmental Monitoring of Commercial
Low—Level Radioactive Waste Disposal Facilities . NUR.EG/CR4162,
U.S. Nuclear Regulatory Commission, Washington, D. C.
Ford, P. J. , and P. J. Turina. 1985. Characterization of
Hazardous Waste Sites—A Methods Manual, Volume I: Site
Investigations . EPA/60014—84--075, Office of Advance Monitoring
Systems Division, Las Vegas, Nevada.
Journal, A. C. 1984. “New Ways of Asse8sing Spatial Distribution
of Pollutants.” In Environmental Sampling for Hazardous Wastes ,
eds. G. E. Schweitzer and J. A. Santolucito. American Chemical
Society Symposium Series 267, pp. 109—118, American Chemical
Society, Washington, D.C.
Miller, W. E., S. A. Peterson, and C. A. Callahan. 1985.
“Comparative Toxicology of Laboratory Organisms for Assessing
Hazardous Waste Sites.” J. Environ. Qual . 14:569—574.
Porcella, D. B. 1983. Protocol for Bloasseasment of Hazardous
Waste Site.8 . EPA 600/2—83-054, Corvallis Environmental Research
Lab, Corvallis, Oregon.
Skaiski, J. K. and J. N. Thomas. 1984. Improved Field Sampling
Design and Compositing Schemes for Cost—Effective Detection of
Migration and Spills at Commercial Low—Level Radioactive or
Chemical Waste Sites . PNL—4935, Pacific Northwest Laboratory,
Richiand, Washington.
Thomas, J.M., J. a. Skalski, J. F. Cline, N. C. McShane, J. C.
Simpson, W. E. Miller, S. A. Peterson, C. A. Callahan, and J. C.
Greene. 1986. “Characterization of Chemical Waste Site
Contamination and Determination of Its Extent Using Bioassays.
Environ. Toxicol. Chea . 5:487—501.
2—134
-------
APPLICATION OF A BIOMARKERS-WASTE CHARACTERIZATION APPROACH
TO THE PREDICTION OF ORG1 NISM RESPONSES FOLL ING EXPOSURE TO
CONTAMINATED MARINE SEDIMENTS
G. G. Pesch, A. R. Malcolm, G. R. Gardner, U. S. Environmental
Protection Agency, Narragansett, Rhode Island, and L. Mills, C.
Mueller, R. Pruell, Science Applications International Corporation,
Narragansett, Rhode Island, and A. Senecal, University of Rhode
Island, Kingston, Rhode Island.
ABSTRACT
This paper reports on the application of short—term biomarkers to
predict organism responses induced by contaminated marine sediments.
Laboratory results were confirmed by observing the same responses to
the same sediments in both laboratory— and field—exposed animals.
Identification of causative agents is being investigated by a
combination of sediment chemical analysis and short—term testing of
fractionated solvent extracts of sediments. Whole extracts of a
contaminated sediment induced concentration—dependent responses in the
Salmonella/inicrosome (Ames) test with strain TA100 in the presence of
an exogenous (S—9) metabolizing system. Increased frequencies of
sister chromatid exchange (SCE) were observed in a marine worm,
Nephtys incisa, exposed to whole sediment in laboratory tests.
Comparable increases in SCE frequency were observed in the same
species sampled from feral populations at a field disposal site
exposed to this same sediment. These responses suggest a tumorigenic
potential for the sediment. This potential has been realized by
induced tumor development in American oysters, Crassostrea virginica,
and winter flounder, Pseudopleuronectes axnericanus, exposed to this
sediment in the laboratory and the field. Partial chemical
characterization of this sediment has identified many organic and
inorganic compounds reported to be carcinogenic, cocarcinogenic,
genotoxic and tumor—promoting. Preliminary tests with fractionated
solvent extracts indicate mutagenic activity (Ames test) to be
primarily associated with fraction 2, a PAH fraction. Effects on
cell—to—cell communication between cultured mammalian cells, a
potential biomarker for tumor promoters, suggest the presence of
tumor—promoting agents in fraction 4 (a highly polar fraction). Other
biomarkers (enzyme inhibition, fertilization impairment, and
macrophage repression) are being explored as short—term tests for
characterizing complex wastes and predicting organism responses.
2-135
-------
ENVIRONMENTAL MONrIORING
Ossi Meyn, Technical Assessment Branch, Office of Solid Waste, U.S.
Environmental Protection Agency, Washington, D. C.
ABSTRACT
Throughout the world are field scientists who are in a unique position
to see subtle changes in their object of study and their environment.
Field scientists keep extensive field notebooks, but seldom
communicate findings such as these, unless and until they have
relevance to one of their projects. The object of this discussion is
the compilations of these findings into a database related to putative
man—made environmental degradation. Some of the uses of such a
database are obvious and I will discuss others. A shotgun collection
of data would be unproductive. Therefore initial studies will be
concentrated in coastal areas.
When proximity of hazardous waste sites was investigated it was found
that many sites on the National Priority List are contaminating marine
waters and sediments via contaminated surface runoff or ground water.
However, many more sites posing severe environmental hazards are not
listed as Superfund sites because they have not been shown to
represent direct threats to human health, thus have low priority for
corrective action. Toxic agents of concern at these sites include
pesticides, metals, hydrocarbons, PCB’s and others.
Possible questions that may arise will be discussed, among them: 1)
Relative sensitivity to toxic substances of bottom dwelling and
pelagic organisms; 2) normal and pathological concentrations of
metals in fish and sea mammals; 3) quantitative studies of the
relationship between sewage outfalls and plankton blooms; and 4)
normal cycles of plant or animal populations.
The ideal computer for this would be an IBM PC—AT or PC compatible,
because of the available generic software. Data collection could be
implemented, using existing electronic mail and “bulletin board”
packages. Access to these media is available to scientists across the
country. The proposed database should be either a relational
database, such as Oracle or a collection of related databases such as
Database III.
2-137
-------
ASSESSMENT OF THE MICROSCREEN PHAGE-INDUCTION
ASSAY FOR SCREENING HAZARDOUS WASTES
Virginia Stewart Houk and David M. DeMarini, Genetic Toxicology
Division, Health Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina 27711
ABSTRACT
The Microscreen phage—induction assay, which quantitatively measures
the induction of prophage in Escherichia coil WP2 CX), was used to
test 14 crude (unfractionated hazardous industrialSwaste samples for
genotoxic activity in the presence and absence of metabolic activation.
Eleven of the 14 wastes induced prophage, and induction was observed at
concentrations as low as 0.4 picograms per ml. Comparisons between the
mutagenicity of these waste samples in Salmonella and their ability to
induce prophase ) indicate that the Microscreen phage—induction assay
detected genotoxic activity in all but one of the wastes that were
mutagenic in Salmonella. Moreover, the Microscreen assay detected as
genotoxic 5 additional wastes that were not detected in the Salmonella
assay. Partial chemical characterizations of the wastes showed high
concentrations of carcinogenic metals, solvents, and chlorinated
compounds, most of which are detected poorly by the Salmonella as€ay.
However, recent studies of the induction of prophage by these chemical
classes have suggested that phage induction may be a sensitive endpoint
for these groups of chemicals. This may explain the enhanced ability
of the Microscreen phage—induction assay to detect genotoxic activity
in 5 additional wasted compared to the Salmonella assay. The
applicability of the Microscreen phage-induction assay for screening
hazardous wasted for genotoxic activity is discussed along with some of
the problems associated with screening highly toxic wastes containing
toxic volatile compounds.
INTRODUCTION
According to estimates by the U.S. Environmental Protection Agency (US
EPA), more than 260 million metric tones of hazardous wastes are
generated annually, a quantity equal to more than 70 billion gallons
(Dietz et al, 1984). Some of the acute health effects linked to
hazardous waste exposures include lung and skin irritations (US EPA,
1980), eye irritations and menstrual problems (Grisham, 1986), and
transitory liver damage (Clark et al, 1982; Meyer, 1983). Included
among the chronic health effects associated with hazardous waste
exposure are chromosomal damage, cancer, and reproductive anomalies
(Maugh, 1979; Vianna and Polan, 1984; Warren, 1981).
2—139
-------
One of the ways by which wastes are classified as hazardous is on the
basis of specific chemical constituents that may be present in the
waste (Federal Register, 1980; Greer, 1984). However, analyzing a
waste for a limited set of compounds may not provide data that reflect
the overall chemical composition of the waste, nor does it take into
account the possibility of chemical interactions (antagonisms,
synergisms, etc.) or the production of metabolites resulting from
degradative pathways. In contrast to chemical analysis, the
identification of potential biological effects elicited by a waste may
provide more useful information for determining human health effects.
Because of the presence of known mutagens and carcinogens in many
wastes, genetic toxicity is an important endpoint that should be
determined for wastes.
The Salmonella assay developed by Dr. Bruce Ames and coworkers (Ames et
al, 1975) has been used more than any other short—term assay for
determining the genotoxicity of complex mixtures and hazardous wastes.
However, hazardous wastes can contain high concentrations of
carcinogenic metals, chlorinated compounds, and solvents (US EPA, 1984)
that are detected poorly by the Salmonella assay (Kier et al, 1986;
Zeiger and Tennant, 1986; Claxton et al, 1987). Thus, the genotoxic
potential of some wastes may escape detection when evaluated by the
Salmonella assay.
The Microscreen phage—induction assay developed by Rossman et al
(1984), has been shown to detect some carcinogenic metals (Rossman et
al, 1984), chlorinated pesticides (Houk and DeMarini, 1987), and
solvents (Houk and DeMarini, in prep.) that are not mutagenic in
Salmonella. It is a rapid and inexpensive assay that quantitatively
measures the induction of prophage A in Escherichia coli WP2(X).
The induction of prophage is one of several events that may occur upon
induction of the SOS response, which is a cellular reaction to DNA
damage. Other manifestations of the SOS response include DNA repair
and mutagenesis, and may agents that induce the SOS response in
bacteria are also genotoxic in a variety of other organisms (D’Ari,
1985).
MATERIALS AND METHODS
In the present study, we have used the Microscreen phage-induction
assay to evaluate the genotoxicity of 14 crude hazardous waste samples
obtained from Edward L. Katz, Hazardous Waste Engineering Laboratory,
US EPA, Cincinnati, OH (Table 1). Three of the waste samples were from
petrochemical, pharmaceutical, and plastics manufacturers (A, B, and C,
respectively). The remaining wastes CD through 0) were composites of
wastes from a variety of industrial sources that were collected by 4
commercial hazardous waste incineration facilities and classified
according to their aqueous or organic properties. Table I provides a
cursory physical description of the wastes, all of which were liquids
or semi—solids.
2-140
-------
Table II shows the results of a partial chemical analysis that was
performed on these samples (US EPA, 1984). These characterizations
should not be viewed as indicative of the overall chemical composition
of the wastes because the analysis was limited to specific organics
and/or metals identified in the EPA Appendix VIII list of priority
pollutants (US EPA, 1984). In addition, for purposes of a previous
study on the performance of hazardous waste incinerators (US EPA,
1984), 7 of the waste samples (B, F, G, J, K, L, and N) were spiked
with carbon tetrachioride and trichioroethylene. Thus, the
concentrations of these 2 chemicals in the 7 waste samples reflect this
addition.
The phage-induction assay was performed as described by Rossman e t al.
(1984) with modifications (Houk and DeMarini, 1987). For a more
detailed explanation of materials and methods, see Houk and DeMarini
(1987, and submitted). The bacterial strains are derivatives of E.
coil B/r. WP2Q ) is a lambda lysogen of WP2s ( trpE , uvrA), and SR714
(trpE, uvrD 3 ) is the indicator strain.
The crude (unfractionated) waste samples were serially diluted in
supplemented minimal medium in sterile 96-well microtiter dishes
(Corning). The lysogenic strain of E. coli was then exposed overnight
(20 h, 37°C) to the diluted wastes in the presence and absence of an
exogenous metabolic activation system (S9). Following exposure, wells
of the microtiter dishes were scored for turbidity (cell survival and
growth) or clarity (cytotoxicity and/or inhibition of cell growth).
Turbidity in the wells of microtiter dishes was often difficult to
discern due to the physical nature of these crude complex mixtures.
Many of the waste samples produced a precipitate (noted by a “P” in
Table III) or an oily film (noted by an “F”) that complicated the
determination of whether the wells were clear or turbid. The cytotoxic
responses of the waste samples are noted by a “T” for clear wells . and a
“t” for slightly turbid wells in Table III.
Wells of the microtiter dishes were then sampled for the presence of
prophage by exposing diluted contents of selected wells to the
indicator strain (SR714) and enumerating the resulting plaque-forming
units (PFU). The appearance of a dose—related increase of induced
(observed minus background) PFU that reached or exceeded the upper
limits of the 99% confidence interval for the control plates indicated
a positive——or genotoxic——response. The background PFU/plate were
calculated by averaging results from eight medium control wells (with
and without S9) for each experiment. The upper limits of the 99%
confidence intervals were 102 PFU/plate in the presence of S9 and 30
PFU/plate in the absence of S9.
2—141
-------
RESULTS
The response for each waste sample based on the criterion described
above is shown in Table III , which also presents raw data from a
representative experiment for each of the wastes. Doses were selected
to illustrate the response range for each waste tested. Eleven of the
14 crude hazardous waste samples induced prophage; only 2 of the wastes
(D and J) required 59 for induction. In addition to showing the
induced PFU/plate, Table III also indicates the fold increase over the
background. For most of the positive waste samples, the dose that
produced either 102 induced PFU/plate (+S9) or 30 induced PFU/plate
(—S9) generally resulted in a 3— or 4—fold increase over the
background, which is consistent with the fold increase observed in a
previous study of pure compounds (Houk and DeMarini, 1987). This is
also consistent with the 3—fold increase recommended by Rossnian et al
(1985) for a positive response.
Dose—response curves for the waste samples are illustrated in Figs. 1—
3. Except for waste samples A and C in the presence of S9 (Fig. 3),
each waste sample produced a dose—response curve exhibiting a linear
portion that spanned ‘ - 1 order of magnitude of dose (Figs. 1 and 2).
Based on the linear portions of these dose—response curves, we
calculated the concentration required by each waste to produce an
induced PFU/plate equal to the upper limits of the 99% confidence
intervals of the medium controls (Table IV). Each of these values
represent the minimum concentration of the waste required to produce a
positive response based on our criterion described previously. The
waste samples were then ranked from the most to the least potent (Table
IV).
Although waste samples A and C were positive in the presence of S9
(Table III), they did ont product simple linear dose responses.
Instead, they produced reproducible, multimodal dose—response curves
that ranged over 4 orders of magnitude (Fig. 3). In addition, waste
sample G produced a nonlinear dose response at the doses tested in the
presence of S9 (Table III; data not plotted).
DISCUSSION
The partial chemical characterization available for these wastes (Table
II) is more extensive than would ordinarily be available for most waste
samples. However, even this level of chemical analysis is inadequate
to distinguish a genotoxic form a nongenotoxic waste or to indicate
which waste might be more genotoxic than another. For example, the
chemical analysis of the petrochemical and pharmaceutical waste samples
(A and B, Table II) does not necessarily suggest the remarkable
cytotoxic and genotoxic potencies exhibited by these two samples
compared to the other waste samples (Tables III and IV). Likewise, the
finding that waste samples K, N, and 0 were not genotoxic is not
readily apparent from their chemical profiles, which are relatively
2-142
-------
indistinguishable form those of the genotoxic waste samples (Table II).
The limitations of the chemical analysis for predicting the biological
effects of these wastes argue in favor of bioassays to characterize the
possible hazardous nature of wastes.
The similar genotoxic potencies exhibited by most of the waste samples
(Fig. 1, Table IV) may have been due to the fact that most of these
wastes were composites of wastes from a variety of sources. Thus,
mixtures of potent and weak wastes may result in samples with similar
average potencies. Consistent with this interpretation is the fact
that the two samples that have vastly different potencies from the rest
(samples A and B) are not composite wastes but are from two distinct
industrial sources, the petrochemical and pharmaceutical industries.
Table V compares the mutagenicity of these waste samples in Salmonella
to the ability of these waste samples to induce prophage. All but one
of the wastes that were mutagenic in Salmonella also induced prophage.
This is consistent with studies of pure compounds which show that for
those compounds tested, most chemicals that are mutagenic in Salmonella
also induce prophage and/or the SOS response (Elespuru, 1984; Oda et
al, 1985; Ohta et al, 1984; Quillardet et al, 1985; Rossman et al,
1986). Thus, at lease some concordance between the assays is not
surprising. The inability to detect waste sample 0 in the phage—
induction assay may have been due to a toxic or inhibitory effect of
the sample on prophage production as evidenced by a reduction in
PFU/plate at concentrations above 100,0000 pg/mi (Table III).
The phage-induction assay detected 5 additional waste samples as
genotoxic that were not mutagenic in Salmonella (Table V). As
mentioned previously, the Microscreen phage—induction assay has been
shown to detect some carcinogenic metals (Rossman et al, 1984),
chlorinated organics (Houk and DeMarini, 1987), and solvents (Houk and
DeMarini, in prep.) that are not mutagenic in Salmonella (Keir et al,
1986; Zeiger and Tennant, 1986). Metals and compounds of these types
are present in most of the waste samples studied here (Table II), and
the ability of some of these compounds to induce prophage may account
for the detection by the phage-induction assay of the 5 additional
waste samples that were not detected by the Salmonella assay. In
addition, accumulating evidence indicates that phage-induction (and the
SOS response in general) is a broader genetic endpoint than reverse
mutation in bacteria, occurring by a variety of mechanisms (Elespuru,
1984; Hofnung and Quillardet, 1986) and involving a number of classes
of genetic damage (Rossman et al, 1985; Elespuru, 1984).
For most assays, complex mixtures must be extracted or fractionated
prior to bioassay to make them compatible with the test system.
However,with the Microscreen assay we were able to test the crude
wastes directly, saving considerable time, effort, and expense. The
unfractionated wastes posed no sterility problems, probably due to
their extreme toxicity. Other wastes, however, may contain microbial
contaminants that would necessitate the preparation of organic
extracts.
2—143
-------
Currently, the US EPA identifies wastes as hazardous based primarily on
physical characteristics and chemical composition of the wastes
(Federal Register, 1980; Friedman, 1985; Greer, 1984). However, the
deficiencies of this definition have led EPA to consider adding health
effects, including mutagenicity, to their definition of hazardous waste
(Federal Register, 1983; 1984a,b). In addition to government, industry
also has recognized the important role that short—term assays could
play in the toxicological assessment of hazardous wastes (Barfknecht
and Naismith, 1984; Guiney, 1985). In combination with chemical
analysis, assessments of the biological effects of wastes could
contribute valuable information regarding the hazardous nature and risk
associated with exposure to industrial wastes and effluents. Our data
suggest that the Microscreen phage-induction assay may provide a
simple, sensitive, and inexpensive way to screen hazardous wastes for
an important biological endpoint, genotoxicity.
2-144
-------
REFERENCES
Ames BN, McCann J, Yamasaki E (1975): Methods for detecting
carcinogens and mutagens with the Salmonella/manimalian—niicrosome
mutagenicity test. Mutation Res 31:347-364.
Barfknecht TR, Naismith RW (1984): Methodology for evaluating the
genotoxicity of hazardous environmental samples. Hazard Waste 1:93—
109.
Clark CS, Meyer CR, Gartside PS, Majeti VA, Specker B, Balistreri WF,
Elia VJ (1982): Environmental health survey of drinking water
contamination by leachate from a pesticide waste dump in Hardeman
County, Tennessee. Arch Environ Health 37:9—18.
Claxton LD, Stead AG, Walsh D (1987): An analysis by chemical class of
Salmonella mutagenicity tests as predictors of animal carciriogenicity.
Mutat Res (in press).
D’Ari R (1985): The SOS System. Biochimie 67:343—347.
Dietz S. Emmet M, DiGaetano R, Tuttle D, Vincent C (1984): National
survey of hazardous waste generators and treatment, storage, and
disposal facilities regulated under RCRA in 1981. EPA 530/SW—84—005.
Prepared by Westat, Inc. Washington, DC: U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response. 3lSp.
Available from: U.S. Government Printing Office, Washington, DC.
Elespuru RK (1984): Induction of bacteriophage lambda by DNA-
interacting chemicals. In de Serres FJ (ed): “Chemical Mutagens:
Principles and Methods for their Detection, Vol 9.” New York: Planum,
pp 213—231.
Federal Register (1980): Hazardous waste and consolidated permit
regulations. Vol 45, No 98, May 19, pp 33066—33133.
Federal Register (1983): Notification requirements; reportable
quantity adjustments. Vol 48, No 102, May 25, pp 23552—23602.
Federal Register (1984a): Hazardous waste management systems. Vol 49,
No 32, Feb 15, pp 5854—5859.
Federal Register (1984b): Proposed deadlines for exposure assessment.
Vol 49, No 227, Nov 23, pp 46304—46312.
Friedman D (1985): An overview of selected EPA RCRA test method
development and evaluation activities. In Petros 3K, Lacy WJ, Conway
RA (eds): “Hazardous and Industrial Solid Waste Testing: Fourth
Symposium.” Philadelphia: American Society for Testing and Materials,
pp 77—84.
2-145
-------
Greer LE (1984): Definition of hazardous waste. Hazard Waste 1:309-
322
Grisham 3W (1986): “Health Aspects of the Disposal of Waste
Chemicals.” New York: Pergamon Press, ch 8, pp 241—318.
Guiney PD (1985): Use of predictive toxicology methods to estimate
relative risk of complex waste mixtures. Hazard Waste Hazard Mater
2:177—189.
Rofnung N. Quillardet P (1986): Recent developments in bacterial
short—term tests for the detection of genotoxic agents. Mutagenesis
1:319—330.
Houk VS, DeMaririi DM (1987): Induction of prophage lambda by
chlorinated pesticides. Mutat Res (in press).
Rouk VS 1 DeMarini DM (in prep) Compatibility and gerzotoxicity of
organic solvents in the Microscreen phage—induction assay.
Xier LD, Brusick DJ, Auletta AE, Von Ralle ES, Brown MM, Simmon VF,
Dunkel V, McCann 3, Mortelmans K, Prival M, Rao TK, Ray V (1986): The
Salmonella typhimurium/manmtalian microsomal assay. A report of the US
Environmental Protection Agency Gene—Tox Program. Mutat Res 168:69 —
240.
Maugh TN (1979): Toxic waste disposal a growing problem. Science 204:
819—823.
Meyer CR (1983): Liver dysfunction in residents exposed to leachate
from a toxic waste dump. Environ Health Perspect 48:9-13.
Oda Y, Nakamura S , Oki I, Kato T, Shinagawa H (1985): Evaluation of
the new system (wnu—test) for the detection of environmental inutagens
and carcinogens. Nutat Res 147:79—95.
Ohta T, Nakamura N, Moriya M, Shirasu Y, Kada T (1984): The SOS-
function—inducing activity of chemical mutagens in Escherichia coli .
Mutat Res 131:101—109.
Quillardet P, de Bellecoinbe C, Hofnung M (1985): The SOS Chromotest, a
colorimetric bacterial assay for genotoxins: validation study with 83
compounds. Mutat Res 147:79-95.
Rossman TG, Molina M, Meyer LW (1984): The genetic toxicology of metal
compounds: I. Induction of prophage in E. coli WP2 5 ( ). Environ
Mutagen 6:59—69.
2-146
-------
Rossman TG, Meyer LW, Butler JP, Daisey JM (1985): Use of the
Microscreen assay for airborne particulate organic matter. In Waters
MD, Sandhu SS, Lewtas J, Claxton L, Strauss G, Nesnow S (eds): “Short-
Term Bioassays in the Analysis of Complex Environmental Mixtures, IV.”
New York: Plenum, pp 9-23.
Rossman TG, Meyer LW, Molina M (1986): Induction of prophage as a
screen for genotoxic agents. Ann NY Acad Sci 463:347—348.
US Environmental Protection Agency (1980): “Everybody’s Problem:
Hazardous Waste.” Washington DC: Office of Water and Waste Management
US Environmental Protection Agency (1984): Performance Evaluation of
Full Scale Incinerators, National Technical Information Center
Publication No. PB85—129500.
Vianna NJ, Polan AK (1984): Incidence of low birth weight among Love
Canal residents. Science 226:1217—1219.
An overview. In Collins JP,
Dilemma: Issues and Solutions.”
Engineers, pp 5—15.
Zeiger E, Tennant RW (1986): Mutagenesis, clastogenesis,
carcinogenesis: expectations, correlations and relations. In Ramel C,
Lambert B, Nagnusson J (eds): “Genetic Toxicology of Environmental
Chemicals, Part B.” New York: Alan R. Liss, pp 75—84.
Warren CS (1981): Hazardous waste:
Saukin WP (eds) “The Hazardous Waste
New York: American Society of Civil
2—147
-------
TABLE I PhYSICAL DESCRIPTIONS OF HAZARDOUS WASTES
Waste Physical Description
State
A Liquid Black, very thin oil from a petrochemical manufacturing
plant
B Liquid Black, oily liquid from a pharmaceutical manufacturer
C Tar Black, pourable tar produced during the production of plastics
D Liquid Composite of aqueous wastes; watery liquid with red oil droplets
F and G Semi-solid Composite of organic wastes; biphasic gray sludge with reddish—
brown liquid
H and I Suspension Composite of aqueous wastes; thin, gray slurry
3 Liquid Composite of organic wastes; gray, thick liquid with suspended
solids
X Liquid Similar to X, but lighter in color and thinner
L and M Tar Composite of organic wastes; black, thin, pourable tar
N and 0 Liquid Composite of aqueous wastes; clear, watery liquid
2-148
-------
. tJLZIL
C,IcZN tI0WS or cHvtrJ.Ls 0 PUD J.S ZDEIrVI?120 ON HAZARDOUS WAS? 8 (uq/g)
p.tro—
ch.aicai
Pt ar *
e.utica
7tiu—
ticu
azard s- a.t.
1 2
Aqu.oul Organic Organic q asouu
tncinsratioll
fact
itty
3
Organic
4
Organic
Aqu.ouu
C 5.,4cai/M.t ui’
u
C
0
2 F
C
I
S
I . N
0
ljd ltnI 14000 550000
IIMY thlorid. 3000
Su—(2—.thy i— 500
-------
TABLE III INDUCTION
OF
PROPHAGE LA) BD
BY HAZARDOUS WASTES
+S9
-S9
Induced
Fold
Induced
Fold
Waste
Dose
(pg/mi)
PFU per
Platea
PFU per
Plateb
in—
crease
Dose
(pg/mi)
PFTJ per
Plate
PFU per
Plate
in-
crease
A 0 21 0 — 0 12 0 ——
5.0 X i0 82 61 3.9 2.0 X i0 30 18 2.5
2.0 X i0 160 139 7.6 4.0 X i0 49 37 4.1
1.2 X i02 320 299 15.2 8.0 X i0 60 48 5.0
2.5 X 10 2 388 367 18.5 1.5 X 106 2(T) 0 ——
1.5 552 531 26.3
50 241 220 11.5
200 182 161 8.7
400 0(T) 0 0
Response + +
B 0 21 0 — 0 12 0 ——
0.05 249 228 11.9 2.0 X iO’ 7 24(P) 12 2.0
0.1 421 400 20.0 4.0 X 49(P) 37 4.1
0.2 582 561 27.2 8.0 X i0 106(P) 94 8.8
0.4 957 936 45.6 1.5 X 106 136(P) 124 11.8
0.8 561(T,P) 540 26.7 3.0 X 10—6 572(t,P) 560 47.8
1.6 2(T,P) 0 — 6.0 x 10—6 24(P) 12 2.0
Response + +
C 0 21 0 —— 0 26 0 ——
1.0 X i02 29 8 1.4 200 24 0 ——
5.0 X i c r 2 56 35 2.7 400 30 4 1.2
0.2 125 104 6.0 800 57 31 2.2
0.8 81 60 3.9 1,600 68 42 2.6
3 53 32 2.5 3,tOO 0(T) 0 ——
12 48 27 2.3 6,300 0(T) 0
50 136 115 6.5
200 9(t) 0 ——
Response + +
D 0 52 0 — 0 12 0 ——
800 76 24 1.5 800 19 7 1.6
1,600 138 86 2.7 1,600 12 0 1.0
3,100 131 79 2.5 3,100 12 0 1.0
6,300 148 96 2.8 6,300 11 0 ——
12,500 218 166 4.2 12,500 12 0 1.0
25,000 23(t) 0 — 25,000 0(T) 0 ——
Response +
F 0 53 0 — 0 9 0 ——
100 63 10 1.2 200 15 4 1.7
200 78 25 1.5 800 33 24 3.7
400 103 50 1.9 1,600 75 66 8.3
3,100 115 62 2.2 3,100 235 226 26.1
12,500 259 206 4.9 6,300 721 712 80.1
25,000 3,755(P) 3,702 70.8 12,500 1,960 1,951 217.8
50,000 416(P) 363 7.8 25,000 962 953 106.9
100,000 0(P) 0 —— 50,000 1(P) 0 ——
Response + +
2-150
-------
+S9
—S9
Induced
Fold
Induced
Fold
Dose
PFU
per
PFU per
in—
Dose
PFU
per
PFU per
in—
Waste (iig/ml)
Plate
Plate
crease
(jig/mi)
Plate
Plate
crease
G 0 89 0 —— 0 19 0 ——
1,250 353 264 4.0 780 20 1 1.0
2,500 346 257 3.9 3,100 100(t) 81 5.3
5,000 402 313 4.5 5,000 950(t) 931 50.0
6,300 400 311 4.5 6,250 5,430(t) 5,411 285.8
12,500 466 377 5.2 10,000 3,450(t) 3,431 181.6
25,000 4,200(P) 4,111 47.2 20,000 2,300(T) 2,281 121.1
40,000 1,800(P) 1,711 20.2 25,000 71(T,P) 52 3.7
50,000 17(P) 0 —— 50,000 0(T,P) 0 — —
Response ÷ +
if 0 53 0 0 9 0 ——
400 33 0 —— 400 10 1 1.1
800 76 23 1.4 800 14 5 1.6
1,600 202 149 3.8 1,600 19 10 2.1
3,100 373 320 7.0 3,100 42 33 4.7
6,300 591 538 11.2 6,300 64 55 7.1
12,500 792 739 14.9 12,500 138 129 15.3
25,000 0(P) 0 —— 25,000 0(P) 0 ——
Response + +
0 53 0 0 9 0
800 88 35 1.7 800 11 2 1.2
1,600 361 308 6.8 1,600 34 25 3.8
3,100 768 715 14.5 3,100 68 59 7.6
6,300 740 687 14.0 6,300 44 35 4.9
12,500 936 883 17.7 12,500 27 18 3.0
25,000 0(P) 0 — — 25,000 0(P) 0 — —
Response + +
J 0 62 0 0 12 0
12 56 0 —— 6 10 0 ——
50 76 14 1.2 12 13 1 1.1
100 111 49 1.8 25 15 3 1.3
200 173 111 2.8 50 26 14 2.2
400 133 71 2.1 100 28(t) 16 2.3
800 12(t) 0 —— 200 10(t) 0 ——
Response +
K 0 62 0 —— 0 26 0 ——
50 69 7 1.1 200 27 1 1.0
100 60 0 400 29 3 1.1
200 66 4 1.1 800 3(t) 0 ——
400 87 25 1.4 1,600 3(t) 0 ——
800 16(t) 0 —— 3,100 1(t) 0 ——
Response
2—151
-------
+89
-S9
.
.
Induced
Fold
Induced
Fold
Dose
PFU
per
PFTJ per
in—
Dose
PFtJ
per
PFU per
in—
Waste
(i g/ml)
Plate
Plate
crease
( ig/ml)
Plate
Plate
crease
L 0 62 0 —— 0 12 0 ——
12 69 7 1.1 800 29 17 2.4
400 116 54 1.9 1,600 69 57 5.8
6,300 204 142 3.3 3,100 101 89 8.4
12,500 464(F) 402 7.5 6,300 266 254 22.2
25,000 2,350(F) 2,288 37.9 12,500 1,692 1,680 141.0
50,000 0(F) 0 —— 25,000 5(F) 0 ——
Response + +
M 0 62 0 —— 0 12 0 ——
3,100 132 70 2.1 1,600 53 41 4.4
6,300 131 69 2.1 3,100 79 67 6.6
12,500 184(F) 122 3.0 6,300 224 212 18.7
25,000 406(F) 344 6.5 12,500 2,085 2,073 173.8
50,000 287(F) 225 4.6 25,000 7,590(F) 7,578 632.6
100,000 0(F) 0 —— 50,000 0(F) 0 ——
Response + +
N 0 87 0 — 0 26 0 ——
37,500 107 20 1.2 37,500 28 2 1.1
50,000 139 52 1.6 50,000 23 0 ——
75,000 135 48 1.6 75,000 23 0
100,000 142 55 1.6 100,000 24 0
200,000 23 0 —— 200,000 16 0
300,000 18 0 —— 300,000 10 0
Response
o 0 87 0 —— 0 26 0
37,500 77 0 — 37,500 12 0
50,000 84 0 —— 50,000 11 0
75,000 94 7 1,1 75,000 12 0
100,000 100 13 1.1 100,000 16 0
200,000 31 0 — 200,000 10 0
300,000 18 0 —— 300,000 6 0
Response
aEach value is the average of two plates. T, toxic (clear well); t, slightly toxic
(slightly turbid well); P, precipitate; F, oily film on surface of medium. The
positive control results were: 2—aminoanthracene (0.3 ig/ml) 716 ± 288 PFU/plate;
2—nitrofluorene (150 ug/ml) 400 ± 264 PFU/plate.
2-152
-------
TABLE IV RELATIVE POTENCIES OF WASTES
Concentration (iig/ml)
required
to pro
duce an
induced
PFU/plate equal to the
upper
limit
of the
99% confidence
interval
based
on the
medium
Waste +S9
controlsa
(rank)
-S9
(rank)
B 0.03 (1) 3 X 10 (1)
A ÷b 4 X i0 (2)
C + 1082 (3)
J 185 (2) —
I 976 (3) 1888 (7)
H 1346 (4) 3355 (8)
L 3Q29 (5) 1200 (4)
D 6076 (6) —
F 6444 (7) 1490 (5)
M 9561 (8) 1638 (6)
G + 3456 (9)
aThis is the concentration (i.tg/ml) required to produce 102
induced PFU/plate (+S9) or 30 induced PFU/plate ( —S9).
bpotency value incalculable due to nonlinear dose response
(see text).
2—153
-------
TABLE V COMPARISON OF PROPHAGE INDUCTION IN Escherichia coli AND
MUTAGENESIS IN Salmonella typhimurium
Prophage
Induction
Mutagenicity in
Waste in E. coli
S. typhimuriulna
A + +
B +
C + +
D 4. —
F + +
+ +
N 4.
1 4
+
&
L + +
A 4 +
N
0 +
a From DeMarini et al (198Th). Wastes were tested in strains TA98 and
TA100. A positive response in either strain, with or without S9, was
sufficient for a positive summary response reported here.
2-154
-------
Fig. 1. Dose—response curves of prophage induction by hazardous waste samples
in the presence and absence of S9. Only the linear portions of the
dose—response curves are shown, and the correlation coefficients
(“r 2 ” values) are 0.99 for all of the curves except for L (0.97)
and F (0.92) in the presence of S9; and for C (0.91), M (0.98), and
G (0.88) in the absence of S9.
Fig. 2. Linear portions of the dose—response curves of prophage induction by
hazardous waste samples A and B in the absence of S9. Correlation
coefficients are 0.94 and 0.99 for A and B, respectively.
Fig. 3. Dose—response curves for waste samples A and C in the presence of S9.
2—155
-------
6000
0.01 0.1
100
100
10
1 10.
1000
+s9
/
0
M
Dose
(pg/mi)
FIGURE 1
F
1000
100
10
6000
1000
I
a)
- )
(0
ci
Lt
0
-D
a)
U
-D
‘ I
100 1000 10000
—s9
G
C
H
I
- —..—.-. t
10000 30000
2—156
-------
100
80
60
40
20
0
0
600
100
10
6
jig/mi)
0.00001
.1 •
0.001
Dose
I . i I I .i. i1
0.01
(jjg/m l)
I
10 100
FIGURE 3
2—157
B
A
2
4
Dose CX
4 J
(0
LL
a
a)
C-)
C
1 -4
a)
(0
F-,
LL
a
a)
C-)
C
‘—4
8
-J
10
FIGURE 2
+s9
A
C
-------
ASSESSMENT OF THE TLC/SALMONELLA ASSAY FOR SCREENING HAZARDOUS WASTES
Virginia Stewart Houk and Larry D. Claxton, Genetic Bioassay Branch,
GTD/HERL, U.S. Environmental Protection Agency, Research Triangle Park,
NC
ABSTRACT
Using a modified version of the TLC/Salmonella assay developed by
Bjorseth et al. (1982), ten complex hazardous wastes were tested for
mutagenic activity. This method couples thin layer chromatography
(TLC) with the Salmonella/mairimalian-injcrosome (Ames) assay for the
detection of mutagenic constituents in complex mixtures. Crude
hazardous wastes and selected hazardous waste extracts were
fractionated on commercially available cellulose TLC plates.
Mutagenicity testing was performed by applying a single overlay of
minimal growth agar containing a tester strain of Salmonella and the
optional metabolic activation system directly onto the developed
chromatogram. The appearance of localized clusters of revertant
colonies or an increase in total revertant growth vis-a—vis control
plates indicated a mutagenic response. Seven of ten hazardous wastes
demonstrated mutagenic activity when tested by this method.
To assess the sensitivity of the modified TLC/Salmonella assay,
fourteen Salmonella mutagens from a wide range of chemical classes and
polarities were tested. The selected compounds included heterocyclics,
aromatic ainines, alkylating agents, antitumor agents, a nitrosamine,
and a nitroaromatic. Eleven of the fourteen mutagens were positive in
this test system. The three compounds refractory to analysis included
a polycyclic aromatic hydrocarbon and two volatile compounds.
INTRODUCTION
Complex hazardous wastes are composed of compounds that manifest
diverse chemical, physical, and toxicological properties. Human
exposures to hazardous wastes have been associated with a number of
adverse health effects, including chromosomal damage, cancer, and
reproductive anomalies (Maugh, 1979; Vianna and Polan, 1984; Warren,
1981). Adequate characterization of the genotoxic potential of
hazardous wastes is therefore crucial to the assessment of risk to
human health.
Assessments of genotoxic risk are based in part upon results from
short—term in vitro assays. One such test, the Salmonella/mammalian-
microsome assay (Ames et al., 1975; Maron and Ames, 1983) has been used
for over a decade to predict the inutagenic potential of pure compounds
and complex mixtures. To enhance detection capabilities, in vitro
assays are often coupled with chemical or physical fractionation
techniques to separate or remove components from the complex
-------
sample prior to testing and to reduce the effects of interactive
mechanisms. For example, in a study performed by Bjorseth et al.
(1982), thin layer chromatography was used to segregate potential
mutagens in a complex environmental mixture, and the Salmonella assay
was then applied directly onto the developed chromatogram. The
subsequent appearance of localized clusters of revertant colonies
suggested the presence of a mutagenic constituent. This assay has been
successfully applied to the analysis of air samples (Bjorseth et al. ,
1982), extracts of typewriter ribbons and carbon paper (Bjorseth et
al., 1982; Moller et al., 1983), and emission samples from coal- and
oil—fired boilers (Alfheim et al., 1983).
We modified the TLC/Salmonella assay described by Bjorseth and his
colleagues by replacing the double-agar system with a single overlay
composed of agar, tester strain, and the metabolic activation system.
This modification simplified the technique, reduced preparation time,
and, additionally, provided direct contact between the bacteria and
partitioned components on the chromatogram, ostensibly enhancing
detection of nonpolar and intermediately polar compounds. A complete
description of this work has been published elsewhere (Houk and
Claxton, 1986).
Ten hazardous wastes were screened for mutagenic activity using this
modified assay. Samples were selected from a variety of dissimilar
waste types, including acids, caustics, tars, emulsions, and sludges.
Solids, semi—solids, and liquids were represented, as were inorganic
and organic wastes.
The detection capabilities of this modified TLC/Salmonella assay were
challenged with fourteen compounds known to be mutagenic in the
Salmonella plate incorporation assay. Mutagens were selected from a
wide range of chemical classes and polarities, and included alkylating
agents, aromatic amines, antitumor agents, a nitrosamine, and a
nitroaromatic.
MATERIALS AND METHODS
The hazardous waste samples evaluated in this study were obtained from
Batelle Columbus Laboratories, Columbus, Ohio, courtesy of Dr. 14.
McKown. Table 1 provides gross chemical and physical characterizations
and a cursory description of each waste tested. Results from gas
chromatography/mass spectrometry ( CC/MS) 1 emission spectroscopy, and
analysis for purgeable and semi—volatile organic compounds have been
published elsewhere (Battelle, 1981; Houk, 1984; Miller et al., 1981;
Warner et al. 1981).
Most samples arrived in physical states that favored direct application
to TLC plates (liquids, emulsions, sludges, and a tar). The organic
still bottoms and the dewatered municipal sludge, however, were solids
and could not be applied to plates in their unadulterated states.
2-160
-------
Ethanol (ETOH) extracts of these wastes were tested. Ethanol and
dimethyl sulfoxide (DMSO) extracts of the coke plant waste were also
obtained and were tested in addition to the crude sample. Extractions
were performed as previously described (Andon et al., 1985).
The TLC/Salmonella technique was a modification of the procedure
developed by Bjorseth et al. (1982) and is described in detail in Flouk
and Claxton (1986). The TLC separations were performed on commercially
available glass—backed cellulose plates (100 mm x 100 mm). Neat
samples or extracts were spotted on the plates at an origin 15 mm from
the bottom edge and 20 mm apart. Doses ranged from approximately 0.1
percent to 16 ul per application point. Plates were developed at room
temperature in chloroform for 15 to 20 mm, at which time the solvent
front had ascended to approximately 10 mm from the top edge of the
plate. When chloroform proved ineffective in fractionating the sample,
other carrier systems wee utilized (see Results and Discussion) -
After the plates were removed from the developing chamber and the
mobile phase had evaporated from the plate, developed plates were
inspected for chromatographic patterns using ambient and ultraviolet
light sources. Chromatograms were placed sorbent side up into over-
sized Petri dishes (150 nun by 15 mm) for mutagenicity testing.
Genotoxic analyses were performed in situ . A mixture containing agar,
minimal growth medium, the bacterial strain, and the optional metabolic
activation system was poured over the TLC plates. The Petri dishes
were incubated for 72 h at 37°C.
Salmonella typhimurium strains TA98 and TA100 were obtained from Dr.
Bruce Ames. Rat liver homogenate (S9) was prepared from rats injected
with Aroclor 1254 as previously described (Ames et al., 1975).
Developed blank TLC plates were used as negative controls. Undeveloped
plates spotted with 2—anthramine, 2—nitrofluorene, or sodium azide were
used as positive controls.
After incubation, the plates were examined for toxicity, as evidenced
by (1) a reduction in the overall number of revertants when compared to
spontaneous values or (2) a discrete toxic zone devoid of bacterial
growth. A mutagenic event was indicated by the appearance of one or
more localized clusters of histidine revertants or as a reproducible
twofold or greater increase in revertant colonies (for the total plate)
over spontaneous background values. Dose—response relationships were
examined by applying four different volumes per plate. Independent
experiments were performed to guarantee reproducibility and to confirm
positive or negative findings. Results were documented by tracing
observed chromatographic patterns and revertant colony locations onto
transparent sheets of cellulose acetate.
2-161
-------
To examine the predictive capabilities of the assay, fourteen
Salmonella mutagens of various classes and polarities were selected for
testing. Increasing volumes of the mutagen solution were spotted in
clockwise fashion at the four corners of a cellulose plate. Plates
were not chromatographed, but were placed undeveloped into Petri dishes
and tested for mutagenic response using the single overlay previously
described.
RESULTS ?.ND DISCUSSION
Analysis of Hazardous Wastes
Results from the mutagenicity testing of the hazardous waste samples
are given in Table 2.
ORGANIC SLUDGE
A relatively nonpolar mutagenic component migrated to a high Rf value
following chromatography of the organic sludge sample. Figure la
illustrates the observed mutagenic response when the chromatogram was
tested with TA98 in the absence of metabolic activation.
The diluted counterpart of this sludge, spiked with eleven organic
compounds, failed to demonstrate similar activity (Table 2). Two
explanations are plausible. The organic compounds which were added to
the original sludge could have interfered with the mutagenic expression
of the resulting sample. On the other hand, diluting the parent sludge
may have masked the original activity, permitting inutagenic expression
of the added compounds,
EDC SPENT CAUSTIC
The EDC spent caustic contained a mutagen(s) that induced a dose-
dependent cluster of revertant colonies in both strains without
metabolic activation. Activity was localized around fluorescent orange
bands that migrated with the solvent front (Figure lb). The migration
of this direct-acting mutagen suggests that the compound is relatively
nonpolar,
HERBICIDE MANUFACTURING ACID
With activation, the herbicide manufacturing acid produced intense
clustering of revertant colonies around the points of sample
application (both strains). Clustering was localized within circular
fluorescent zones (Figure ic). At doses greater than 1 ul, a zone of
toxicity surrounded by a ring of revertant colonies was evident. These
results indicate that indirect—acting mutagen(s) are present that are
toxic to Salmonella when administered at sample doses greater than 1
ul.
2—162
-------
The mutagenic constituent present in this herbicide waste was not
desorbed by chloroform and consequently did not ascend the TLC plate.
Thus, two nonpolar mobile phases described for pesticide separations
were selected to try to enhance migration. The first was a solvent
system for separation of chlorinated pesticides (heptane—acetone, 98:2)
described by Stephens and Chan (1980). The second was a system that
has proven useful in the separation of organophosphorus pesticides
(hexane-acetone, 5:1) (Stephens and deVera, 1979). Both solvent
systems failed to partition the herbicide waste. A polar mobile phase
(dichloromethane/ethanol/water, 10:20:1) described by Tomingas et al.
(1977) subsequently proved successful in desorbing the mutagenic
compound(s). Most revertant activity remained localized midway up the
plate within the fluorescent zone (Figure 2).
The initial localization of revertant activity around the points of
sample application suggests that the mutagens) is very polar. Changes
in the mobile phase system from nonpolar to polar have supported this
assumption. These findings demonstrate the usefulness of testing
different solvent systems to more clearly define chemical properties.
COKE PLANT WASTE
The neat coke plant waste pro uced approximately a two—fold increase in
the number of spontaneous His revertants when tested with TA98 in the
absence of exogenous activation. Ethanol and DMSO extracts of this
sample produced clusters of colonies at points of sample application
(figures not included). Independent analysis using the standard plate
incorporation assay supports our findings of mutagenic activity in the
extracts (Andon et al., 1985). Additional verification of the
mutagenic potential of this coke plant waste comes from GC/MS analysis,
which revealed the presence of the mutagens/carcinogens benzo(a) pyrene
(l.l2ug/g), benzo(b) and benzo(k)fluoranthene (0.69 ug/g), and chrysene
and benzo(a)anthracene (0.5 ug/g) (Miller et al., 1981).
INK PIGMENT SLUDGE
No mutagenic activity was evident when ink pigment sludge was tested
with TA98. With TA100, results proved inconclusive. Three independent
experiments were performed with increasing dosages, but convincing
results could not be obtained.
LATEX PAINT WASTE
Results from the TLC/Salmonella assay of the latex paint waste were
negative. However, elemental analysis of this waste indicated the
presence of several carcinogenic and mutagenic metals, including
cadmium, arsenic, and selenium (Miller et al., 1981). In general, the
Salmonella reversion assay does not detect mutagenic metals w; hout
bioassay modifications, so it follows t iat the TLC/Salmonella assay
would also prove unsuitable for the routine s ’re’ ing of metal—
containing wastes. 2-163
-------
COAL GASIFICATION TAR
The coal gasification tar was too viscous to be drawn into a pipette
tip. Instead, the tip was touched onto the surface of the sticky
substance and a small amount of tar was transferred to the TLC plate.
Consequently, applied sample volumes are unknown.
When chromatographed, this sample streaked the length of the plate with
no definition of component parts. Mutagenicity testing suggests that
this complex sample contains indirect—acting frameshift mutagens that
are not particularly water—soluble. This conclusion is based upon the
fact the TA98 (-FS9) revertants are localized along the length of the
visualized streak and are not profuse across the plate (Figure id).
There also appear to be direct—acting frameshift mutagens that are more
water—soluble, as evidenced by the diffuse response observed on plates
without metabolic activation (figure not shown).
ORGANIC STILL BOTrOMS, ETCH EXTRACT
There was no evidence of mutagenic or toxic activity in either strain.
DEWATERED MUNICIPAL SLUDGE, ETCH EXTRACT
Revertant clustering was observed in the area of the origin when the
chromatograni of the municipal sludge was tested with metabolic
activation (both strains). These results suggest the presence of
indirect—acting mutagens that are polar and probably water—soluble.
OIL REFINING WASTE
In independent duplicate experiments there appeared a slight dose-
response clustering around the origin in TA98 with activation. This
effect was evident only at low doses (from 0.5 to 2 ui) and disappeared
at volumes of 4 ul and greater. The explanation for this response may
be linked to the influence of a fluorescent spot that became larger as
sample volumes increased. This separated component appeared to be
toxic to th bacteria. An overall reduction of colony counts vis—a—vis
contxol plates was also noted.
Supportive evidence of this sample’s toxic potential comes from a study
using the CHO cytotoxicity test (Andon et al., 1985). This sample
produced a greater than 50% reduction both in the cellular AT?
concentration and in the viability index in Chinese hamster ovary
cells. Because oil refining wastes have been shown to contain
significant levels of lead and chromium (U.S. Environmental Protection
Agency, 1980), the observed cytotoxic response may ostensibly be linked
to these toxic metals.
2-164
-------
Analysis of Model Compounds
Results from the testing of the Salmonella mutagens are given in Table
3. Eleven of the fourteen test compounds were positive. Three
compounds --benzo(a)pyrene (BaP), epichlorohydrin, and dimethylcarbamyl
chloride--registered negative responses, even at very high dose levels
(Table 3). Previous TLC/Salmonella research on BaP also resulted in
negative findings (Ejorseth et al., 1982). As reported by Alfheim et
al., (1983) and Moller et al. (1983), this assay system appears
unsuitable for the detection of non—polar organics such as
unsubstituted PAils.
The lack of demonstrable mutagenic activity shown by epichiorohydrin
and dimethylcarbamyl chloride is presumably linked to their high
volatility. These unstable compounds probably volatilized from the
chromatography plate prior to testing, thereby precluding detection by
the Inutagenesis assay.
A nonmutagen (1-naphthylamine, 2—80 ug) was also tested and did not
register a mutagenic response.
CONCLUSION
Our findings demonstrate that the TLC/Salmonella assay described in
this paper can detect mutagens in complex hazardous wastes. Of ten
wastes tested, seven were found to contain mutagenically active
compounds. The assay proved to be applicable to a broad range of
chemically and physically dissimilar waste types. Liquid or near—
liquid samples could be applied to the TLC plates without preparatory
workup. Solid wastes, however, had to be converted to a form suitable
for TLC application and development. The system proved sensitive to
microliter volumes of crude test sample.
Additionally, our investigation has shown that detection capabilities
cover a wide range of mutagen classes and polarities. Mutagenic
activity was detectable in the microgram range. Compounds refractory
to analysis by the TLC/Salmonella assay included two mutagens that were
volatile and an unsubstituted PAH, a class that does not ordinarily
evoke a mutagenic response in a spot test.
Chemical and physical properties of detected mutagens may be inferred
from parameters such as localization on the TLC plate (expressed as an
Rf value), polarity of the mobile phase, and response to visualization
techniques (fluorescence, color, etc.). Fluorescent mutagens were
present in the organic sludge, the EDC spent caustic, the herbicide
manufacturing acid, and the coal gasification tar. Polar mutagens
(those with low Rf values) were detected in the herbicide manufacturing
acid, the municipal sludge, and the oil refining waste, while nonpolar
mutagens were detected in the organic sludge and the EDC spent caustic.
2—165
-------
The TLC/Salmonella assay offers several important advantages over other
coupled fractionation/bioassay schemes. The procedure is very rapid.
Fractionation of the samples by chromatography can be completed within
10 to 20 minutes. Mutagenicity results are documented after 72 hours.
Samples can be applied to the system in their neat state; thus,
extraction procedures are obviated except for samples that cannot be
chromatographed in their crude form (e.g., solids). Compounds that may
exert synergistic or antagonistic effects are separated from the
complex sample during chromatography and are detected independently.
Problems with sample toxicity are minimized.
In summary, the TLC/Salmonella assay represents a simple screening
tool for the qualitative detection of mutagens in complex environmental
sax les. The speed and economy with which results can be obtained make
it useful for routine testing and for processing large numbers of
samples. The presumptive identification of mutagenic constituents in a
sample can lead to a systematic selection of that sample for more
conclusive chemical or biological evaluation.
2-166
-------
References
Afheim, I., J.G.T. Bergstrom, D. Janssen, and M. Moller (1983)
Mutagenicity in emissions from coal- and oil—fired boilers, Environ.
Health Perspect., 47, 177—187.
Ames, B.N., J. McCann, and E. Yainasaki (1975) Methods for detecting
carcinogens and Inutagens with the Salmonella/mainxnalian—microsome
mutagenicity test, Mutation Res., 31, 347—364.
Andon, B.M., N. Jackson, V. Houk, and L. Claxton (1985) The evaluation
of chemical and biological methods for the identification of mutagenic
and cytotoxic hazardous waste samples, in: J.K. Petros, Jr., W.J.
Lacy, and R.A. Conway (Eds.), Hazardous and Industrial Solid Waste
Testing: Fourth Symposium, American Society for Testing and Materials
STP 886, Philadelphia, pp. 204—215.
Battelle Columbus Laboratories, Final Report on the Collaborative Study
for the Evaluation of a Selected Method for Hazardous Waste Analysis,
Dec. 8, 1981.
Bjorseth, A., G. Eidsa, J. Gether, L. Landmark, and N. Moller (1982)
Detection of mutagens in complex samples by the Salmonella assay
applied directly on thin layer chromatography plates, Science, 215, 87—
89.
Claxton, L.D., M. Kohan, A. Austin, C. Evans (1981) The Genetic
Bioassay Branch protocol for bacterial mutagenesis including safety and
quality assurance procedures, Health Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
Houk, V_S. (1984) Screening complex hazardous wastes for mutagenic
activity using the TLC/Axnes assay, Master’s Thesis, Department of
Environmental Sciences and Engineering, University of North Carolina,
Chapel Hill, NC.
Houk, V.S., and L.D. Claxton (1986) Screening complex hazardous wastes
for Inutagenic activity using a modified version of the TLC/Salmonella
assay, Mutat. Res. , 169, 81—92.
Maron, D. and B.N. Ames (1983) Revised methods for the Salmonella
mutagenicity test, Mutation Res., 113, 173-212.
Maugh, T.H. (1979) Toxic waste disposal a growing problem, Science,
204, 819—823.
2—167
-------
Miller, H.C., K.H. James, W.K. Dickson, M.D. Neptune, and M.H. Carter
(1981) On evaluation of methodology for the survey analysis of solid
wastes, in: L.A. Conway and B.C. Malloy (Eds.), Hazardous Solid Waste
Testing: First Conference, merican Society for Testing and Materials,
Philadelphia, pp. 240-266.
Moller, M. I. Alfheim, G. Lofroth, and E. Agurell (1983) Mutagenicity
of extracts from typewriter ribbons and related items, Mutation Res.,
119, 239—249.
Sparacino, C.M. (1983) Preparation and fractionation/characterization
of hazardous waste samples for toxicological studies, Draft Final
Report, Analytical and Chemical Sciences Division, Research Triangle
Institute, Research Triangle Park, N.C.
Stephens, R.D. and J.J. Chan (1980) The characterization of hazardous
wastes by thin layer chromatography, in: J.C. Touchstone and D. Rogers
(Eds.) Thin Layer Chromatography: Quantitative Environmental and
Clinical Applications, John Wiley and Sons, New York, pp. 363-369.
Stephens, R.D. and E.R. deVera (1979) Analysis of hazardous wastes,
U.S. Environmental Protection Agency, Municipal Environmental Research
Lab, Cincinnati, Ohio.
Tc*uingas, R., G. Voitmer, and R. Bednarik (1977) Direct fluorometric
analysis of aromatic polycyclic hydrocarbons on thin layer
chromatograms, Sci. Total Environ., 7, 261-267.
U.S. Environmental Protection Agency, Resource Conservation and
Recovery Act, Subtitle C, Background Document, “Petroleum Refining”,
U.S. Government Printing Office, Washington, D.C., 19 May 1980.
Vianna, N.J. and A.K. Polari (1984) Incidence of low birth weight among
Love Canal residents, Science, 226, 1217—1219.
Warner, J.F., B.J. Ridy, G.A. Jungclaus, MM. McKown, M.P. Miller, and
R.M. Riggin (1981) Development of a method for deten ining the
leachability of organic compounds from solid wastes, in: RA. Conway
and BC. Malloy (Eds.) Hazardous Solid Waste Testing: First Conference,
American Society for Testing and Materials, Philadelphia, pp. 40—60.
Warren, C.S. (1981) Hazardous Waste: An Overview, in: J.P. Collins
and W.D. Saukin (Eds.) The Hazardous Waste Dilemma: Issues and
Solutions, American Society of Civil Engineers, New York, pp. 5—15.
2-168
-------
FIGURE 1: SEPARATION OF MUTAGENIC COMPOUNDS BY THE TLC/SALMONELLA
ASSAY.
Samples were spotted at increasing volumes across the plate,
chrotnatographed with chloroform, and tested for mutagenic activity.
Sample application points and chromotographed spots are indicated (see
text for discussion). Negative control plates (spontaneous revertants)
are also illustrated. o, application point, , solvent front.
(a) Organic Sludge (0.25, 0.5, 1.0, 2.0 ul) tested with TA98, —S9.
(b) EDC Spent Caustic (2, 4, 6, 8, ul) tested TA100, —S9.
Plate was broken during excision of top corners.
(c) Herbicide Manufacturing Acid ( 0.5, 1.0, 2.0, 4.0 ul) tested with
TA100, +S9.
(d) Coal Gasification Tar (unknown volumes) tested with TA98, +S9.
2-169
-------
TABLE I
DESCRIPTION OF HAZARDOUS WASTE SAMPLES
Waste
Physical
slate
pH a
-Total
solids’
( )
Description
Organic sludge
Liquid
NM
NM
Sludge from an industrial solveni
recO%ery process
Organic sludge, spiked
Liquid
NM
NM
Sludge described abose. diluted and
fortified with ii organics
EDC spent caustic
(walar portion)
Liquid
3.2
cl .0
Clear, tan liquid waste horn ethylene
dieblonde production
Herbicide mfg. acid
Liquid
0.5
42
Thin, gray aqueous suspension from an herbicide production process
Coke plant waste
Liquid
&o
4
Thin, brown fluid with suspended solids.
95 DM50-soluble, water-in,olvble
Ink pigment sludge
Semi-solid
8.7
14
Pourable black jelly from ink production
Lates paint mfg. waste
Suspension
7.5
41
Thick polymeric emulsion of pigment, 5O
soluble in DMSO
Coal gasification tar
Tar
‘
8.1
79
Black, viscous tar produced during coal
gasification
Organic still bottoms
Solid
3.8
83
Thick, black slurry from tank bottoms
Dew atered municipal
sludge
Solid
7.5
44
Dark, wet cake of chemically treated
municipal sludge
Oil-refining waste
Semi-solid
7.0
59
Dark liquid with black flcccuknt and
oil droplets, DMSO- and water-soluble
Analysis performed by Research Triangle Institute (Sparacino, 1983).
NM, not measured.
2—170
-------
TABLE 2
RESULTS FROM THE TLC/SALMONELLfr 5 ANALYSIS OF HAZARDOUS WASTES
%k’a te
Extract a
TLC plate
dose range
(,.L I/application)
Mutagencity
TA98
TA IOO
+S9
—S9
+S9
—S9
Organic sludge
Organic sludge.
NEAT
NEAT
0.25—2
0.25—16
+ b
+ c
+ d
+ d
.
spiked
EDC spent
NEAT
0.5-8
—
+ C
+ b.c
caustic
Herbicide
NEAT
0.1—4
+ b.c
b.c
mfg. acid
Coke plant waste
NEAT
EtOH
DMSO
0.5-10
0.5-6
0.5-8
+ C
—
b
—
d
Ink pigment
NEAT
0.5—4
—
—
mc i
m d
sludge
Latex paint
NEAT
0.5—12
—
—
—
—
mfg. waste
Coal gasification
NEAT
unknown
+ b.c
+ b
+ d
tar
-
Organic still bottoms
Dewatered municipal
EtOH
EtOH
0.1—8
1—16
—
+ d
—
—
+ d
—
sludge
Oil-refining waste
NEAT
0.5—4
+ d
T
—
—
a Extract: NEAT, unadulterated; EtOH , ethanol; DMSO, dimethyl sulfoxide.
b 2-fold or greater incTease in revertant counts (for the total plate) over spontaneous background values.
Clustering of colonies; distinct localization.
d Clustering of colonies: localization detected b grid analysis.
md. inconclusive findings.
T. toxic.
TABLE 3
RESULTS FROM TLC/SALMONELLA ANALYSIS OF MODEL COMPOUNDS
Compound
Responsive
dose range
(jAg/application)
Mutage
mc activity
Result
Strain
Activation
Sodium azide
2-16
+ b
TA IOO
— S9
MMS
2-16
+ b
TA IOO
—S9
-Propiolactone
100—400
+ b
TA100
—S9
2-Nitrofluorene
2—16
+ C
TA98
—S9
( ‘yclophosphamide
2-16
20-80
weak +
+C
C
TA IO O
TA IO O
+S9
+S9
2-AAF
2-16
20—80
+ C
+c
TA9S
TA I OO
+S9
+S9
Daunomycin
0.125—1
2—16
.
+ C
toxic
TA98
TA9S
—S9
—S9
.
Strcptozotocin
2—16
+ d
TA IOO
— S9
4.NQO
MNNG
0.5—16
0.5—16
2-16
+ d
+ d
+ d
TA98
TA IOO
TA IO O
—S9
—S9
—S9
2-Anihramine
2—16
2—16
+ d
+ d
TA98
TA100
.
+ 59
+S9
Benzo(o pyrene
2—80
2—80
—
—
TA98
TA IO O
-f S9
+S9
Epichlorohydnn
2-200
—
TA IOO
—S9
Dimethylcarbamyl chloride
2—2000
—
TA IOO
— S9
a Responsive dose range does not necessarily reflect the lowest possible or highest achievable dose. It merely indicates those doses
tesied.
b Profusion of reveriant colonies across the plate.
Localized clusters of reveriant colonies at points of sample application.
d Dose-dependent ring of revertant colonies around a central toxic zone.
2-171
-------
(a)
(c)
(b)
(d)
-------
(b)
FIGURE 2: TLC/SAU4ONELLA ANALYSIS OF THE HERBICIDE MANUFACTURING ACID
C}IROMATOGRAPHED WITH DIFFERENT SOLVENT SYSTEMS
Fluorescent zones (denoted by a dashed line) are located around sample
application points in (a) and midway up the plate in (b). Negative
control plates (spontaneous revertants) are also illustrated.
(a) Sample Volumes: 0.1, 0.25, 0.5 ul
Mobile Phase: Chloroform
Mutagenicity Testing: TA100 with activation
(b) Sample Volumes: 0.1, 0.25, 0.5 ul
Mobile Phase: DCM/ETOM/Water, 10:20:1
Mutagenicity Testing: TA100 with activation
o, application point; - - - -, solvent front.
(a)
2—173
-------
METHODOLOGY FOR EVALUATING POTENTIAL HUMAN HEALTH EFFECTS
OF MICROORGANISMS THAT DEGRADE HAZARDOUS WASTES
Susan E. George, Michael J. Kohan, U.S. Environmental Protection
Agency, Health Effects Research Laboratory, Genetic Toxicology
Division, Research Triangle Park, North Carolina; Debra B. Walsh, Larry
D. Claxton, Environmental Health Research and Testing, Research
Triangle Park, North Carolina
ABSTRACT
Microorganisms are being developed to environmentally degrade hazardous
wastes. Before such organisms are deployed, methods need to be
designed to monitor the organisms and waste by—products for potentially
adverse human health and environmental effects. Initial work in our
laboratory involved the study of the biological effects of a mutant
microorganism that biodegrades polychiorinated biphenyl compounds
(PCB’s). We have designed several mouse model systems to examine the
ability of these organisms to colonize the intestines of the mouse and
compete with the resident intestinal flora. These include two 48 hour
studies with and without ampicillin pretreatment, a 14 day study in
conventional and metabolism cages, and exami atiog of the 9 fecal
material. Mice were dosed by gavage with 10 , 10 , or 10 organisms
and sacrificed at time intervals up to 48 hours following exposure. The
hazardous waste degrading microorganisms, Pseudomonas aeruginosa , was
recoverable on a selective medium in all methods evaluated. The normal
intestinal flora was also monitored. This included enumeration of the
lactose positive enteric microorganisms, total aerobic and anaerobic
counts, and the obligately anaerobic, predominantly Grain—negative rod
populations. Depending on the method utilized, an alteration in one or
more of the normal flora was observed in a dose related manner.
INTRODUCTION
Microorganisms are being isolated and developed to biodegrade hazardous
wastes such as polychlorobiphenyls (PCB) (18, 24), formaldehyde (18),
chlorobenzoates (10, 4) and pesticides (21, 20). The ability of
microorganisms to detoxify the target compounds is being enhanced
through mutagenicity or recombinant DNA techniques. The ultimate goal
is to apply these organisms to a site contaminated with an
environmentally hazardous compound and metabolize it to a nonhazardous
state, in situ . The need to transport the compound to a central
location such as a landfill or incinerator would be unnecessary, thus
eliminating the associated risks.
Very little research has been reported on the potential human health
effects associated with these biodegradative microorganisms. The
proposed release of these bacteria constructed through the use of
recombinant DNA techniques into the environment has received national
attention (6, 7, 9). Federal and state agencies now regulate the
2—175
-------
environmental application of these bacterial strains (17, 19). The
purpose of this study is to propose and evaluate methods that will
examine the potential human health effects of these recombinant,
mutant, and naturally occurring microorganisms used to detoxify
hazardous environmental chemicals. The methods presented in this study
examine the potential of these biodegradative microbes to colonize the
intestines of mice and detexinine how they compete with the normal
flora. An alteration of the normal flora may indicate a change in the
intestinal metabolism that could have an effect on digestion,
absorption, and the overall physiology of the host. The results
reported depict the competition and colonization potential of a PCB
degrading microorganisms, P. aeruginosa , that was isolated from a
commercially available product for PCB degradation. Manuscripts
submitted and in preparation describe this research in greater detail.
MATERIALS AND METHODS
Chemicals . All chemicals utilized in this study were obtained
commercially and of reagent grade. Vancoinycin MCi, kanamycin sulfate,
ampicillin, hemin, and vitamin K 1 were purchased from Sigma Chemical
Company, St. Louis, MO. Mercuric chloride was obtained from
Mallinckrodt, Paris, KY.
Media. P. aeruginosa stain BC16 was enumerated on Pseudomonas
isolation agar (Difco, Detroit, MI) containing mercuric chloride
(ffgCl 2 ) (30 ug/mi). MacConkey agar (Difco) was prepared according to
the manufacturer’s direction. Brucella blood agar (BEA) was prepared
as follows: Brucella agar (BBL Microbiology Systems, Cockeysville, MD)
as directed, 5% (v/v) laked, defibrinated sheep blood (Environmental
Diagnostics, Burlington, NC), vitamin K 1 (1.0 ug/mi) arid hemin (5
ug/m1 3 0 926, 11). Where indicated, vancomycin HCL (7.5 ug/mi) and
kanamycin (75 ug/ini) were added to BBA after autoclaving. Yeast
extract-tryptone (YT) (15) medium without added NaCl was used as
culture medium where indicated.
Bacterial strains. P. aeruginosa was isolated from BI-Chem 1006 PB
(Sybron Chemicals, Inc., Salem, VA) as directed by the manufacturer.
Bacterial strains in this product were designed to biodegrade
polychlorinated bipheriyl compounds. Identification was confirmed by
Gram straining and using the API2OE method (Analtab Products,
Plainview, NY).
culture conditions . All incubations for bacterial growth were done at
37°C. Inocula for dosing was prepared by sedimenting an overnight 25
ml YT culture of P. aeruginosa . The cells were resuspended in 5 ml
Dulbeccos phosphate buffered saline (PBS) (GIBCO Laboratories, Chagrin
Falls, OR) prior to animal dosing. All other incubations were done
aerobically for 48 h or anaerobically for 72 h.
2—176
-------
Animal studies . Sixty—day old strain CD-i male mice (Charles River
Laboratories, Raleigh, NC or Kingston, NY) were used in the studies.
Animals were given sterilized food (Sterilizable Rodent Biox,
Continental Grain Company, Chicago, IL) and waster ad libitum . Animals
were housed in cages with wood shaving bedding unless otherwise
indicated. When metabolism cages (Nalge Company, Rochester, NY) were
employed, animals were fed sterile pelietized chow (Dustless Pellets—45
mg. R.C., Bioserv Inc. 13 Frenghtown , J) and water ad libitum . Animals
were administered 0, 10 , 10 , or 10 CFU/ml of P. aeruginosa by
gavage. At time intervals (0, 3, 6, 12, 24, 48 for the 48 h study, 14
days for the metabolism cage study, and 3, 12, 24, and 48 h for the
ampicillin pretreatment 48 h study) the animals were sacrificed and the
intestines surgically removed and immediately placed into 5 ml reduced
PBS. The intestines were homogenized using a Polytron tissue
homogenizer (Brinkman Instruments, Westhury, NY) under a constant
stream of nitrogen and then placed into an anaerobic chamber (Coy
Laboratory Products, Ann Arbor, MI) containing an atmosphere of 5% Co 21
10% H 2 , and 85% N . Serial dilutions of the homogenate were made in
reduced PBS and ah anaerobic plating and incubation was done in the
chamber. Aerobic plating was done aerobically on selective media
(Figure 1.)
Where indicated, feces were collected from regular cages on a daily
basis (1, 2, and 3 days) or metabolism cages at 0, 1, 2, 4, 6, 8, 10,
12, and 14 days after dosing. Approximately 1.0 g of fecal material
was placed into 5 ml PBS, vortexed, and the dosed microorganisms
selected on Pseudomonas isolation agar containing HgC1 2 (30 ug/mi).
Three independent animal studies were performed using the above
protocol. The first examined the survival and competition potential of
the dosed microorganism over a period of 48 h. Results are an average
of values enumerated from four animals per dose—time combination. The
second study compared the flora from animals housed in regular (five
animals per dose—time combination) cages for 14 days after dosing. In
the third study, animals were pretreated with 5 mg of ampicillin on
days 1 and 2 then 1 mg of ampicillin on day 3 prior to dosing with the
microorganism (four animals per dose—time combination). A general
outline for the animal studies is shown in Figure 1.
Statistical analysis . Analyses of variance and T—tests were performed
using the RS/l statistical software package (BBN Software Products
Corporation, Irving, Texas) on an IBM—AT personal computer.
RESULTS
Recovery of P. aeruginosa in the feces . Fecal material was collected
from the cages over a three day period. Recovery of the microorganism
was examined by selection on Pseudomonas isolation agar containing
HgC1 2 . The results are shown in Table 1. P. aeruginosa was recovered
only from animals at the higher dose.
2-177
-------
A more i depth study was done in metabo1 sm cages, using only the high
dose, 10 CFU, where the fecal material was collected and the organisms
enumerated. P. aeruginosa was recovered form the fecal material for
only two days after dosing (Table 1).
Survival of P. aeruginosa in the gastrointestinal tract . A more
quantitative assay for the survival of the dosed microorganism was
needed so the intestines were analyzed for recovery and competition.
Three separate studies were performed to look at the competition of the
PCB degrading microorganism in the mouse intestine: 1) enumeration
over 48 hours, 2) ampicillin pretreatment followed by enumeration over
48 hours, and 3) enumeration after 14 days in regular and metabolism
cages. In both 48 hour studi s, the survival of the organism was
similar with co nts in the 10 CFU/g intestines range or animals
administered 1 l0 CFU (Table 2). Animals dosed with 10 CFU yielded
less than 10 CFU/g intestines after 48 hours in both ampicillin
treated and untreated animals.
In the 14 day study, P. aeruginosa appeared to olonize the 4
gastrointestinal tract in animals dosed with 10 CFU yielding 4.0 X 10
CFtJ/g intestines (Table 2). Animals housed in metabolism cages, where
fe a1 material was not in contact with the animals, yielded less than
10 CFU/g intestines.
Effect of P. aeruginosa on the normal intestinal flora . Four
populations of normal flora were monitored in the 48 hour and 14 day
studies. They included 1) lactose position enteric organisms, 2) total
aerobic counts, 3) total obligately anaerobic predominantly Gram-
negative rods, and 4) total anaerobic counts.
In the 48 h ur study, the lactose positive counts ranged from 5.0 X
to 3.8 X 10 CFIJ/g intestines. Results from the 14 day study gave
number of similar magnitude. Ampicillin pretreatnient 7 markedlY 9
increased these counts yielding a range from 5.4 K 10 to 4.5 X 10
CFU/g intestines. The results are tabulated in Table 3.
The total aerobic count, which included aerobes and facultative
anaerobes was also increased in animals pretreated with ampicillin.
The counts are shown in Table 4. Ainpicillin pretreatment selected for
several different species including Enterobacter cloacae (predominant),
Proteus vulgaris , and Lactobacillus spp. Though numbers were lower in
the animals not pretreated with ampicillin, a more diverse group of
organisms was observed and 5 are y were E. cloacae or P. vulgaris
observed. Values in the 10 —10 CFU/g intestines range were observed
in the animals without antpicill3n t eatment whereas those that received
anpicillin had counts in the 10 —10 CFU/g intestines range.
unpici11in pretreatment greatly reduced the obligately anaerobic Grain-
negative rod counts (Table 3). These organisms were selected for by
resistance to the antibiotics kanamycin and vancomycin. included in
2-178
-------
this group were Bacteroides spp. and Fusobacterj,uin pp. (8).
Ampicillin treatment reduced the counts 1 fro 10 —10 CFU/g intestines
in the 48 hour and 14 day studies to 10 —10 CFU/g intestines in the
ampicillin pretreatment study. Results are shown in Table 5.
The total anaerobic counts from mice dosed with P. aeruginosa are
tabulated in Table 6. Included in this count were facultative and
strict anaerobes. Facultative organisms included Streptococcus spp.,
Staphylococcus spp., Lactobacillus spp., and organisms from the family
Enterobacteriaceae (2). Strict anaerobes included the Clostridia,
Eubacteria, Bacteroides spp., and the Fusobacteria as well as others
(8, 2, 16).
In the 48 hour study, counts range from 2.2 X 108 to 8.7 x io8 cFU/g
intestines. The 14 day study in regular and metabolism cages showed
similar values. Again, ampicillin i creased the observed c?unts.
Values ranged from a low of 2.7 X 10 to a high of 2.2 X 10 CFU/g
intestines. The flora enumerated were similar to that observed in the
total aerobic counts.
In order to determine whether the presence of P. aeruginosa caused an
alteration in the normal murine intestinal flora, statistical analyses
were performed on the data collected in Tables 3—6. P—values are
reported in Table 7.
DISCUSSION
This study was designed to develop and evaluate methods to determine
some of the indirect adverse human health effects potentially
associated with microorganisms that degrade hazardous wastes. Four
different studies were proposed, three of which involved enumeration of
the biodegrading microorganisms from the gastrointestinal tract and one
study that examined the fecal material for the organism’s presence.
Further effects on the normal flora were investigated. They included
enumeration of 1) the enteric lactose positive microorganisms, 2) the
total aerobic counts, 3) obligately anaerobic Gram—negative rod counts,
and 4) the total anaerobic counts.
The initial study involved the examination of the mouse fecal material
for the presence of the dosed biodegrading microorganism. This method
was utilized by Levy and Marshall (13) and Levy et al. (14) to
determine colonization of recombinant Escherichia coli strain in mice.
Similar studies were done in human volunteer (1, 14, 22, 23). The
results were not very promising wi h recovery being observed in regular
cages only at the highest dose, 10 CFU (Table 1). Further studies in
metabolism cages extended the recovery time to two days at the same
dose. Several questions arose such as whether oxygen or dessication
had an effect on the microbial content. In order to avoid these
potential problems, the mouse intestines were examined.
2—179
-------
None of the mice dosed by gavage with P. aeruginosa appeared unhealthy
during these studies. No diarrhea was observed and animal weights did
not decrease. No abnormal pathology was observed. Therefore, all of
these methods were able to evaluate the direct health effects on mice
such as morbidity or mortality.
In the 48 hour study, p. aeruginosa was recoverable from the mouse
intestines at all three doses. By the end of 48 hours, organisms were
countable on the selective medium.
Ampicillin pretreatment was incorporated into the study in order to
alter the normal flora and enhance the survival of the biodegradative
microbe, which was resistant to ampicillin. Cohen et al. (5) and Laux
et aL (12) used streptomycin to select for colonization by E. coli
derivatives. Ampicillin pretreatment did not appear to have a
selective effect on the survival of strain BC16 at 48 hours. Survival
counts were similar for both the experiments (Table 2). An increase in
survival at the other time points was noted in ampicillin treatment
animals.
A difference in normal flora counts was observed in the two 48 hour
experiments. An increase in total counts was observed for the
ainpicillin treated animals in all populations monitored with the
exception of the obligately anaerobic Gram—negative rods (Table 3—6).
Antibiotics have been reported to alter the normal flora in the
gastrointestinal tract populations (3, 25). Ainpicillin treatment
decreased those counts by at least four orders of magnitude (Table 5).
The increase in the other counts was due to an increase in the E.
cloacae population. This organism is facultative so therefore can grow
on MacConkey agar and Brucella blood agar both anaerobically and
aerobically.
The resulting alterations in the normal flora between the two 48 hour
studies differed. A dose effect (p less than 0.05) was observed for
the lactose positive and obligately anaerobic Grain—negative rod counts
in animals without ampicillin pretreatment (Table 7). The total
anaerobic population counts were affected (p less than 0.1) in the
ainpicillin treatment experiment (Table 7).
Survival and alteration of the intestinal flora was also examined at
the end of 14 days in animals housed in regular and metabolism cages.
This provided resul s from a longer term of study. P. aeruginosa was
recovered (4.0 X 10 CFU/g intestines) 14 days after dosing in animals
housed in regular cages (Table 2). The results varied from those
observed in animals housed in metabolism cages. Mice are coprophagic
and the observed results were probably due to this phenomenuin. Again
the alterations in the normal flora were different. The obligately
anaerobic Grain—negative rod and total anaerobic populations showed a
significant different (p less than 0.1) upon exposure to P. aeruginosa
(Table 7). Only the total aerobic populations were affected (p less
than 0.1) when mice were housed in regular cages.
2-180
-------
The comparison of intestinal flora from animals housed in regular or
metabolism cages provided a good model to compare repeated exposure
versus a single dose. The organisms that were excreted were re—
administered through this coprophagic behavior. Because all fecal
material was separated from the animals in the metabolism cages, no re—
dosing could occur through this route.
The 48 hour study provided a short term method for examining the fate
of the biodegrading microorganism and its effect on the normal flora in
both conventional mice as well as ones whose normal flora had been
altered. The ampicillin treatment may give researchers insight into
how compromized humans (those on antibiotic therapy or
immunosuppressant drugs) may be effected upon exposure to this class of
microorganisms.
These studies provide methods to examine the colonization and
competition potential of microorganisms that degrade hazardous wastes.
If the normal flora are altered due to the presence of these strains,
then the effect on intestinal flora metabolism, absorbtion, and the
host’s physiology need to be considered. In vitro methods are
currently being developed to investigate colonization, competition and
metabolic changes associated with exposure to the biodegraders using
both mouse and human intestinal flora.
ACKNOWLEDGEMENTS
We are grateful to Ms. Angalena Monroe for her technical assistance in
the laboratory and with the data entry.
We would also like to thank Sybron Chemicals, Inc. for supplying our
laboratory with a sample of Bi-Chems 1006 PB.
The research described in this article has been funded wholly or in
part by the Health Effects Research Laboratory, U.S. Environmental
Potential Agency. Although this article has been reviewed by the
Health Effects Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC, it does not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
2—181
-------
REFERENCES
1. Anderson, ES. 1975. Viability of, and transfer of a plasmid
from, B. coil 1(12 in the human intestine. Nature 255:502—504.
2. Benno, Y., K. Suzuki, K. Suzuki, K. Narisawa, W. R. Brucwe, and T.
Mitsuoka. 1986. Comparison of the fecal microflora in rural
Japanese and urban Canadians. Microbiol. Inununol. 30:521-532.
3. Burton, G.C., D.C. Rirsh, D.C. Blenden, and J.L. Zeigler. 1974.
The effects of tetracycline on the establishment of Escherichia
coli of animal origin, and in vivo transfer of antibiotic
resistance, in the intestinal tract of man, p. 241—253. In F.A.
Skinner and J.G. Carr (eds.), The Normal Flora of Man. Academic
Press, New York, NY.
4. Chattergee, D.K., S.T. Kellogg, S. Hamada, and A.M. Chakrabarty.
1981. Plasmid specifying total degradation of 3—ch].orobenzoate by
a modified ortho pathway. J. Bacteriol. 146:639-646.
5. Cohen, P.S., R.W. Pilsucki, M.L. Myhal, C.A. Rosen, D.C. Laux, and
V.J. Cabelli. 1979. Colonization potentials of male and female
B. coil 1(12 strains E. coli B and human fecal E. coli strains in
the mouse CI tract. Recomb. DNA Tech. Bull. 2:106-113.
6. Coiwell, R.R., P.R. Brayton, D.J. Grimes, D.B. Roszak, S.A. Hug,
and L.N. Palmer. 1985. Viable but non—culturable Vibrio
chlorerae and related pathogens in the environment: Implications
for release of genetically engineered microorganisms.
Bio/Tech.nology 3:817—820.
7. Day, P.R. 1985. Engineered organisms in the environment: A
perspective on the problem, p. 4-10. In H.O. Halvorson, D.
Pramer, and M. Rogul (eds.), Engineered Organisms in the
Environment: Scientific Issues. American Society for
Microbiology, Washington, DC
8. Draser, B.S. and P.A. Barrow. 1985. The bacterial flora of the
normal intestine, p. 19—41. In J.A. Cole, C.J. Knowles, and D.
Schlessinger (eds.), Intestinal Microbiology, Aspects of
Microbiology 10. l merican Society for Microbiology, Washington,
DC
9. Fox, J. 1985. Anticipating deliberate release. A growing
intellectual industry. ASM News 51:619—620.
10. Hartman, J., W. Reineke, H.—J. lcnackmuss. 1979. Metabolism of 3—
chioro—, 4—chloro—, and 3,5-dichlorobenzoate by a Pseudomonad.
Appl. Environ. Microbiol. 37:421—428.
2—182
-------
11. Holdeman, L.V., E.P. Cato, and W.E.C. Moore, eds. 1977. Anaerobe
Laboratory Manual, 4th Edition. VPI Anaerobe Laboratory,
Blacksburg, VA
12. Laux, D.C., V.J. Cabelli, and P.S. Cohen. 1982. The effect of
plasmid gene expression on the colonizing ability of E. coli HS in
mice. Recomb. DNA Tech. Bull. 5:1-5.
13. Levy, S.B., and B. Marshall. 1981. Risk assessment studies of E.
coli host—vector systems. Recomb. DNA Tech. Bull. 4:91—98.
14. Levy, S.B.., B. Marshall, D. Rowse-Eagle, and A. Onderdonk. 1980.
Survival of Escherichia coli host—vector systems in mammalian
intestine. Science 209:391—394.
15. Miller, J.H. 1972. Experiments in Molecular Genetics. Cold
Spring Harbor Laboratory, Cold Spring Harbor, NY
16. Moore, W.E.C., E.P. Cato, and L.V. Holdeman. 1978. Some current
concepts in intestinal bacteriology. Amer. J. din. Nutrit.
31:533—542.
17. Office of Science and Technology. 1986. Coordinated framework
for regulation of biotechnology; announcement of policy and notice
for public comment. Federal Reg. 51:23302-23393.
18. Powledge, T.M. 1983. Prospects for pollution control with
microbes. Bio/Technology 1:743-755.
19. Rissler, J.F. 1984. Research needs for biotic environmental
effects of genetically—engineered microorganisms. Recoxnb. DNA
Tech. Bull. 7:20—30.
20. Serdar, C.M., D.T. Gibson, D.M. Munnecke, and J.H. Landcaster.
1982. Plasmid involvement in parathion hydrolysis by Pseudomonas
diminuta . Appi. Environ. Microbiol. 44:246—249.
21. Sheela, S. and S.B. Pai. 1983. Metabolism of fensulfothion by a
soil bacterium, Pseudomonas alcaligenes Cl. Appi. Environ.
Microbiol. 46:475—479.
22. Smith, H.W. 1975. Survival of orally administered E. coli K12 in
alimentary tract of man. Nature 255:500—502.
23. Smith, H.W. 1978. Is it safe to use Escherichia coli Kl2 in
recombinant DNA experiments? J. Infect, disease 137:655-660.
24. Straley, J.P., J.A. Christiansen, and A.L. Kopecky. 1982.
Improved biological degradation of chlorinated hydrocarbons using
mutant bacteria. Proc. Internati. Water Conf. 43:523—532.
2-183
-------
25. Sutter, V.L., and S.M. Firiegold. 1974. The effect of
antimicrobial agents on human faecal flora: studies with
cephalexin, cyclacillin and clindamycin, p. 299—240. In F.A.
Skinner and J.G. Carr (eds.), The Normal Flora of Man. Academic
Press, New York, NY
26. Sutter, V.L., V.L. Vargo, S.M. Finegold, and X.S. Bricknell.
1975. Wadsworth anaerobic bacteriology manual, 2nd edition. The
Regents of the University of California, Los Angeles, CA
2-184
-------
Table 1. Recovery of P. aeruginosa from mouse fecal material.a
Time(days) 0 i0 io 6 1o
Regular 1 _b +
Cages
2
3
Metabolism 0 ndC rid
Cages
1 nd rid +
2 nd nd +
3 rid rid
4 nd nd
6 nd nd
8 nd nd
10 rid nd
12 nd rid
14 nd
a Mice were dosed with P. aeruginosa and at designated time intervals,
approximately 1.0 gram of fecal material was collected and placed
into 5.0 ml PBS. A 0.1 ml aliquot was plated onto Pseudomonas isolation
agar containing HgC1 2 , incubated 24 hours, arid the organism enumerated.
b (—), organism detected; (+), growth of dosed strain was evident.
c not determined
2-185
-------
Table 2. Survival of P. aeruginosa in rrcuse intestines.a
Dose (CFU/animal)
Time(h) 0 i0 i 6
48 hour 0
-------
Table 3. Total lactose positive counts in P. aeruginosa dosed mice:
enumeration from an intestinal homogenate.&
Dose (CFU) x iO
Time 0 106
(h) CPU x i0 4
48 hour 0 18 b 37 b
study
3 6.1 2.4 0.5 2.4
6 1.2 10.1 5.5 12.8
12 1.9 0.5 0.8 1.4
24 5.9 3.6 4.1 2.3
48 0.5 1.5 0.6 7.1
(h) CFU x i0 7
48 hour 3 156.1 67.9 184.3
study
ampicillin 12 110.8 86.7 n.d. 95.9
pretreatment
24 449.7 5.4 n.d. 202.3
48 273.1 189.8 n.d. 337.9
(days) CPU x i0 3
14 day study
metabolism 14 n.d. n.d.
regular 14 1 0 e n.d. n.d.
a Mice were dosed with 0, 106, or CFU and sacrificed at indi-
cated time intervals. The intestines were removed, homogenized, and
dilutions of the homogenate prepared in PBS. A 0.1 ml. aliquot of
the appropriate dilution was spread plated onto MacConkey Agar and
plates were incubated 24—48 hours. Counts are an average of values
from 4 animals unless otherwise indicated.
b Counts are an average of values from 2 animals.
c not determined
d Counts are an average of values from 10 animals.
e Counts are an average of values from 5 animals.
2—187
-------
Table 4. Total aerobic counts in P.aeruginosa dosed mice:
enumeration from an intestinal homoqenate.a
Dose
48 hour
study
48 hour
study
ampicil].in
pretreatment
14 day study
Time
0
i 6
(h)
CPU x io
0
4 • 8 b
60 b
26 b
150 b
3
5.8
234 b
9.1
12.9
6
2.9
45 b
14.7
7.1
12
1.5
2 2 b
1.7
1.5
24
9.8
12 • 9 b
13.1
10.8
48
7.3
17 • 2 b
12.6
225
(h)
CPU x IO
3
200 1 b
876 b
m.d.
2478 b
12
1162 b
964 b
n.d.
107 0 b
24
7861 b
80 b
n.d.
3115 b
48
411 gb
240 • 8 b
m.d.
5291 b
(days)
CPU x
regular
14
0 • 22 d
n.d.
n.d.
O.23
14
078 e
m.d.
n.d.
265 e
a Mice were dosed with 0, i0 , 106, or cFu and sacrf iced at indi-
cated time intervals. The intestines were removed, homogenized, and
dilutions of the homogenate prepared in PBS. A 0.1 ml. aliquot of
the appropriate dilution was spread plated onto Brucella Blood Agar
and plates were incubated 48 hours. Counts are an average of values
from 4 animals unless otherwise indicated.
b Counts are an average of values from 2 animals.
c not determined
d Counts are an average of values from 10 animals.
e Counts are an average of a values from S animals.
2-188
metabolism
-------
Table 5. Total obligately anaerobic predomiantly Gram negative rod counts in
P.aeruginosa dosed mice: enumeration from an intestinal homogenate.a
Dose (CFU)
Time 0 io6 10
(h) CFtJ x io8
48 hour 0 33 b 18 b 20 b
study
3 47 b 2 • 9 b 18 b 31 b
6 36 b 37 b 38 b 4 5 b
12 09 b 13 b 17 b 2.6
24 20 b 13 b 2.0 2.4
48 11 b 11 b 2.3 2.4
(h) CFU x i0 2
48 hour 3 >i0 4 16.3 fl.d.C 10
study
ai tpicillin 12 35.8 0.5 n.d. 3.4
pretreatment
24 180.2 10.4 n.d. 683.9
48 >io 4 672.3 n.d. 663.6
(d) CFU x io
14 day study
metabolism 14 1108 d n.d. n.d.
regular 14 880 d n.d. n.d. 1388 d
a Mice were dosed with 0, iO , 106, or CFU and sacrificed at in-
dicated time intervals. The intestines were removed, homogenized,
and dilutions of the homogenate prepared in PBS. A 0.1 ml. au-
quot of the appropriate dilution was spread plated onto Brucella
Blood Agar supplemented with vancomycin and kanamycin. Counts are
an average of values from 4 animals unless otherwise indicated.
b Counts are an average of values from 2 animals.
C not determined
d Counts are an average of values from 5 animals.
2-189
-------
Table 6. Total anaerobic counts in P.aeruginosa dosed mice: enumeration
from an intestinal homogenate.a
o io 3 io 6 10
2 3 b
79 b
48 b
40 b
6.9
6 0 b
4.2
6.4
6.2
8 7 b
7.6
6.2
12
2.2
2 4 b
2.4
2.6
24
2.6
2 • 7 b
3.5
4.1
48
34
35 b
5.9
6.3
48 hour
study
ampicillin
pretreatment
845 d n.d. n.d. 3.25
514 d n.d. n.d. 3.84
a Mice were dosed with o, io , i0 6 , or i0 CFU and sacrificed at indi-
cated time intervals. The intestines were removed, homogenized, and
dilutions of the homogenate prepared in PBS. A 0.1 ml. aliquot of
the appropriate dilution was spread plated onto Brucella Blood Agar.
Counts are an average of values from 4 animals unless otherwise in—
cated.
b Counts are an average of values from 2 animals.
C not determined
d Counts are an average of values from S animals.
(h)
Time Dose
48 hour
study
0
3
6
7
(h)
CFU x 10
3
131.8
2.7
n.d.C
203.4
12
145.6
42.1
n.d.
117.6
24
145.7
9.4
n.d.
105.9
48
203.8
223.5
n.d.
192.6
Cd)
CFU x10 8
14 day study
metabolism 14
regular 14
2—190
-------
Table 7. P—values for the effects of P.aeruginosa dose on the normal flora.
ampicillin
14 day
pretreatment
study 1 ’
14
day
48 hour 48 hour
StUdYa studya
metabolism
cages
regular
cages
Lactose
positive
colonies
Total
aerobic
colonies
Obligately
anaerobic
Gram-
negative
rods
Total
anaerobic
colonies
Dose
Time
Nonadditivity
0.012
0.002
0.000
0.271
0.451
0.757
0.615
n.d.c
n.d.
0.338
n.d.
n.d.
Dose
0.292
0.270
0.234
0.100
Time
0.044
0.374
n.d.
n.d.
Nonadditivity
0.866
0.684
n.d.
n.d.
Dose
0.013
n.d.
0.027
0.236
Time
0.000
n.d.
n.d.
n.d.
Nonadditivity
0.347
n.d.
n.d.
n.d.
Dose
0.292
0.073
0.063
0.338
Time
0.044
0.079
n.d.
n.d.
Nonadditivity
0.866
0.513
n.d.
n.d.
a P—values from a two—way analysis of variance of
enumeration of the normal intestinal flora.
the data obtained from
b P—values from a T—test of the data obtained from enumeration of the normal
intestinal flora.
2-19 1
C not determined
-------
Figure 1. Protocol for animal study
Dose strain CD-i 60 day old
male mice with microorganisms
1
Sacrifice animals
Remove intestines and homogenize
under N2
Place homogenate into anaerobic
chamber and prepare dilutions
Plate 0.1 ml homogenate
dilution onto selective media
Anaerobic media Aerobic media
S Brucella laked blood agar • Pseudomonas isolation
• Brucella laked blood agar agar
+ vancomycin and • MacConkey agar
kanamycin • Brucella laked blood agar
2-192
-------
BIOASSAY DETERMINATION OF SOIL ASSIMULATIVE CAPACITY
Spencer A. Peterson, 3oseph C. Greene, William E. Miller, Corvallis
Environmental Research Laboratory, U.S. Environmental Protection
Agency, Corvallis, Oregon
ABSTRACT
Bacteria, earthworms and ants were used to define the assimulative
capacity of various soil types for heavy metals (Cu, Cd, Zn) wood
treating and petrochemical waste products. Mixed soil microbial
respiration activity to Copper increased proportionally (less toxic)
to the increase in organic content of the soils studied. Zinc
toxicity to the mixed microbial respiration activity was only reduced
in the high organic Rifle Soil. The earthworm Eisenia fetida was
exposed to varying concentrations of Cu, Cd, and Zn added to filter
paper and artificial soil substrates. Zinc toxicity was similar in
each substrate. The bioavailability of Cd and Cu to E. Fetida
decreased approximately two and four—fold respectively in artificial
soil relative to their avaliability in the filter paper aqueous
contact test. Harvester ants, Pogonomyrmex owyheei , were exposed to
1.0% (wt/wt) aqueous concentrations of wood treating, drilling fluid
and oil slop wastes, applied to filter paper and Ritzville
silt—loam—soil. The toxicity rank order was similar for both the
filter paper and Ritzville soil substrates. However, Ritzville soil
assimulative capacity ranged from a two—fold reduction of toxicity for
wood treating waste to a five—fold toxicity reduction for both the
drilling fluid and slop oil wastes. These results demonstrate the
capability of bioassays to define the assimulative capacity of soils
for selected waste products. They also show that the assimulative
process is very complex and that the capacity varies with the waste
product, and that it is not easily predicted from a soils, silt, sand,
clay, or organic content.
2-1 q3
-------
ENFORCEMENT
chairpersons
Kenneth Jennings Robert Stevens
Environmental Scientist Chief
Office of Waste Programs Enforcement California Department
U.S. EPA of Health Services
401 N Street, S.w. Hazardous Materials
Washington, D.C. 20460 Laboratory
2151 Berkeley Way
Berkeley, CP 94704
-------
RCRA LIP ND DISPOSAL RESTRICTION PROGRAM
Victor Hays, Office of Enforcement, U.S. Environmental Protection
Agency, Washington, D.C.
ABSTRACT
The Congress decided in its reauthorization of RCRA in 1984 to
restrict the disposal of various classes of hazardous wastes e.g.,
solvents and dioxins in lang—based units like landfills. EPA
promulgated a complex land disposal restriction rule to implement this
new law. Solvents and dioxins, for example, have been restricted from
land disposal since November of 1986.
Laboratories play an important role in this matter because wastes and
soil must be tested to ascertain whether they meet a specified
treatment standard for land disposal. The testing procedure is called
the Toxicity Characteristic Leaching Procedure (TCLJP). It is an
improved Extraction Procedure (EP) although it is questionable whether
it is an effective measure of leaching potential where hydrophobic
chemicals in natural soils are the concern.
The following is an overview of the Land Disposal Restrictions
program.
INTRODUCTION
On November 8, 1984 Congress passed the Hazardous and Solid Waste
.P xnendments (HSWA) to the Resource Conservation and Recovery Act.
I mong other things, these amendments require EPA to evaluate all
hazardous wastes according to a strict schedule to determine which
wastes should be restricted from land disposal. For wastes that are
restricted, the amendments require EPA to set levels or methods of
treatment which substantially diminish a waste’s toxicity or reduce
the likelihood that a waste’s hazardous constituents will migrate from
the disposal site. Beyond specified dates, restricted wastes which do
not meet treatment standards are prohibited from land disposal.
On November 7, 1986 EPA promulgated the first phase of the Land
Disposal Restrictions (LDR) by restricting the land disposal of
solvent (FOO1—F005) and dioxin containing wastes. This paper will
discuss development and implementation of this first phase of LDR.
In the final rule, EPA defined land disposal to include, but not be
limited to, any placement of hazardous waste in:
— Land fills
— Surface Impoundments
— Waste sites
— Injection wells
— Land treatment facilities
— Salt domes or salt beds
— Underground mines or caves
— Concrete vaults or bunkers
3 —1
-------
The LDR rule covers hazardous wastes placed in land disposal units
after the effective dates of the prohibitions. Wastes disposed of
before November 7, 1986 do not have to be removed from land disposal
for treatment. There are extensions and variances which extend th
effective date of particular wastes under certain circumstances.
HSWA required EPA to set levels or methods of treatment which
substantially diminish the toxicity of waste or reduce the potential
for migration of hazardous constituents. These levels or methods
referred to as treatment standards, un.ist minimize short and long term
threats to human health and the environment.
To establish treatment standards, EPA identified wastes with similar
characterizations (physical and chemical). EPA then categorized these
similar wastes into “waste treatability groups.” EPA then evaluated
identified technologies used to treat the wastes to determine the best
dei nstrated available technology (BDA T) for each waste treatability
group.
Once BD T is identified, EPA then established the treatment standards
as either a specific technology or as a performance level (i.e., the
concentration level of hazardous constituents that is representative
of treatment by BD T). For the November 7, 1986 rule covering
solvents and dioxins, this is expressed as a concentration level of
hazardous constituents in an extract of the waste using the Toxicity
Characteristic Leaching procedure ( LP).
The TCLP is an analytical method used to determine whether the
concentrations of hazardous constituents in a waste extract or an
extract of the treatment residual meet the applicable treatment
standards. EPA promulgated the TCLP for use only in solvents and
dioxins final rule, and only then treatment standards are expressed as
concentration levels of hazardous constituents in an extract. The
TCLP is an “improved” Extraction Procedure (EP) toxicity test. There
is some question of the LP effectiveness when testing hydrophobic
chemicals in natural soils.
In response to the LDR, the Office of Waste Programs Enforcement
embarked on a campaign of implementing and enforcing the new
regulations. This included development of a comprehensive LDR
inspection guidance manual to educate and prepare enforcement
personnel.
The LDR guidance manual is composed of 3 main sections. The first
section consists of the statutory and regulatory overview. This
covers requirements for each potential handler of restricted solvent
waste, as well as a brief discussion on applicable extensions and
variances. The second section focuses on major concerns with regard
to enforcing these new restrictions, It brings to light areas of
potential non—compliance and instructs on proper response and
documentation. The third section is an inspection checklist. This
checklist is comprehensive to include all possible situations
encountered at any type facility handling restricted solvent waste.
3—2
-------
The checklist gives the inspector a tool to track and record
information during an inspection and to gather necessary information
to determine a facility’s status of compliance with LDR.
A more detailed discussion on the LDR Guidance Manual as well as other
enforcement efforts will be presented.
3—3
-------
RCRA LPJBORA’IORY AUDIT INSPECTION (LAI) PROGRAM
Edwin Pryor, Department of Enforcement, U.S. Environmental Protection
Agency, Washington, D.C.
ABSTRACT
The Resource Conservation Recovery Act (R RA) requires the owner or
operator of a surface impoundment, landfill or land treatment unit
that is used to manage hazardous waste to implement a ground—water
monitoring program capable of determining a facility’s impact on the
uppermost aquifer. The Agency has developed guidance entitled, RCRA
Ground—Water Monitoring Technical Enforcement Guidance Document, which
details the technical aspects of ground—water monitoring system design
and operation deemed essential by the Agency in order for a
ground—water monitoring system to meet the goals of the RCRA program.
Once it has been established that owner/operators have adequately
designed and constructed their ground—water monitoring systems and
that these systems are providing representative ground—water samples,
EPA must confirm that these samples are being properly analyzed.
The Laboratory Audit Inspection (LAl) will be conducted to determine
the quality of owner/operator’s laboratory analyses. At present, the
LAX is envisioned to be a two phased inspection. The first phase will
determine whether the laboratory is properly staffed and equipped, if
there are adequate quality assurance/quality control procedures and
whether the samples are properly logged and tracked throughout the
laboratory. The second phase of the LAX, if needed, will be a more
detailed review of the analytical procedures being performed. This
phase of the inspection will require an intensive effort by qualified
chemists to evaluate the equipment and the methodologies that are used
for particular analyses.
INTRODUCTI
The Resource Conservation Recovery Act (RCRA) requires the owner or
operator of a surface impoundment, landfill or land treatment unit
that is used to manage hazardous waste to implement a ground—water
monitoring program capable of determining a facility’s impact on the
uppermost aquifer. The Environmental Protection Agency has developed
a guidance entitled, RCRA Ground—Water Monitoring Technical
Enforcement Guidance Document, which details the technical aspects of
ground—water monitoring system design and operation deemed important
by the Agency in order for a ground—water monitoring system to meet
the goals of the RCRA program. Once it has been established that
owner/operators have adequately designed and constructed their
ground—water monitoring systems and that these systems are providing
representative ground—water samples, EPA must confirm that these
samples are being properly analyzed.
The Office of Waste Programs Enforcement, RCRA Enforcement Division,
is developing a RCRA Laboratory Audit Inspection (LAX) program. The
goal of the inspection is to determine whether the laboratory that the
3—5
-------
owner/operator is using for ground—water sample analyses is properly
equipped, maintained and staffed, whether there are adequate quality
assurance/quality control procedures and whether samples are properly
logged and tracked throughout the laboratory.
A Workgroup of Headquarters, Regional, and State laboratory and
program personnel was assembled in March 1987 to begin development of
the LAX. The purpose of this outline is to summarize the scope of the
LAX as developed in the first workgroup meeting. The content and
format may change as the inspection is further developed.
XJ LINE
I) Introduction
A. Purpose
B. Intended audience
C. Relationship to the RCRA ground—water monitoring program
0. Resources available for conducting laboratory audits
E. Laboratory audits in other EPA programs
F. Not a laboratory certification program
G. Scope of the LAX
H. Relationship with other inspections
I. Overview of the inspection
II) Laboratory Audit Inspection Process
A. Schematic of the LAX process
B. Targeting of laboratories to receive an LAX
1) potential use of Performance Evaluation (PE) samples
2) results of other inspections
C. Review of owner/operator’s sampling and analysis plan
1) identification of the laboratory performing the analysis
2) parameters the laboratory is responsible for
3) review of the laboratory’s QA,VC plan
4) identification of detection limits
5) highlight activities that need to be confirmed during the
laboratory visit
6) uSe of the sampling and plan review checklist
D. Conducting the on—site inspection
1) initial briefing with laboratory management
2) laboratory walk—through
3) use of the laboratory audit checklist
4) debriefing at the conclusion of the inspection
E. Inspection report preparation
1) siinm ry of inspection findings
2) reccu ndations
III) Sampling and Analysis Plan Revew Checklist
IV) Laboratory Audit Checklist
3—6
-------
OPERATION AND NAINTEN NCE GUIDANCE FOR RCRA GROUND—WATER
MONITORING SYSTEMS
Ken Jennings, Office of Enforcement, U.S. Environmental Protection
Agency, Washington, D.C.
ABSTRACT
The performance of wells and sampling devices changes over time
resulting in the generation of spurious data. An operation and
maintenance program is therefore necessary to ensure that conditions
are constant, and thus data from one time remain comparable to those
of another. The necessity for operation and maintenance is further
underscored by the fact that initial design of a monitoring network
may change owing to human or natural factors. The continued
reliability of data from monitoring wells is an important facet of
generating ground—water quality information, and one about which
laboratories should be made aware.
The Office of Waste Programs Enforcement is currently involved in the
development of guidance to assist the Regions and States in the
performance of new types of inspections, one of which is the operation
and maintenance inspection. What follows is an outline of the key
areas the workgroup will use as a starting point for writing the
document. A final document is scheduled for release in late 1987.
OUTLINE
1.0 Introduction
1.1 Preface — audience, closure, operating permits
Fundamental: Is company in compliance with regs. and permit?
1.2 Importance of O&M Program
1.3 Relationship with Other Inspections
— timefraiues, WE, linkages, triggers
1.4 O&N Inspection Scope — long term, high use well systems.
Emphasis on monitoring wells, piezometers
1.5 Overview of Inspection: (consistent with body)
— flowchart (products)
2.0 Preparation of the Inspection
2.1 Objectives — to identify data deficiencies, problems
2.2 Review of Data, Documents, and Plans, Identify Data Gaps
2.3 Interviews with Involved Parties
2.4 Inspection Goals
3—7
-------
2.5 Inspection Strategy — develop project plan: outline
equip nt selection, health and safety plan, notification,
sampling plan (split sample)
3.0 Conducting the Inspection
3.1 Discuss Entry — identification
3.2 Objective: fill data gaps, (within O&N context only)
3.3 Project Plan Outline
3.4 Inspections Methods
4.0 Reporting Results
4.1 Report Format
4.2 Checklist
4.3 - nditinns
3—8
-------
UNCONVENTIONAL TECHNIQUES AND UNCOMMON ANALYSES
IN HAZARDOUS WASTE ANALYSIS
Douglas Kendall, National Enforcement Investigations Center,
Environmental Protection Agency, Denver, Colorado
ABSTRACT
One of the more productive approaches to hazardous waste analysis in
our laboratory has been to attempt to identify all major and minor
components in a waste sample, an approach called compositional
analysis. This approach has also been applied to the polluted
groundwater from several Superfund sites, where an attempt was made
to identify all the contributors to the total organic carbon. One
example which will be presented is the identification of
parachlorobenzene sulfonic acid in groundwater. Examples of the use
of infrared spectroscopy and X—ray fluorescence spectroscopy in
hazardous waste analysis will be presented.
INTRODUCTION
The National Enforcement Investigations Center of the U.S.E.P.A.
provides technical support for the enforcement activities of the
Agency. Among the chemical analyses performed for this purpose are,
of course, many of the routine tests described in SW-846. However,
for many of our - projects it is necessary to provide additional
information for such purposes as assessing hazards, tracing wastes
to their source, aiding in cost recovery efforts, and helping with
remedial efforts. It has often been necessary and beneficial to
employ techniques, such as infrared spectroscopy and X—ray
diffraction and fluorescence, which are not commonly encountered in
environmental analysis. Since it has been fruitful and necessary to
completely analyze samples, analytes which are not listed wastes
have been encountered. This paper will describe of some of the less
standard methods and analytes which have been encountered.
X—RAY FLUORESCENCE
X—ray fluorescence instrumentation has been a valuable addition to
the metals capabilities of our laboratory. Energy dispersive XRF
has proven very valuable for sample screening by providing a
semi—quantitative elemental analysis. Simply by placing a powder,
liquid, tar, or sludge in a small plastic cup with a mylar window
and performing a 5 minute analysis, it is possible to determine
qualitatively the composition of a waste sample. This allows
choosing subsequent analyses in a selective manner and greatly
increases efficiency. For instance, solid samples can be eliminated
from consideration for EP toxicity tests if the total amount of each
of the EP metals is less than the extraction limits (corrected for
dilution).
3—9
-------
Wavelength dispersive X—ray fluorescence has been used to make
quantitative measurements on samples, such as waste oils and
contaminated soils, which presented preparation difficulties for
solution measurements.
As a specific example, a comparison study between XRF and ICP will
be described. The samples which were studied were soil samples
which had been heavily contaminated with heavy metals form mining
and smelting activity. After grinding, the soils were pressed into
pellets, with a boric acid binder, for analysis by XRF. For ICP
analysis the samples were fused with RON. This fusion has proven to
be very useful in that siliceous minerals and organic matter are
rendered soluble, but the temperature is low enough that lead,
cadmium, selenium and other volatile metals (except for mercury) are
retained.
Twelve metals were determined by both ICP and XRF, with
concentrations in the range from 20 mg/kg to one percent. The
results form the two methods compared quite well, with an average
relative standard deviation of 20 percent. In our experience, the
sampling error for soil sampling is at least 30% larger than the
analytical error. This study showed that X—ray fluorescence Is
suitable for the rapid analysis for a large number of soil samples
contaminated with metals. Calibration is more difficult with XRF,
but sample preparation is fa8ter, so for large numbers of similar
samples XRF is superior.
INFRARED ANALYSIS
Infrared spectroscopy is one of the most widely applicable
techniques for chemical analysis and compound identification. Among
the many advantages of infrared analysis are the often minimal
sample preparation and the ability to qualitatively identify a wide
variety of both organic and inorganic species. The application in
this laboratory of infrared spectroscopy (IRS) to the analysis of
several hundred hazardous waste samples collected from dump sites
has greatly increased the quality of the analytical results and the
number of components identified.
The applications of IRS to hazardous wastes will be divided Into two
groups. First is the screening of samples; second is the
identification of components best determined by Infrared.
Screening samples by IRS is an integral and key step in the analysis
of hazardous wastes by our laboratory. Rather than look only for a
selected set of compounds, the approach is to determine the major
components of a waste present at the one percent level or higher.
In order to do this, samples were first analyzed by rapid
techniques, including IR, with wide applicability. Subsequent
3—10
-------
analyses are then directed towards those techniques which will
supply the additional information sought.
Usually, the screening and subsequent analyses are done on the
separated phases of the sample, after which the results are combi’ned
to give a view of the entire sample. After phase separation, each
phase is scanned in the as received condition and after drying in an
oven. The spectrum of the volatile portion can be obtained by
subtraction.
The screening spectra from each of the separated phases are then
interpreted. Although some spectra are complex, a surprising number
of spectra are readily interpretable. Often there are only one, two
or three major components in a given phase and the experienced
analyst can readily sort them out. Even if specific compounds can
not be identified, the identification of functional groups or
compound type, such as phenolic compounds, can be very valuable.
Infrared spectroscopy can be used to complement and extend the
commonly used methods of hazardous waste analyses. In several years
experience at NEIC much less than half (probably closer to ten
percent) the organic content of hazardous waste samples can be gas
chromatographed and thus identified by GC/MS. While often the
non—chromatographic components are not the most hazardous, it is
important to identify all major components if hazard is to be
assessed and if futile efforts to identify additional compounds by
GC/MS are to be avoided. The application of IRS to the nonvolatile
organic portion of hazardous waste is the best way to identify such
components as polymers, glycols, etc.
Infrared spectroscopy also can aid with inorganic analyses,
especially with compound speciation. IRS can easily identify some
species such as carbonate and insoluble cyanides often missed by
other methods. Since almost all species with covalent bonds absorb
in the mid—infrared region, IRS has much to offer for inorganic
analysis.
FACTOR ANALYSIS
Environmental analyses often produce much data, and obtaining the
maximum amount of information from this abundance of data is not a
trivial task. Factor analysis is a very useful approach for
reducing the dimensionality of the data so that it is much easier to
display graphically or otherwise interpret. Factor analysis employs
an objective mathematical method to examine the correlations between
variables. A reduced number of new variables are constructed which
contain most of the information of the original variables.
As a simple example, consider some of the data from the analysis of
contaminated groundwater near a hazardous waste dump. Over 30
3—11
-------
parameters were measured on water from 25 wells. To plot each of 30
parameters versus all of the others using two—dimensional plots
would require 375 plots ( 30 x 25 / 2). As an illustration of the
power of factor analysis to reduce the dimensionality of this data
set, the data from the determination of ten metals was treated by
factor analysis. Only three variables were needed to express 96Z of
the total variance from these ten metals. Cadmium, magnesium,
manganese, nickel and potassium were all highly correlated, and
could be combined into one new variable which expressed most of the
changes in these five metals. A second new variable could be
constructed from aluminum, chromium, copper and iron which accounted
f or most of the variation In these metals. Calcium was not
correlated with any other variables and remained as an independent
variable. Thus almost all of the data for ten elements was reduced
to three dimensions, and could be viewed in three plots.
Incidentally, the reason that calcium was not correlated with any
other variable was that large amounts of sulfate were present at
this site, and the sulfate controfled the solubility of calcium.
PAA—CHLOROBENZEN E SULPONIC ACID
A detailed example will now be presented which illustrates the
identification and quantification of a unique pollutant. A waste
site in California received a great deal of hazardous material,
particularly sulfuric acid. The groundwater leaving the site
followed a well defined path down a canyon, toward a small town.
Our laboratory undertook a comprehensive study of the polluted
groundwater as extracted from a number of on and off site wells. It
soon became clear that the amount of organic carbon In the
groundwater far exceeded the total carbon accounted for in the CC/MS
analyses. The gas chromatographable compounds determined in the
purge and trap, semivolatiles and pesticide analyses accounted for
less than ten percent of the total organic carbon.
Besides scientific interest, there were at least three other reasons
for wanting to identify the major organic components in the
groundwater. First, identification of specific compounds would
allow assessment of the hazardous potential of the groundwater.
Secondly, identification of specific compounds might tie a
particular waste generator to the waste site, and perhaps aid In
cost recovery efforts. Third, remedial efforts at the site Include
organic carbon removal, and compound identification could help with
treatment selection and cost estimation.
Among the many parameters determined in the groundwater were the
standard ones of chloride and sulfate. The latter was very high, of
course, since sulfuric acid had been dumped. In addition, X—ray
fluorescence was used to measure total sulfur and total chlorine in
the groundwater samples. Comparison of these results showed that
significant amounts of chlorine and sulfur were present in addition
3—12
-------
to the amounts present as chloride and sulfate. It was possible
that organic species containing chlorine or sulfur were present.
Several approaches were tried to isolate nonvolatile organic
compounds, but only the most successful approach will be described.
A portion of a highly contaminated groundwater sample which was
drawn from a well was taken to dryness In a vacuum oven. The dried
residue was then extracted with boiling ethanol in order to separate
polar organics and ionic organics from inorganic salts. This type
of extract has been used to isolate surfactants from complex
samples. After filtering and drying, the residue was dissolved in
ethanol, water and concentrated hydrochloric acid. Methylene
chloride was used to extract organic acids. An infrared spectrum of
the extract showed that an aromatic sulfonic acid was present which
contained very little if any aliphatic character. Comparison of
this spectrum with that of an authentic sample of para—chlorobenzene
sulfonic acid (PCBSA) showed that this was Indeed the sulfonic acid
that was present.
The presence of the PCBSA has been confirmed in many ways. Two of
the most definitive ways will be mentioned. Diazomethane was used
to make the methyl ester of the sulfonic acid. GC/MS was used to
detect and identify the methyl ester. Another derivative was also
prepared and then confirmed by GC/MS. This was the pentafluoro—
benzyl derivative, prepared from pentafluorobenzyl bromide. This
derivative has been used to detect organic acids and phenols in
water (1).
Quantification was important in order to see if the para—
chlorobenzene sulfonic acid was a major organic component in the
groundwater. Ion—paired reverse—phase liquid chromatography was
found to give good separations and allow control of the elution by
varying the amount of methanol in the methanol — water eluent. The
ion—pairing reagent was tetrapropylammonium ion. The liquid
chromatography gave accurate quantification, as shown by spike
recoveries, and good precision.
The para—chlorobenzene sulfonic acid proved to be the highest
concentration organic compound in the groundwater at this site, with
some wells having over 50% of their organic carbon as this
component. One well had 0.4% PCBSA.
The situation just described was very favorable for identifying the
PCBSA, since it was present at such high concentrations. The method
used to identify this compound would not be applicable to all
situations and required some skill to execute. However, it is clear
that many of the organic compounds at this and similar sites can be
separated and eluted by liquid chromatography. All that is needed
is a means to identify peaks. Mass spectrometry detection is making
great progress in this regard. Also, work Is being done to
3—13
-------
interface infrared spectroscopy with liquid and thin layer
chromatographies, and shows proolse of eventually being applicable
to problems such as those just described.
COMPOS ITIONAL ANALYSIS
Our overall approach to chemical analyses for enforcement purposes
is best described as compositional analysis. Even when the standard
tests for hazardous characteristics are done, it is often desirable
and necessary to obtain further information in order to assist with
hazard assessment, Identify waste sources, etc. In our laboratory,
compositional analysis refers to the determination of all of the
components present in a sample at the one percent level or greater.
For most samples this provides all the information that is needed.
The first step In compositional analysis Is phase separation
followed by a number of widely applicable tests which serve to
screen the samples; eliminate those that require no further work and
direct any further analytical work towards specific goals. For
Instance, a qualitative XRF analysis can determine if solid or heavy
oil samples need EP tests, and can give a good overall indication of
what metals and other elements are present in the sample. In many
cases, infrared spectra give a good indication of the overall
composition of the sample. A Karl Fischer titration Is done to
determine the water concentration. The initial screening provides
sufficient information to characterize many samples, at least in
broad terms and often in some detail. Further work, such as purge
and trap GC/MS or EP Toxicity tests can be selected for those
samples which require them and which have a high probability of
yielding violations or additional information from such tests.
An Important part of our approach to waste analysis is to report the
Information from the analytical tests in a form which can be
understood by engineers, investigators, or nontechnical colleagues.
Rather than report long lists of compounds which were not detected,
we produce a summary of our results. The significant findings for
each sample are presented clearly, along with a physical description
of the sample. Only those determinations which make a waste sample
hazardous or which define Its overall composition are presented.
For example, rather than list a large number of hydrocarbons from .a
GC/MS analysis, we report saturated hydrocarbons or oil. Along with
this summary table, a series of detailed tables present the complete
analytical results.
REFERENCES
Kawahara, F. K., Anal. Chem. 40 (1968) 1009.
3—14
-------
LEACHING AND
PHYSICAL
METHODS
thai rperson
Gail Hansen
themist
Office of Solid Waste
U.S. EPA
401 N Street, S.W.
Washington, D.C. 20460
-------
MODIFICATION TO THE TCLP PROCEDURE FOR PROBLEM MATRICES
Paul J. Marsden, Staff Scientist, S-CUBED, P.O. Box 1620, La Jolla, California
92038; Llewellyn R. Williams, Deputy Director, Quality Assurance and Methods
Division, Environmental Monitoring and Systems Laboratory, U.S.
Environmental Protection Agency, 944 E. Harmon, Las Vegas, Nevada 89114;
Gail Hansen, Chemist, Office of Solid Waste, U.S. Environmental Protection
Agency, 401 M Street SW, Washington, DC 20460
ABSTRACT
The Toxicity Characteristic Leaching Procedure (TCLP) was developed as a
technique to measure the potential pollutants to leach from unsecured
hazardous waste sites. The method was validated in an interlaboratory study
last year. As TCLP has been applied to an increasing number of environmental
samples, certain difficult matrices have been observed.
Possible modifications to the TCLP are being tested for three specific matrices,
oily wastes, wastes which require size reduction, and stabilized wastes. Oily or
nonaqueous liquid wastes cause problems in the TCLP because the matrix can
clog the glass fiber filters thus reducing the amount of contaminants leached
from a sample by the buffer solution. Samples with solids pose two potential
difficulties because (1) solids can contain pollutants that will not come in
contact with the buffer, or (2) new surfaces exposed during pulverization steps
can change the buffering capacity of the sample. Samples that have been
stabilized require physical testing to establish that the contaminants are
actually immobilized in a stable form. If the waste is truly stabilized, then it
should be tested without size reduction. If the waste is poorly stabilized and
can be expected to break up under the effects of weathering (freeze/thaw,
wetting/drying, or compression), then the waste must be pulverized prior to
TCLP.
Analytical data from samples representing each problem matrix will be
presented to demonstrate the effect of proposed method modifications on the
precision and accuracy of TCLP.
4—1
-------
FURTHER DEVELOPMENT OF THE LIQUID RELEASE TEST
Paula A. Hoffman, R. S. Truesdale, P. F. Overby, Research Triangle
Institute, Research Triangle Park, NC and M. B. Meyers, Office of
Solid Waste, Environmental Protection Agency, Washington, D. C.
ABSTRACT
In response to requirements under the 1984 Congressional Amendments to
RCRA, the Office of Solid Waste, Methods Section, developed the Liquid
Release Test (LRT). The test method is intended to determine whether
or not liquids can be released from liquid—loaded sorbent materials
under simulated landfill pressures. On December 24, 1986 the EPA
proposed rules specifying the LRT for use in evaluating whether or not
sorbents employed to remove free liquids would release such liquids
under land disposal environment conditions.
Previous LRT development work led to the selection of the
Zero—Headspace Extractor (ZHE), originally designed for the Toxicity
Characteristic Leaching Procedure, for the LRT. This device uses gas
pressure to force a piston against the sample, in effect squeezing any
releasable liquid from the material.
A single laboratory evaluation of the test’s ruggedness and precision
demonstrated that the test is adequately precise and rugged. An
initial collaborative study indicated a need for several modifications
to the procedure. These modifications included changing the method
for piston—position set up and clarification of the written procedure
to prevent misinterpretation of the procedural steps.
This paper discusses the development of the LRT and further work to be
completed on this test in 1987. This additional work includes
conducting a second collaborative study of the improved procedure,
testing the use of colored filter paper for improved visual detection
of a liquid release, investigating shortening the test time by using
increased pressures, and summarizing and responding to public comments
submitted in response to the EPA’S proposed rules banning free liquids
in containers from hazardous waste landfills.
4-7
-------
VALIDATION OF TOXICITY CHARACTERISTIC LEACHING PROCEDURE
(TcLP) AND APPLICATION TO INDUSTRIAL WASTES
Patty L. Ragsdale, Director of Technical Affairs and Robert E.
Meierer, Compuchem Laboratories, Inc., Research Triangle Park, North
Carolina
ABS TRACT
This paper describes one laboratory’s analytical method validation
process as it was applied to Toxicity characteristic Leaching
Procedure (TCLP) leachate samples. This process involves a
determination of the analyzability of the individual analytes in
simple standards and the extractability of the analytes from the
TCLP leachate. Calculation of method detection limits and report
limits are also part of the validation process described in this
paper.
Data is presented which demonstrates how detection limits and report
limits were established by analyzing replicate spiked samples. The
process for determining appropriate spiking levels for the replicate
samples is also described.
Following completion of the validation process, the tested methods
were applied to leachates prepared from actual industrial waste
samples. The data obtained from the industrial waste leachates is
presented with a discussion of the results. New problems
encountered in the analyses of real samples are discussed with
recommendations for resolution of those problems in a production
environment.
INTRODUCTION
The EPA has been charged with devising the means to be able to
assess certain characteristics or properties of wastes in order to
classify the waste as hazardous. A waste is deemed hazardous when,
if improperly managed, harm to human health or the environment would
occur. Hazardous characteristics, once exhibited by a waste would
thereby require that waste to be managed in a specified way.
The identification of wastes with hazards resulting from the
potential to leach specific toxicants is defined as the Extraction
Procedure Toxicity characteristic. The leaching test, called the
EP, results in an extract analyzed for normal drinking water
parameters; eight metals, four pesticides, and two herbicides. The
regulatory thresholds for those constituents were set at 100 times
the drinking water standard levels.
One of the shortcomings of the EP Toxicity test was its lack of
optimization for organic compounds.
4—9
-------
EPA proposed in the Federal Register of June 13, 1986, to amend the
EP Toxicity Qiaracteristic in order to more properly assess the
bility of both organic and inorganic compounds. The proposed
leaching procedure was called the Toxicity Characteristic Leaching
Procedure or TCLP. The proposal also expanded the list of
constituents to be assessed in the leachate. Table 1 presents the
expanded list together with the regulatory levels for the
contaminants.
Table 1 — Toxicity C2iaracteristic Contaminants and Regulatory levels
Contaminant Regulatory Level (mg/i)
Acrylonitrlle 5.0
Arsenic 5.0
Barium 100.0
Benzene 0.07
Bis (2—chioroethyl) ether 0.05
cadmium 1 • 0
Carbon Disulfide 14.4
Carbon Tetrachioride 0.07
Qilordane 0.03
thlorobenzene 1.4
Qilorofora 0.07
Chromium 5.0
o— reso1 10.0
a—creso l 10.0
p—creso l 10.0
2,4—D 1.4
1, 2—Dichlorobenzene 4.3
1,4—Dicb lorobenzene 10.8
1, 2—Dichloroethane 0.40
1 ,l—Dichloroethylene 0.1
2,4—Dinitrotoluene 0.13
Faidrin 0. 0003
Reptachior (and its hydroxide) 0.001
Ilexach lorobenzene 0.13
Ilezachiorobutadiene 0.72
xach1oroethane 4.3
Isobutanol 36.0
Lead 5.0
Lindane 0.06
Mercury 0.2
Methoxychior 1.4
Methylene Chloride 8.6
Methyl Ethyl Xetone 7.2
Nitrobeazene 0.13
Pentachiorophenol 3.6
Phenol 14.4
Pyridine 5.0
4—10
-------
Table 1
Toxicity Characteristic Contaminants and Regulatory Levels
(continued)
Contaminant Regulatory Level (mg/i)
Selenium 1.0
Silver 5.0
1,1,1, 2—Tetrachioroethane 10.0
1,1,2, 2—Tetrachioroethane 1.3
Tetrach loroethylene 0.1
2,3,4,6—Tetrachiorophenol 1.5
Toluene 14.4
Toxaphene 0.07
1,1,1—Trlchioroethane 30.0
1,1, 2—Trichi.oroethane 1.2
Trichioroethylene 0.07
2,4, 5—Trichiorophenol 5.8
2,4, 6—Trichiorophenol 0.30
2,4,5—TP (Silvex) 0.14
Vinyl chloride 0.05
This paper presents our laboratory’s approach to the validation of
this leaching procedure and the associated analytical methods.
Our laboratory’s validation efforts were initiated through
participation in two EPA—sponsored studies. The first was a
collaborative study on the TCLP conducted by S—Cubed. Subsequently,
we were involved in another study program, funded by EPA under the
direction of the Dynamac Corporation. One phase of the latter study
being reported here was to determine if selected SW—846 methods
could be used as written, to detect the various hazardous chemicals
at or below the regulatory levels in a synthetic, laboratory—
generated leachate.
EXPERIMENTAL DESIGN FOR SYNTHETIC LEACHATE TESTING
Preparation of Synthetic Leachate
The synthetic leachate consisted of the TCLP Extraction Fluid #1.
Each liter of leachate was prepared by adding 5.7 ml glacial acetic
to 500 ml of laboratory reagent water followed by the addition of
64.3 ml of 1.0 N sodium hydroxide and dilution to 1.0 liter. The pH
of the solution was verified to be within the required range of 4.93
+1— 0.05.
4—11
-------
Table 2 - Concentrations of TCLP Synthetic Leachate Spiking Levels
Compared to Regulatory Levels
Regulatory Spiking
Level (mg/i) Concentration (mg/i)
Volatile Compounds - Analyzed Direct Injection
Acrylonitrile 5.0 2.0
Carbon Disulfide 14.4 2.0
Isobutanol 36 4.0
Volatile Compounds - Analyzed Purge & Trap
Benzene 0.07 0.05
Carbon Tetrachioride 0.07 0.05
Chlorobenzene 1.4 0.05
Chloroform 0.07 0.05
1,2-Dichioroethane 0.40 0.05
1,1—Dichloroethylene 0.1 0.05
Plethylene Chloride 8.6 0.05
Methyl Ethyl Ketone 7.2 0.05
1,1,1 ,2-tetrachloroethane 10.0 0.05
1,1 ,2,2—Tetrachloroethane 1.3 0.05
Tetrachloroethylene 0.1 0.05
Toluene 14.4 0.05
1,1,1—Trichioroethane 30 0.05
1,1,2—Trichloroethane 1.2 0.05
Trichioroethylene 0.07 0.05
Vinyl Chloride 0.05 0.05
4—12
-------
Table 2 - Concentrations of TCLP Synthetic Leachate Spiking Levels
Compared to Regulatory Levels (Continued)
Regulatory Spiking
Level (mg/i) Concentration (mg/i)
Sernivoiatiie Compounds
o-Cresol 10.0 0.200
m-Cresoi 10.0 0.200
p-Cresol 10.0 0.200
1,2-Oichlorobenzene 4.3 0.100
1,4-Dichlorobenzene 10.8 0.100
Bis(2-chloroethyl) Ether 0.05 0.05
Hexachlorobenzene 0.13 0.100
Hexachlorobutadiene 0.72 0.100
Hexachioroethane 4.3 0.100
2,4-Dinitrotoluene 0.13 0.100
Nitrobenzene 0.13 0.100
Phenol 14.4 0.200
Pentachiorophenol 3.6 0.200
Pyridine 5.0 0.100
2,3,4,6—Tetrachiorophenol 1.5 0.200
2,4,5—Trichiorophenol 5.8 0.125
2,4,6—Trichiorophenol 0.3 0.125
Pesticide Compounds
Chiordane 0.03 0.002
Endrjn 0.0003 0.0005
Heptachior 0.001 0.0003
Ljndane 0,06 0.0002
Iflethoxychior 1 .4 0.001
Toxaphene 0.07 0.005
4—13
-------
Preparation of Spiking Standards
The spiking levels selected for each analyte and the corresponding
regulatory level are presented in Table 2. A number of the spiking
concentrations are significantly lower than the regulatory limits.
This is because those regulatory levels were outside the working
analytical range of the method. In those cases, the spiking level
was set to be within the upper half of the analytical range.
The base—neutral extractable compounds were nominally spiked at
0.100 mg/i with bis(2—chloroethyl)ether being spiked at 0.050 mg/i
which is the regulatory level. The acid extractable compounds were
nominally spiked at 0.200 mg/i with the trichlorophenols being
spiked at 0.125 mg/i.
The trichiorophenols were expected to coelute and the spiking levels
for these compounds were set such that the combined concentration
would be equal to the regulatory level set for the 2,4,6—isomer.
Throughout the TCLP validation phases, the trichlorophenols did not
coelute, but m—cresol and p—cresol did.
All analyte, surrogate and internal standard spiking solutions were
prepared within our Standards Laboratory and, with the exception of
a few volatile compounds, neat materials were used.
Teats Included
For each method evaluated the tests depicted in Table 3 were
performed.
QC Performed
As presented in Table 3, surrogates were spiked into all test
samples including blanks. The surrogates and concentration levels
used in EPA’s Contract Laboratory Program (CLP) were employed for
the study. Table 4 presents the surrogates employed, the
concentration of the surrogates in the TCLP extract and, for
advisory purposes, the CLP surrogate acceptance criteria.
A synthetic leachate blank containing surrogates was processed to
evaluate the leachate background. Method blanks containing only
surrogates in laboratory reagent water were extracted and analyzed
to demonstrate that the processing and processing environment did
not introduce contamination.
A blank spike, containing surrogates and analytes of interest was
processed to verify that the entire analytical process was in
control.
4—14
-------
Table 3 - Tests Performed for Each F 1ethod Evaluated
- 1 TCLP Extract method Blank
Prepared from Synthetic Leachate Fluid
(Synthetic Leachate Fluid + Surrogates)
- 3 TCLP Extract Spikes
Prepared from Synthetic Leachate Fluid
(Synthetic Leachate Fluid + Surrogates + !\nalytes)
- 1 Blank Water Spike
Prepared from Laboratory Reagent Water
(Reagent Water + Surrogates + I-’tnalytes)
- 2 method Blanks
Prepared from Laboratory Reagent Water
(Reagent Water + Surrogates)
4—15
-------
Table 4 - Surrogate Spike Concentrations and CLP Acceptance Criteria
Surrogate Compound
Concentration
(ugh)
Percent Recovery
Acceptance Range
d8-Toluene
4-Bromo?luorobenzene
d4-1 ,2—Dichloroethane
d5—Nitrobenzene
2-Fluorobiphenyl
dl 4-Terphenyl
d5-Phenol
2-Fluorophenol
2,4 ,6-Tribromophenol
Dibutyichiorendate
50
50
50
50
50
50
1 00
100
100
88 - 110
86 - 115
76 - 114
35 - 114
43 - 116
33 — 141
10 - 94
21 — 100
10 — 123
24 - 154
0.1
4—16
-------
Triplicate spikes of the synthetic leachates, containing surrogates
and compounds of interest were processed to obtain a measure of
precision and accuracy.
Analytical Method Employed
The primary source for the analytical methods used in the study was
the 2nd Edition of SW-846. Methods 8080, 8240 and 8270 were
employed as written. Modifications of 8240 and 8150 were also
tested as part of the study, but were not considered part of
Compuchem’s validation process due to their lack of official status.
Instrumentation
8240 Analyses were conducted using a Finnigan OWA equipped with a
Tekmar purge and trap device and 6’ x 1/4” 1% SP—l000 on 60/80 mesh
Carbopack B column.
8270 analyses were conducted using a Finnigan OWA equipped with a 30
meter wide bore, thick film DB—5 fused silica capillary column.
8080 analyses were conducted using two Varian gas chromatographa
equipped with electron capture detectors. Additionally, one
instrument was equipped with a 2mm x 6’ column packed with 3% OV—lOl
on 100/120 mesh Supelcoport and the other instrument was equipped
with a 4mm x 6’ column packed with 1.95% SP—225011.5% SP—2401 on
100/120 mesh Supelcoport.
Instrument QC
Prior to analysis of volatile samples by Method 8240, an acceptable
BFB tune and a multipoint calibration were obtained. The BFB
acceptance criteria presented in CL? methodologies were applied.
The nominal concentrations of the three standards which comprised
the multipoint calibration were 40, 80 and 160 ug/l.
Prior to analysis of semivolatile samples by Method 8270, an
acceptable DFTPP tune and a multipoint calibration were obtained.
DFTPP acceptance criteria presented in CLP methodologies were
applied. The nominal concentrations of the acid extractable
compounds in the five standards comprising the multipoint
calibration were 50, 100, 150, 200 and 250 ugh. The nominal
concentrations of the base—neutral extractable compounds in the
multipoint standards were 20, 50, 80, 120 and 160 ugh.
The %RSD for seven compounds (methylene chloride, 2—butanone,
trichloroethylene, benzene, 1,1, 2—trichioroethane, 1,1,2, 2—tetra—
chloroethane, and tetrachioroethylene) in the volatile multipoint
calibration were greater than the 10% required by the method for use
of the average response factor in subsequent sample quantitations.
4—17
-------
For these compounds, quantitations were based on linear fit. All
other compounds were quantitated Using the average response factor
from the multipoint calibration.
The %RSD for each analyte in the semivolatile multipoint
calibrations was less than 25% as required by the method for use of
the average response factor in subsequent sample quantitations.
z! ! !i Results
AnRlyte and surrogate recoveries for each spiked synthetic leachate
aemple and the blank spike processed for semivolatile, purgeable
volatile, direct Injectable volatile and pesticide analytes are
presented in Tables 5 — 8 respectively.
As illustrated, all analytes were recovered well from the synthetic
leachate. Semivolatile recoveries ranged from 50% for phenol to
113% for pentachiorophenol. Purgeable volatile recoveries ranged
from 72% for 2—butanone to 107% for l,2—dlchi.oroethane. Pesticide
recoveries ranged from 81% for heptachior to 106% for toxaphene.
The direct injected volatiles exhibited the greatest variation in
recovery with acrylonitrile at 48% and isobutanol at 122%. The
standard deviation for all compounds except isobutanol and
2—butanone were less than 10%. The standard deviations of those
compounds were 21.2 and 12.3, respectively.
Onj METHOD VALIDATION STUDIES
The Federal Register of June 13, 1986 which presented the proposed
rule on TCLP Indicated that the quantitation limits for the
constituents of Interest were based on a water matrix. Further, It
states “Since TCLP extracts would also be aqueous In nature, EPA is
proposing to use the quantitation limit as observed In water.”
In our laboratory we have a formalized method validation program
based on various EPA guidance documents. Our typical method
validation program involves generating precision and accuracy data
on quadruplicate replicates processed through the entire analytical
process. The spiking level Is typically in the middle of the
analytical working range. Then, using information contained in
Appendix B of October 26, 1984, Federal Register a Method Detection
Limit (MDL) study is performed. For this requirement, a minimum of
seven (7) replicates are spiked with each analyte being validated at
a level of three to five times the estimated detection limit. Each
replicate is processed through the entire analytical process. Mean
recoveries are determined as well as the standard deviation of the
replicate analyses. The MDL is then derived by applying a Student’s
“T” value to the standard deviation.
4—18
-------
This procedure is performed using laboratory reagent water for
aqueous matrix procedures and Ottawa sand or verified blank dirt for
procedures designed for solid matrices.
The final aspect of our method validation program employs guidance
contained in an Analytical Chemistry article entitled “Principles of
Environmental Analysis.” The article authored by the ACS Committee
on Environmental Improvement, included a discussion on the MDL as
well as the Limit of Quantitation (LOQ). The latter is the level
above which quantitative results may be obtained with a specified
degree of confidence. Whereas, the MDL determination for the
analysis of a minimum of seven replicates utilizes a Student’s “T”
value in the 2.5 to 3.1 range times the standard deviation, the LOQ
determination recommends a multiplier of ten (10) to eliminate any
degree of uncertainty. If analytes are detected below this level
and, when employing CC/MS, a resulting mass spectrum meets
identification criteria, the quantitative value is characterized as
an estimate and is flagged as such. This is the policy we have
attempted to follow for all of the SW—846 methods we have validated.
As part of the validation program for SW—846, Third Edition, we
compared precision and accuracy data of the analytes involved in the
TCLP synthetic leachate study with those same compounds in
laboratory reagent water. A comparison of those data is presented
In Table 9. The data obtained during our validation of SW—846,
Second Edition is also included.
APPLICATION TO INDUSTRIAL WASTES
During the past year we have had occasion to perform compositional
analyses of some Industrial wastes as well as TCLP and ZHE (Zero
Head Space) extractions followed by semivolatile and volatile
analyses, respectively. Examinations of the data obtained for both
analyses shows that compounds which were present iii the raw sample
at concentrations greater than 1000 ug/kg were present in the TCLP
and ZHE leachates at measurable concentrations. Analytes present in
the industrial wastes included benzene, toluene, chlorobenzene,
chloroform, 1, 2—disclorobenzene, tetrachioroethylene, o—cresol,
p—cresol, hexachlorobenzene and 2,4, 6—trichiorophenol.
The major problem encountered in the analysis of real industrial
waste samples was buffering during the final pH adjustment of
several samples during the semivolatile extraction. For three
samples, the pH had Increased to greater than 2 by the end of the
first acid extraction. These samples required readjustment of the
pH at that point and three more extractions with methylene
chloride. By verifying the the pH at the end of the first base and
acid extractions, viable extracts in which all surrogate recoveries
were within acceptance ranges were obtained for each leached sample.
4 -19
-------
Table 5 — Synthetic Leachate Spike Percent Recovery Data for Method 8270
Synthetic Leachate
Concentration Blar c Synthetic Leachate Splices ——— Spike
of Spike Added Water Spike 11 12 13 Average S.D.
(uqJl)
Analytical Ccm ouri1s:
1 ,2—Dichlorobenzene 100 84 87 89 75 84 6.2
I ,4—Dichlorcbenzene 100 80 84 85 69 79 7.3
2,3,4,6—Tetrachiorophenol 200 102 106 108 106 107 0.9
2,4,6—Trichiorophenol 125 68 68 67 69 68 0.8
2,4,5-Trichiorophenol 125 61 65 63 67 65 1.6
2,4-Olnitrotoluene 100 94 99 97 90 95 3.9
Bis(2—chloroethyl)ether 50 77 84 84 7? 82 3.3
Hexachlorthenzene 100 97 116 109 106 110 4.2
Hexachlorabutadiene 100 86 86 91 70 82 9.0
Hexachloroethane 100 79 75 80 65 73 6.2
m-Cresol 200 76 69 85 78 7? 6.5
Nitrobenzene 100 81 88 92 79 86 5.4
o—Creso]. 200 82 89 81 84 85 3.3
Pentachiorophenol. 200 101 110 113 116 113 2.4
Phenol 21)0 48 4? 50 52 50 2.1
Pyridine 100 70 72 70 62 68 4.3
p-Creeol 200 76 69 85 78 77 6.5
Surrogate Cc ounds:
2,4,6—Tribrcm phanol 200 105 106 113 110 110 2.9
2-Fluoroblphenyl 100 89 98 92 83 91 6.2
2—Fluorophenol 200 61 62 66 60 63 2.5
d I O—Pyrene 100 124 127 117 131 125 5.9
d14—Terphenyl 100 126 134 122 129 128 4.9
d5—Nitrobenzene 100 78 92 84 73 83 7.8
d5—Phenol 200 44 48 50 46 48 1.6
4—20
-------
Tthle 6 — Synthetic Leachate Spike Percent Recovery Data for Purgeable Vo].atiles by Nethod 8240
Synthetic Leachate
Concentratim Blank Synthetic Leachate Spikes ————— Spike
of Spike Added Water Spike #1 #2 #3 Average S.D.
(ugh)
Analytical Compounds:
Vinyl Chloride 50 112 94 88 92 91 2.5
thyler€ Chloride 50 100 94 84 86 88 4.3
1,1-Dichloroethylene 50 116 96 88 86 90 4.3
1,2-Dichioroethane 50 128 116 102 102 10? 6.6
2-Butanone 50 96 88 70 58 72 12.3
1,1,1-Trichioroethane 50 124 106 96 94 99 5.2
Carbon Tetrachioride 50 128 106 96 96 99 4.7
Trichloroethylene 50 116 98 88 84 90 5.9
B izene 50 110 96 86 84 89 5.2
1,1,2—Trichioroethane 50 116 102 86 84 91 8.1
1,1,1,2—Tetrachioroethane 50 124 110 98 96 101 5.2
1,1,2,2—Tetrachioroethane 50 114 98 82 76 85 9.3
Ietrachloroethylene 50 118 98 88 86 91 5.2
Toluene 50 118 104 96 94 98 4.3
Q 1orobenzene 50 122 108 98 96 101 5.2
Irrogates:
d -1,2-Dichloroethane 50 114 114 100 100 105 6.6
Bromofluorobe,izene 50 108 106 96 84 95 9.0
-Ioluene 50 108 108 98 96 101 5.2
4—21
-------
Table 7 - Synthetic Leachate Spike Percent Recovery Data for Direct Inject Volatiles by Method 8240
Synthetic Leathate
Concentratien Blar* Synthetic Leachate Spikes - --- — Spike
of Spike Added Water Spike 11 12 13 Average S.D.
(ug h)
Analytic :
Carbm disulficM 2.0 57 93 104 94 97 5.0
Acrylanitrile 2.0 57 50 59 36 48 9.5
Isthutan 1 4.0 4.3 123 96 148 122 21.2
Table 8 — Synthetic Leachate Spike Percent Recovery Data for Method 8080
Synthetic Leachate
Concentraticm 8lar c Synthetic Leachate Spikes - Spike —
of Spike Added Water Spike 11 12 13 Average S.D.
(ugh)
Pesticide Spike Series 11:
Toxa iene 5.00 112 106 108 104 106 1.6
Dibutyil ii te (surrogate) 1.00 97 93 89 93 92 1.9
Pesticide Spike Series 12:
Endrin 0.50 93 98 96 99 96 1.2
Heptath].or 0.30 81 81 82 81 81 0.5
Lindane 0.20 86 87 86 88 87 0.8
Nethoxychior 1.00 98 99 99 99 99 0.0
Dibutyichiorendate (surrogate) 1.00 64 91 86 86 88 2.4
Pesticide Spike Series 13:
thiordane 1.00 88 73 87 86 82 6.4
Oibutylch].orendate (surrogate) 1.00 88 79 89 89 86 4.7
4—22
-------
Table 9 — Comparison of Results of Three Validations
Synthetic Leachate Study 2nd Edition StiJ—846 3rd Edition S iI—846
Average S.D. Average S.D. Average S.D.
Søi ivolatile Compounds:
Phenol 51] 2.1 64 12.3 40 7.3
Bis(2—chloroethyl)ether 82 3.3 88 4.4 75 17.7
1,4—Dichlorobenzene 79 7.3 111 12.6 78 8.0
1,2—Oichlorobenzene 84 6.2 98 6.9 71 6.7
o—Cresol 85 3.3 98 9.2 77 10.8
p—Cresol 77 6.5 96 14.9 84 10.5
Hexachioroethane 73 6.2 96 13.4 70 7.8
Nitrobenzene 86 5.4 95 8.8 81 8.1
Hexachiorobutadiene 82 9.0 106 1.7 77 8.2
2,4,6—Trichiorophenol 68 0.8 111 2.6 81 17.4
2,4,5—Trichiorophenol 65 1.6 121 2.5 84 16.5
2,4—Dinitrotoluene 95 3.9 101 3.6 85 13.4
I-Iexachlorobenzene 110 4.2 102 1.4 83 10.5
Pyridine 68 4.3 58 15.8 28 9.2
2,3,4,6—Tetrachiorophenol 107 0.9 124 2.2 86 20.5
Purgeable Volatile Compounds:
Vinyl Chloride 91 2.5 76 3.2 89 12.1
Nethylene Chloride 88 4.3 89 1.8 129 16.1
1,1—Dichloroethylene 90 4.3 82 3.9 87 5.6
Chloroform 51 2.4 90 2.5 94 5.6
1,2—Dichloroethane 10? 6.6 81 12.2 90 2.7
2—Butanone 72 12.3 89 1.6 NA NA
1,1,1—Trichioroethane 99 5.2 85 4.6 80 3.4
Carbon Tetrachioride 99 5.2 82 4.9 74 4.1
Trichloroethylene 90 5.9 82 3.1 78 3.1
Benzene 89 5.2 86 3.2 87 3.6
1,1,2—Trichioroethane 91 8.1 84 6.4 88 6.6
1,1,2,2—Tetrachloroethane 85 9.3 78 3.4 81 10.7
Tetrachioroethylene 91 5.2 79 8.5 74 4.0
Toluene 98 4.3 84 3.0 86 2.0
Chlorobenzene 101 5.2 85 2.5 87 2.4
NA = Not Available, Validation Not Complete At This Time
4—23
-------
CONCLUS IONS
The experimental data indicated that the TCLP methods would work for
the analytes of interest. The results of the various validation
studies indicate that extraction fluid #1 is similar to water for
the parameters tested. Subsequent application of those methods to
actual industrial waste sample leachates proved that the methods are
viable for the TCLP. As long as the pH of each sample is checked at
the end of the initial base and acid extraction, the semivolatile
recoveries, as indicated by surrogate recoveries, should be similar
to those obtained for water samples.
4—24
-------
PERFORMANCE OF THE TOXICITY CHARACTERISTIC LEACHING PROCEDURE
Lynn R. Newcomer, Chief Chemist, Wilson Laboratories, Sauna,
Kansas; W. Burton Blackburn, S—Cubed, LaJolla, California; Gall A.
Hansen, U.S. Environmental Protection Agency, Washington, D.C.
ABSTRACT
The U.S. EPA has recently promulgated the use of a leaching
procedure (TCLP) to be used as a criterion for disposal of hazardous
wastes under the Land Disposal Restrictions Rule. The TCLP has also
been included In the proposed rule for Identification and Listing of
Hazardous Wastes.
Summarized are results of eight different studies designed to
evaluate the TCLP method. Study results indicate that the TCLP
procedure can be applied consistently by a diverse group of
laboratories.
INTRODUCTION
The Resource Conservation and Recovery Act (R.CRA) directs the U.S.
Environmental Protection Agency (EPA) to identify and regulate
wastes which pose a hazard to human health and the environment. In
order to meet this mandate, the EPA listed a number of wastes as
hazardous and identif led four hazardous waste characteristics. One
of these, the Extraction Procedure (EP) Toxicity Characteristic,
addre8ses wastes which exhibit leaching potential into ground water
supplies If Improperly managed.
In 1984, Congress amended RCRA directing the EPA to make changes in
the existing Extraction Procedure to predict more accurately the
leaching potentials of hazardous waste8 and expand its application
to a greater number of toxic constituents. In response, the EPA
developed and proposed the Toxicity Characteristic Leaching
Procedure (TCLP). The second generation leaching procedure, TCLP,
was published as a draft protocol on December 20, 1985; it was
officially published in the Federal Register (Vol. 51, No. 9,
January 14, 1986) as part of the proposed Land Disposal Restrictions
Rule, and again in the Federal Register (Vol. 51, No. 114, June 13,
1986) as part of the proposed rule for Identification and Listing of
Hazardous Wastes. The final land disposal restrictions rule was
published In the Federal Register (Vol. 51, No. 216, November 7,
1986).
The TCLP protocol calls for an 18—hour extraction of a waste sample
with either an. acetic acid or sodium acetate solution followed by
the determination of metals, pesticides, and semi—volatile and
volatile organic compounds in the leachate. Metals, pesticides and
semi—volatile organic compounds are extracted using a bottle or jar
4—25
-------
similar to the EP procedure. For the extraction of volatile organic
compounds (VOCs), a new device knowii as a Zero Headapace Extractor
(Z}IE) is used.
To date, there have been various single and multi—laboratory studies
conducted to evaluate the TCLP procedure. Approximately 23
laboratories representing government, industry, research and
commercial organizations participated in various phases of the
evaluation studies. At least 15 different types of wastes were used
including metal manufacturing, power plant, refinery,
electroplating, textile and Publicly Owned Treatment Works (POTW)
wastes.
It is the intent of this paper to combine and condense data from
these various studies and to discuss the method (TCLP) performance,
i.e., comparison to the EP procedure, ruggedness, precision and
reproducibility. Data presented herein represent a fairly
comprehensive data base from analyses of the various wastes.
In order to condense available data, only results of analytes
detected in greater than 90 percent of the samples are reported.
Outliers, as determined by the Dixon Test (1) at the 95 percent
confidence level, are also excluded. In several of the studies, the
Dixon Test had not been applied to the study results. Where raw
data were available, the Dixon test was applied by the author.
Most of the original Studie8 conducted under EPA contact or by
private industry as referenced in this paper are available for
public review in the EPA R RA Docket (S—212) at the EPA, Washington,
DC.
TCLP VERSUS EP
Two studies comparing the TCLP and EP extractions are sni insrized.
The sponsoring industries selected 13 different wastes for
extraction and analysis. Only results reported by more than one
laboratory and on analytes which were detectable are included here.
Inter—Industry Collaborative Study
The Inter—Industry Collaborative Study (2) was sponsored by six
trade associations with the goal of evaluating the TCLP on wastes of
interest to the organizations. Only one waste, waste #6, is
reported since it was the only waste analyzed by all six
participating laboratories. This waste consisted of a composite of
baghouse dust from a steelniaking operation and sinter waste from a
lead smelting facility. The results of this study are summarized in
Table 1. For the five metals presented, there does not appear to be
a large difference in the amount of extractable analytes. Although
the R.SDEP for lead is significantly greater than the lead RSDTCLP,
4—26
-------
the other four metals indicate similar extraction repeatability.
Since the raw data for this study were not available, it is not
known whether outliers were discarded. The study concluded that
although the TCLP and EP are not precise methods, they are similar
in precision.
Table 1. Inter—Industry Collaborative Study Summary Results of EP
and TCLP Metal Extracts of Waste #6
Parameter
XEP
XTCLP
SEP
STCLP
%RSDEP
%RSDTCLP
Arsenic
Barium
Cadmium
Chromium
Lead
0.048
5.9
0.19
0.090
293
0.039
14.5
0.17
0.069
256
0.024
3.77
0.061
0.065
471
0.021
10.5
0.067
0.048
241
50
64
32
72
161
44
72
40
69
94
Average
76
64
Note: !t ts mg/L
X = Means values for 11—12 extractions
S Standard Deviation
ZRSD Relative Standard Deviation
Electric Power Re8earch Institute
A second study comparing TCLP and EP procedures was sponsored by the
Electric Power Research Institute (EPRI) (3). Seven different
utility wastes (including fly ashes, bottom ashes, flue gas sludges)
were extracted In duplicate by three laboratories. The six extracts
of each waste were divided equally among the three laboratories and
analyzed in quadruplicate. A summary of the results are reported In
Table 2.
During the period of public comment, one of the frequently mentioned
concerns of the TCLP was the possible aggressive leaching potential
of the acetate extraction medium. The average total metal
extractables appear to be somewhat higher for the TCLP extraction
(0.50 mg/L vs 0.39 mg/L or 0.29 mg/L vs 0.17 mg/L) if zinc, waste
W—2, results are not included; the relatively large zinc values tend
to buffer significant differences of the smaller values. Eighty
percent of the results in Table 2 showed a ration of XTCLp:XEP of
0.8—2.0 and 15 percent fell within the 2.0 to 3.0 range
(disregarding the data pairs containing “ND”). Although these
4—27
-------
Table 2. EPRI Results of EP and TCLP Metal Extracts of Seven Utility
Wastes
Metals
XE?
XTCLP
SEP
STCLP
%RSDEP
%RSDTCLP
Arsenic, W—2
W—7
Barium, v—i
W—3
W—7
Cadmium, V—i
W—2
W—5
W—6
W—7
Chromium, W—1
W—2
W—3
W—4
W—5
W—7
Lead, W—2
Selenium, W—7
Zinc, V—i
W—2
W—3
W—5
W—6
W—7
ND
0.051
0.406
0.430
0.177
0.013
0.224
0.033
0.005
0.006
0.427
0.016
0.008
ND
0.030
MD
ND
0.061
0.171
5.36
0.092
1.48
0.174
0.151
0.317
0.149
0.327
0.819
0.446
0.016
0.233
0.028
0.004
0.006
0.470
0.921
0.010
0.004
0.042
0.059
0.181
0.135
0.238
5.37
0.164
1.6
0.306
0.234
0.034
0.178
0.219
0.060
0.004
0.038
0.133
0.002
0.001
0.077
0.023
0.006
0.026
0.024
0.142
0.3
0.129
0.12
0.052
0.073
0.228
0.021
0.147
0.336
0.138
0.004
0.100
0.005
0.002
0.002
0.070
0.405
0.005
0.002
0.036
0.018
0.122
0.036
0.171
0.38
0.174
0.22
0.278
0.140
67
44
51
31
31
17
25
48
18
l8
145
75
85
39
83
6,,
140L
8
30
48
72
14
45
41
31
25
43
17
38
32
15
44
46
55
86
30
67
27
72
7
106
14
91
60
Average
0.3
0.50
0.08
0.13
51
45
Without
Zinc, W—2
Average
0.17 0.29
Note: Units mg/L
Alkaline fly ash
W—2 Acidic fly ash
W—3 Alkaline bottom ash
v-4 Neutral bottom ash
W—5 Forced oxidized flue gas desulfurization sludge
W—6 Flue gas desulfurization sludge
W —7 Neutral fly ash
ND = 50% of the results were not detected
One lab reported results lOX lover than thee other two labs
Analyses of one lab’s extracts were approximately lOX greater
than those of the other two labs
Approximately 40 analyses were performed on each waste type.
4—28
-------
ratios tend to support a conclusion that the TCLP is more
aggressive, all extractable results reported are well below
hazardous waste threshold levels. If the TCLP is more aggressive,
it would appear to be waste and metal specific.
The EPRI study concluded that the TCLP appears to provide better
extraction reproducibility but that inter—laboratory variations are
a significant factor in TCLP analyses.
For the majority of wastes analyzed in these two studies, the TCLP
and EP are similar in extraction efficiency and method precision.
Where differences exist, the TCLP tends to provide better precision,
is easier to perform and produces a more aggressive leaching medium.
RUGGEDNESS
Two ruggedness studies have been performed to determine the effect
of various perturbations on specific elements of the TCLP protocol.
Ruggedness testing determines the sensitivity of small procedural
variations which might be expected to occur during routine
laboratory application. Method variations which are observed to
significantly affect analytical results need to be carefully
controlled.
Both studies followed the partial factorial design described by
Youden. In this design, seven conditions were slightly altered and
eight extractions were performed. This design provided sufficient
Information to identify those areas of the method which were
affected by procedural variations.
Metals
A study by ENSECO(4) investigated the following conditions for
metals results on two wastes:
1) Liquid/Solid ratio — 19:1 vs 21:1
2) Extraction time — 16 bra vs 18 bra
3) Headspace — 20% vs 60%
4) Buffer #2 acidity — 190 meq vs 210 meq
5) Acid—washed filters — yes vs no
6) Filter type — O.7,um glass fiber vs. 0.45gm vs
polycarbonate
7) Bottle type — borosilicate vs flint glass
Of the seven method variations examined, acidity of the extraction
fluid had the greatest impact on the results. Four of 13 metals
from an API separator sludge/electroplating waste (API/FM) mixture
and two of three metals from an ammonia lime still bottom waste were
extracted at higher levels by the more acidic buffer. Because of
the sensitivity to pH changes, the method requires that the
4—29
-------
extraction fluids be prepared so that the final pH is within + 0.05
units as specified. —
Although not directly apparent from this study, other studies
referenced have indicated that flint glass, no acid—washed filters
or other filter materials may also affect results. Therefore, glass
fiber filters (acid washed for metals analyses) are required and
borosilicate glass is recommended.
Volatile Organic Compounds
A separate ruggedness study was performed by ERCO/ENSECO (5) to
investigate method variations on volatile organic compounds (VOCs)
in API/EW and ammonia lime still bottom wastes. The following
parameters were tested:
1) Liquid/Solid ratio — 19:1 vs 21:1
2) Headspace — 0% v 5%
3) Buffer #1 acidity — 60 meq vs 80 meq
4) ZHE device — ADM vs Millipore
5) Method of storing
extract — Syringe vs Tedlar bag
6) Aliquotting — yes vs no
7) Pressure behind piston — 0 psi vs 20 psi
The only parameter having a significant effect on the results was
the choice of the extraction device. The original Zero Headspace
Extractors (ZHEs) manufactured by Millipore had problems with
leaking valves and with piston movement which resulted in loss of
volatile compounds. Milhipore has since corrected the valve and
piston problems so that both the Millipore and ADM ZHEs should
perform similarly. No problems were reported with the retrofitted
Milhipore ZREs used in the S—Cubed Collaborative Study referenced
later in this paper.
The ERCO/ENSECO study reported that the ZHE was “adequately rugged
with respect to the parameters Investigated.”
PRECISION
A substantial amount of data ha8 been generated from TCLP precision
(reproducibility) studies. Both single— and multi—laboratory
studies to assess method precision have been sponsored by private
industry and by the EPA. Precision results from recent studies are
encouraging. The general consensus from these studies is that the
precision of the TCLP is comparable to or exceeds that of the EP
procedure and that method precision Is adequate. One of the more
significant contributions to poor precision appears to be related to
sample homogeneity and inter—laboratory variation which is not
surprising due to the nature of waste materials.
4—30
-------
l4etals
The largest source of TCLP precision data comes from the metals
analyses of TCLP extracts. Twenty—three laboratories performed
extractions and analyses on 15 different wastes. The results from
these analyses are presented in the tables below. Tables 3 and 4
contain single and multi—laboratory results, respectively. Results
included in tables 1 and 2 also provide precision information. The
general range for mean %RSDs in Tables 1, 2, 3 and 4 is 22—74%.
Although not necessarily an indication of a precise method, the
range is not unreasonable considering the waste types and the
relatively low levels of metals determined. Less than five percent
of these wa8tes would be regulated as hazardous wastes under 40 CFR,
Part 261, based on the data presented in these Tables. The results
indicate that a single analysis of a waste may not be adequate for
waste characterization and identification requirements, a comment
reinforced by the Inter—Industry Collaborative Study and in the
Background Document (6).
Semi—Volatile Organic Compounds
Semi—volatile organic compounds have not been studied as extensively
as the metals. There are, however, two studies, a single—laboratory
evaluation (4) and a multi—laboratory collaborative study (7), which
provide data from several different waste types. Results of these
studies are summarized in Tables 5 and 6.
Single—laboratory precision was excellent with greater than 90
percent of the results exhibiting an RSD less than 25 percent. Over
85 percent of all individual compounds in the multi—laboratory study
fell in the RSD range of 20—120 percent. The actual ranges are
included In Table 6.
The single—laboratory evaluation (4) reported somewhat better
reproducibility for semi—volatiles with the more acidic extraction
fluId, #2. However, the multi—laboratory study (7) did not confirm
the same relationship between extraction fluid acidity and precision.
Both studies concluded that the TCLP provides adequate precision.
It was also determined that the high acetate content of the
extraction fluid did not present problems (i.e., column degradation
of the gas chromatograph) for the analytical conditions used.
Volatile Organic Compounds
Both single and multi—laboratory data are available for VOC
precision. Three studies using a combination of three different
waste types are summarized.
A single—laboratory study conducted by S—Cubed (8) involved the use
4—31
-------
Table 3. Single—Laboratory TCLP Metals, Precision (Williams, et al)
Waste
Extraction
Fluid
Metal
X
S
%RSD
Ammonia Lime
Still Bottoms
API/EW
Mixture
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
Barium
Chromium
Zinc
Barium
Chromium
Zinc
0.283
0.332
0.049
0.062
0.178
0.257
0.696
1.01
0.125
14.4
136
347
0.013
0.015
0.023
0.012
0.126
0.071
0.020
0.01
0.062
3.6
7.5
14.2
4.6
4.6
47
19
71
28
2.8
0.6
49
25
5.5
4.1
XRSD Range 0.6—71
Mean ZRSD 22
Note: All extractions were performed in triplicate.
Extraction Fluid #1 pH 4.9
#2 = pH 2.9
Units mg/L
4—32
-------
Table 4. Multi—Laboratory TCLP Metals, Precision (Blackburn and Show)
Waste
Extraction
Fluid
Metal
X
S
%RSD
Ammonia Lime
Still Bottoms
API/EW
Mixture
Fossil Fuel
Fly Ash
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
#1
#2
Cadmium
Chromium
Lead
Cadmium
Chromium
Lead
Cadmium
Chromium
Lead
0.053
0.023
0.015
0.0032
0.0030
0.0032
0.0046
0.0005
0.0561
0.105
0.0031
0.0124
0.080
0.093
0.017
0.070
0.0087
0.0457
0.031
0.017
0.0014
0.0037
0.0027
0.0028
0.0028
0.0004
0.0227
0.018
0.0031
0.0136
0.069
0.067
0.014
0.040
0.0074
0.0083
60
76
93
118
90
87
61
77
40
17
100
110
86
72
85
57
85
18
%RSD Range = 17—118
Mean %RSD = 74
Note: X = Mean results from 6—12 different laboratories
Units = mg/L
Extraction Fluid #1 = pH 4.9
#2 = pH 2.9
4—33
-------
Table 5. single—Laboratory Semi—Volatiles, Precision
Waste
Compound
Extraction
Fluid
X
S
%RSD
Ammonia
Lime Still Phenol #1 19000 2230 11.6
Bottoms #2 19400 929 4.8
2—Methylphenol #1 2000 297 14.9
#2 1860 52.9 2.8
4—Nethyiphenol #1 7940 1380 17.4
#2 7490 200 2.7
2,4—Dimethyiphenol #1 321 46.8 14.6
#2 307 45.8 14.9
Naphthalene #1 3920 413 10.5
#2 3827 176 4.6
2—Methy lnaphthalefle #1 290 44.8 15.5
#2 273 19.3 7.1
Dibenzofurafl #1 187 22.7 12.1
#2 187 7.2 3.9
Acenaphthy lefle #1 703 89.2 12.7
#2 663 20.1 3.0
Fluorene #1 151 17.6 11.7
#2 156 2.1 1.3
Phenanthrene #1 241 22.7 9.4
#2 243 7.9 3.3
Anthracene #1 33.2 6.19 18.6
#2 34.6 1.55 4.5
Fluoranthrene #1 25.3 1.8 7.1
#2 26.0 1.8 7.1
API/EW Phenol #1 40.7 13.5 33.0
Mixture #2 19.0 1.76 9.3
2,4—Diniethylpheflol #1 33.0 9.35 28.3
#2 43.3 8.61 19.9
Naphthalene #1 185 29.4 15.8
#2 165 24.8 15.0
2—Methy lnaphtha lefle #1 265 61.2 23.1
#2 200 18.9 9.5
%RSD Range 1—33
Mean %RSD = 12
Note: Units = pg/L
Extractions were performed in triplicate
All results were at least 2X the detection limit
Extraction Fluid #1 p11 4.9
4—34
-------
Table 6. Multi—Laboratory Semi—Volatiles, Precision
Waste
Compound
Extraction
Fluid
X
S
%RSD
Ammonia Lime
Still Bottoms (A)
API/EW
Mixture (B)
Fossil Fuel
Fly Ash (C)
BNAs
BNAs
BNAs
#1
#2
#1
#2
#1
#2
10043
10376
1624
2074
750
739
7680
6552
675
1463
175
342
76.5
63.1
41.6
70.5
23.4
46.3
Mean %RSD = 54
Note: Units = ig/L %RDS Range for Individual Compounds
X = Mean results from 3—10 labs
Extraction Fluid #1 = pH 4.9 A, #1 0—113
#2 = pH 2.9 A, #2 28—108
B, #1 20—156
B, #2 49—128
C, #1 36—143
C, #2 61—164
4—35
-------
of two waste types plus a VOC free sand/water mixture blank. In
order to assure that volatile compounds were present to extract,
each waste was separated into two fractions and spiked with a
mixture of VOCs. One fraction was spiked at ppm levels and the
other at ppb levels. The spikes were added to the system at two
different times and locations in the extraction process. Group 1
compounds (Table 7) were spiked into the sample immediately after
placing the waste in the ZHE (both ADM and Millipore extractors were
used). The initial waste liquid was removed by the ZifE, as
specified in the TCLP protocol, and collected for analysis. Group 2
compounds were added with the extraction fluid as the fluid was
being pumped into the Z}IE. At the end of the 18—hour extraction
procedure, the extraction fluid was collected from the ZHE and
combined with the liquid collected from the first solid/liquid
separation. Results of the analyses are presented in Table 7. To
reduce the amount of data, compound groups were combined for the
API/EW and sand/water blank wastes.
Although the data presented in Table 7 are reported in percent
recoveries, the reproducibility of percent recoveries is an
indicator of precision. The precision results (RSDs) are
excellent. Only tetrachioroethane and trichioroethylene showed
atypical recoveries. The report suggested that dehydrohalogenation
may have occurred converting tetrachloroethane to trichioroethylene
and Rd. Since both Groups 1 and 2 RSDs fall in the same range, it
is unlikely that the spiking scheme contributed significantly to the
excellent precision.
A second study directed by Oak Ridge National Laboratory (9)
evaluated the performance of ZREs with two wastes. Ten extractions
were performed on fortified wastes that were spiked with halocarbona
and aromatics in a scheme similar to that described in the S—Cubed
study. The liquids from the initial liquid/solid separations and
the final extracts were analyzed individually. Pour of each set of
10 extractions were analyzed by Oak Ridge National Laboratory while
the other six of each set were analyzed by another laboratory. The
results are su trlzed In Table 8.
In general, precision results are similar to those reported In the
study Involving seal—volatile organic compounds (Table 6). Although
both wastes were spiked with a mixture of compounds, the oily
characteristics of the API/EW waste apparently prevented
partitioning of the volatilea into the extraction fluid. Many of
the analytea were either not detected or showed a wide range of
concentrations. This can be seen from the larger RSDs for several
of the compounds, methylene chloride and benzene. Although the
variations may be greater for oily wastes which reduce the
effectiveness and applicability of the method, the variations are
not out of line when compared to other multi—laboratory results. It
4—36
-------
Table 7. Single—Laboratory VOCs, Precision
Waste Compound Spike Level X S %RSD
Group 1 Compounds 1
Ammonia Acrylonitrile 200 ppb 69 5.3 7.6
Lime Still 1 ppm 62 3.6 5.8
Bottoms Carbon disulfide 200 ppb 47 3.3 7.1
1 ppm 20 3.2 16
2—Butanone 200 ppb 75 6.0 8.1
1 ppm 66 6.6 10
Benzene 200 ppb 93 3.0 3.3
1 ppm 46 8.3 18
Toluene 200 ppb 87 4.0 4.7
1 ppm 47 7.1 15
Chlorobenzene 200 ppb 74 2.2 3.0
1 ppm 52 5.7 11
Group 2 Compounds 2
1,1—Dichloroethylene 200 ppb 80 4.3 5.4
1 ppm 89 4.2 0.7
Chloroform 200 ppb 90 2.9 3.2
1 ppm 94 8.0 8.5
1,2—Dichloroethane 200 ppb 93 3.0 3.2
1 ppm 95 7.7 8.1
1,1,1—Trichloroethane 200 ppb 81 3.3 4.1
1 ppm 88 3.2 3.6
Carbon tetrachioride 200 ppb 74 7.0 9.5
3 1 ppm 86 4.8 5.6
Trichloroethylene 200 ppb 136 6.9 5.1
1 ppm 151 13 8.6
1,1,2—Trichloroethane 200 ppb 86 1.6 1.9
1 ppm 88 4.1 4.7
Tetrachioroethylene 200 ppb 63 3.7 5.9
1 ppm 69 5.0 7.3
1,1,1,2—Tetrachioroethane 200 ppb 88 1 12
lppm 96 0 11
1,1,2,2—Tetrachloroethane 200 ppb 16 4.6 29
1 ppm 15 8.4 56
4—37
-------
Table 7 (Continued)
Waste
Compound
Spike
Level
X
S
%RSD
Surrogates
d4_1,4_Dich loroethane
200
ppb
96
3.4
3.5
1
ppm
95
5.6
5.9
Broinofluorobeazene
200
ppb
97
5.8
6.0
1
ppm
99
1.2
1.2
d 8 —Toluene
200
1
ppb
ppm
101
101
3.9
4.1
3.9
4.1
API/EW
Mixture
Group 1 Compounds
Group 2 Compounds
Surrogates
I
5
1
5
1
5
ppn4
ppm 2
ppm 1
ppm
ppm
ppm
51
56
38
35
101
102
ND
ND
ND
ND
ND
ND
7.6
9.0
8.5
7.6
3.0
2.5
Sand,
Spiked
Blank
Group 1 Compounds
Group 2 Compounds
Surrogates
20
200
1
20
200
I
20
200
1
ppb
ppb 2
ppm 2
ppb 2
ppb 1
ppm
ppb
ppb
ppm
62
54
92
91
93
34
95
102
99
ND
ND
ND
ND
ND
ND
ND
ND
ND
8.2
10
7.2
1.5
73
20
2.0
4.6
3.6
4
Mean ZRSD = 9
ZRDS Range = 1—56
Note: X = Mean Z recovery for five replicates
1
2 Compounds spiked into the waste
Compounds spiked into the extraction fluid
Recoveries indicate possible transformations due to dehydrohalogenation
Surrogate and Blank data not included
ND Not determined
4—38
-------
Table 8. Multi—Laboratory (Two Labs) VOCs, Precision
Waste
Compound
Liquid
X
S
%RSD
Ammonia Lime
Still Bottoms
API/EW
Mixture
Benzene
Bromodichioromethane
Bromoform
Carbon tetrachioride
Chlorobenzene
Chloroform
Methylene chloride
Tetrachloroethylene
Toluene
Benzene
Toluene
Methylene chloride
Free liquid
Leachate
Free liquid
Free liquid
Free liquid
Free liquid
Free liquid
Leachate
Free liquid
Free liquid
Free liquid
Leachate
Free liquid
Leachate
Free liquid
Leachate
Free liquid
Leachate
36.5
90.3
24.7
30.1
36
33
32
17.2
235
59.8
35.2
65.2
274
299
208
444
514
452
11.2
47.7
16.3
18.4
47.1
25.8
85.2
11.2
136
31.1
11.6
22.3
137
262
71.5
210
545
424
30.7
52.8
66.2
61.2
34.6
19.4
19.7
65.2
57.7
52.1
32.9
34.2
50.0
87.4
34.4
47.4
106
93.7
Mean %RSD = 53
%RDS Range = 19—106
Note: Units = pg/L
X = Mean results from 9—10 extractions
4—39
-------
does, however, indicate that not all wastes may be effectively
characterized by the TCLP method.
A third study, coordinated by S—Cubed (Blackburn and Show), was a
collaborative study in which 23 laboratories participated. Only 11
laboratories participated in the ZHE testing. Two wastes, API/EW
and mine tailings wastes, were fortified with a mixture of VOCs.
All participating laboratories had prior experience in using the
ZUEs (ADM and Millipore). A aui ary of the results is included in
Table 9.
Precision results for VOCa tend to occur over a considerable range.
However, the range and mean RSD compare very closely to the same
collaborative study metals results (Table 4). Blackburn and Show
concluded that at the 95Z level of significance:
o Recoveries among laboratories were statistically similar.
o Recoveries did not vary significantly between the two sample
types.
o Each laboratory showed the same pattern of recovery for each
of the two samples.
One can, therefore, conclude that the two samples contributed
equally to the leachable VOCs and that the RSD range does not
preclude acquiring consistent results.
TCLP AND MUNICIPAL POTW SLUDGES
Two studies have been conducted to determine the impact of the
TCLP on Publicly Owned Treatment Works (PON) wastes. S—Cubed
(10) and the EPA’s Office of Water (11) have coordinated testing
of twelve PON sludges. These wastes were analyzed as split
samples by S—Cubed, the EPA contract laboratory, and by the POTW
laboratory or by a POTW contracted laboratory. S—Cubed
determined EP and TCLP concentrations of the wastes.
Statistically it is difficult to evaluate the data in terms of
RSDs. However, some general observations are In order. Table 10
presents a suanary of these studies.
Results reported from POTW wastes indicate that analytes of
Interest, TCLP constituents, are present In relatively low
concentrations. Even though the results do vary, indications are
that the TCLP method can adequately and reproducibly be used for
municipal wastes. The only compound approaching the threshold
limit was 2—Butanone found in one waste. An earlier study by the
EPA (12) of six municipal sludges, found that benzene and
chloroform also approached the threshold limit for two sludges.
It appears that of the moat critical application of the TCLP to
P01W wastes Is the ZHEIVOC analysis.
4—40
-------
Table 9. Multi—Laboratory (11 Labs) VOCs, Precision
Waste Compound X S %RSD
Mine Tailings Vinyl chloride 6.36 6.36 100
Methylene chloride 12.1 11.8 98
Carbon disulfide 5.57 2.83 51
1,1—Dichioroethene 21.9 27.7 127
1,1—Dichioroethane 31.4 25.4 81
Chloroform 46.6 29.2 63
l,2—Dichloroethane 47.8 33.6 70
2—Butanone 43.5 36.9 85
1,1,1—Trichloroethane 20.9 20.9 100
Carbon tetrachloride 12.0 8.2 68
Trichloroethene 24.7 21.2 86
1,1,2—Trichloroethene 19.6 10.9 56
Benzene 37.9 28.7 76
1,1,2,2—Tetrachloroethane 34.9 25.6 73
Toluene 29.3 11.2 38
Ch lorobenzene 35.6 19.3 54
Ethy lbenzene 4.27 2.80 66
Trichlorof luoromethane 3.82 4.40 115
Acrylonitrile 76.7 110.8 144
Ammonia Lime Vinyl chloride 5.00 4.71 94
Still Bottoms Methylene chloride 14.3 13.1 92
Carbon disulfide 3.37 2.07 61
1,1—Dichioroethene 52.1 38.8 75
1,1—Dichloroethane 52.8 25.6 49
Chloroform 64.7 28.4 44
1,2—Dich loroethane 43.1 31.5 73
2—Butanone 59.0 39.6 67
1,1,1—Trichioroethane 53.6 40.9 76
Carbon tetrach]oride 7.10 6.1 86
Trichioroethene 57.3 34.2 60
1,1,2—Trichloroethane 6.7 4.7 70
Benzene 61.3 26.8 44
1,1,2,2—Tetrachioroethane 3.16 2.1 66
Toluene 69.0 18.5 27
Chlorobenzene 71.8 12.0 17
Ethylbenzene 3.70 2.2 58
Trichiorofluoromethane 4.05 4.8 119
Acry]onitri le 29.4 34.8 118
Mean ZRSD = 75
%RDS Range 17—144
Note: Units = pg/L
4—41
-------
Table 10. TCLP Summary of Twelve POTW Wastes
Parameter
Laboratory
% of Analytes Detected
TCLP EP
Metals (8 EP Metals)
Semi—Volatiles
Volatiles
EPA
POTW
EPA
POTW
EPA
P OTW
27 11
36 NA
2 NA
2 NA
4 NA
10 NA
Analytes Found in at
Least 50% of the Wastes
Range of Analyte
Concentrations (mgIL)
Proposed Regulatory
Threshold (mg/L)
Barium
Cadmium
Toluene
2—Butanone
Ethy lbenzene
Methylisobutylketone
Xy lene
p—Creso]
Phenol
0.2—3.9
0.02—0.21
0.0008—0.95
0.0165—2.2
0.00060.0085
0.0048—0.064
0.0045—0.05
0.17—1.5
0.0008—0.39
100
5.0
14.4
7.2
NL
NL
NL
10
14.4
Note: NA Not analyzed
I lL No limits have been proposed
4—42
-------
The results from the POTW study are encouraging in light of the
previous studies. Some of the POTW tests were performed by POTW
laboratories or other commercial laboratories which did not have
experience with TCLP equipment. Quality control guidelines, while
Included in the method, were not otherwise specified. In other
words, the results may be more typical of laboratory variations
among all types of laboratories. It could, however, be argued
that since the POTW sludge had relatively low levels of
constituents of concern, that a true test of laboratories’
abilities to perform the test had not been measured.
The POTW study results do not appear to differ greatly from the
method performance studies.
CONCLUSION
From strictly an analytical application, the TCLP provides a
workable method which performs equally well or better than the EP
procedure. With practice and good laboratory skills, the
extraction procedure can become routine. As stated earlier,
difficulties in achieving sample homogeneity and in collecting
representative samples from waate sources, may imply that a single
TCLP analysis of a waste could provide misleading information.
The TCLP was designed to provide reproducible extraction results,
not to determine total constituents of a waste. It Is likely that
some waste constituents may give highly variable results which are
matrix dependent and unpredictable. Distribution coefficients
(lQjs) or sorption coefficients of compounds will vary with the
sample matrix. In general, TCLP performance appears to satisfy
regulatory needs.
ACKNOWLEDGEMENT
The authors wish to acknowledge Mr. Todd A. KIinmell for his
technical critique of the paper. Mr. Kimniell, former U.S. EPA
Project Director for the Development of the TCLP procedure is
currently employed by NUS Corporation, Gaithersburg, Maryland.
Gail A. Hansen Is the current Project Officer at the U.S. EPA.
4—43
-------
R.EFERENCES
Touden, W.J. and Steiner, E.H. Statistical Manual of the Association
of Official Analytical Chemists . AOAC, 1975.
“Inter—Industry Collaborative Study of the Toxicity Characteristic
Leaching Procedure, Addendum to Compilation of Phase IA and
Phase II Data.” Docket Number F—86—TC—FFFFF, September 15,
1986.
Mason, Benjamin J. and Carlile, David W. “Round—Robin Evaluation
for Selected Elements and Anionic Species from TCLP and EP
Extractions.” A Pre—Publication Version of EPRI Report No.
EA—4870, Draft Report, April 25, 1986.
Williams, Liewellyn R., Francis, Chester W.; Maskarinec, Michael P.,
Taylor, David R., and Rotbman, Nancy. “Single—Laboratory
Evaluation of Mobility Procedure for Solid Waste.” EMSL, ORNL,
S—Cubed, ENSECO.
Henry, Betsy. “Evaluation of the ZHE TCLP Protocol.” Final Report,
Contract No. 68—01—7075, ERCO/ENSECO, June 11, 1986.
“Background Document, Resources Conservation and Recovery Act,
Hazardous and Solid Waste Amendments of 1984, Land Disposal
Restrictions Rule, Solvents and Dioxins.” U.S. EPA, November
7, 1986.
Blackburn, W.B. and Show, I. “Collaborative Study of the Toxicity
Characteristics Leaching Procedure (TCLP).” Draft Final
Report, Contract No. 68—03—1958, S—Cubed, November 1986.
Taylor, David R. and Shurtleff, Arthur B. “Precision Evaluation of
the Toxicity Characteristic Leaching Procedure (TCLP) for
Volatile Contaminants.” Final Report, Contract No. 68—01—7266,
S—Cubed, July 2, 1986.
Maskarinec, M.P. and Francis, C.W. “Precision Analyses for the
Zero—Headspace Extractor.” Draft Interim Report, Contract No.
DE—ACO5—840 R211400, Oak Ridge National Laboratory, January 15,
1986.
Taylor, C.L.; Blackburn, W.B.; Swanson, G.R. “Analytical Data for
POTW Sludge Testing.” Contract No. 68—01—7266, S—Cubed,
October 1986.
Walker, John M. “Cooperative Testing of Municipal Sewerage Sludges
by the Toxicity Characteristics Leaching Procedure and
Compositional Analysis.” Draft, Residuals Management Branch,
U.S. EPA, May 15, 1987.
4—44
-------
Walker, John M. “Report on Six POTW Sludges Tested for
Compositional and TCLP Analysis.” Memorandum, U.S. EPA, July
11, 1986.
4—45
-------
DETERMINATION OF ORGANIC COMPONENTS
IN LEACH1 .TES - A SURVEY
James A. Poppiti, Eric Johnson, Finnegan MP T, San Jose, CA
ABSTRACT
The mobility of organic compounds in the environment is dependent on
compound solubility, sorption effects, pH behavior, and pH of the
mobile phase. The mobility of model organic compounds has been
evaluated to determine the effect of each of these factors.
Model compounds were applied to soil columns containing varying
amounts of humic matter. The columns were eluted under different pH
conditions and the concentration of the model compounds was determined
in the resulting leachates by gas chromotography/rnass spectrometry.
The leachate volume and compound’s concentration in the leachate
indicate the compound’s mobility under the test conditions. The
effect of soil composition, pH of the mobile phase, and compound pH
behavior were evaluated using this approach. Leaching behavior and
mobility of the compounds studied are explained in terms of these
factors.
INTRODUCTION
The effect of physical/chemical factors on migration of toxic
compounds in the environment has not been widely investigated. It is
known that water soluble organics, such as solvents, quickly migrate
from waste to groudwater, and that the migration results from the
leaching of the organics from the waste. Little is known, however,
about the mechanism controlling the migration of chemicals contained
in the leachate.
The migration of toxic organic chemicals in the environment was
simulated by applying three organic chemicals to sand and soil
columns. Migration was evaluated by varying the pH of the leaching
media, the amount of humic material contained in the column, and
compound solubility behavior.
EXPERIMENTAL
Two sets of three columns (6 cm by 1.5 cm id.) were prepared. Each
column set consisted of a column filled with sand, a 1:1 sand/soil
mixture, and soil. One set of columns was eluted, first with water,
and then with .1N, pH 5, Acetic Acid buffer. The other set was eluted
with water and then .1N, pH 8.5, Acetate buffer. After collection of
20 mL of eluate, 10 mg each of Acetophenone, p—Nitroaniline, and
p—Chlorobenzoic acid were added to the columns. Eluates from the
columns were collected as 50 mL aliquots. A total of 400 mL of eluate
was collected from each column. All chemicals were obtained from
Aldrich Chemical and were used without purification.
Eluates were analyzed by GC/MS on a Finnegan MAT INCOS 50 GC/MS
system. Samples were injected directly onto a J&W DB—5 column using
4—47
-------
sp it injection technique. The CC was progranuned from 100 to 200°C at
10 C per minute. Quantitation was performed using Finnegan Target
Compound Analysis (TCAj software using a three point calibration.
RESULTS
The column experiments were designed to evaluate the effects of
leachate pH, stationary phase organic content, and acid/base compound
behavior on compound migration. A low (pH 5) and high (pH 8.5) pH
were chosen to represent the range of pHs that might be reasonably
encountered in the environment. The amount of soil in each column
was varied to determine what effect, if any, organic carbon contained
in soil would have on migration. The three compounds used were
selected based on their water solubility and acid/base/neutral
characteristics.
In all substrate cases, the Acetophenone was the first compound to
elute and was essentially completely eluted within the first 150 niL
regardless of pH. Each 50 n iL passing through the column is roughly
equivalent to 11 inches of rain fall. The behavior of the
chlorobenzoic acid and nitroaniline were greatly affected by the pH of
the leaching media and to a lesser extent column substrate
composition. As expected, the acid eluted more quickly at high pH and
the aniline at low pH. The make—up of the column had almost no effect
on elution of these compounds.
DISCrJSSICt
Column composition had very little effect on migration behavior of the
compound tested. In this case, the concentration of each compound in
the eluate was determined only by the compounds solubility and pH
behavior. While this result is not entirely surprising, the widely
accepted view that soil attenuates compound migration may not be
warranted.
4—48
-------
EVALUATION OF THE TCLP FOR DETERMINING THE RELEASE
PYTENTIAL OF OILY WASTES
Robert S. Truesdale, Steven L. Winters, Research Triangle Institute,
Research Triangle Park, NC; Gail Ann Hansen, Office of Solid Waste,
U.S. EPI OSW, Washington, D.C.
ABSTRACT
Oily wastes are important because they are produced in large
quantities and often contain toxic organic and inorganic constituents.
Oily waste constituents may be released fron landfills by several
mechanisms. They may be released as immiscible liquids or as aqueous
colloidal suspensions of these liquids. Infiltrating water may leach
various constituents from the wastes, resulting in release of
contaminants as dissolved compounds. Case histories demonstrate the
potential for oily wastes to contaminate ground water by any or all of
these mechanisms; the relative importance of each depends on the
properties of the waste and the characteristics of the environment in
which it has been disposed.
The TCLP was designed to model release of contaminants from a
reasonable worst—case mismanagement scenario; codisposal of 5 percent
industrial waste with 95 percent municipal refuse in a sanitary
landfill. To evaluate how well the TCLP simulates release of
hazardous waste constituents under this scenario, the procedure was
performed on 11 different wastes. However, oily wastes were not
evaluated. The objective of this research effort is to determine the
applicability of the TCLP to oily wastes by answering the following
questions: 1) Does the TCLP adequately estimate the amount and
quality of hazardous contaminants that will be released as organic
liquids from a Subtitle D landfill?; 2) Does TCLP adequately
estimate the quality of aqueous leachate that will be generated by the
dissolution of oily wastes in a Subtitle D landfill?; and 3) Are
there any procedural problems that may be encountered when applying
the TCLP to oily wastes? This presentation describes current progress
of Phase I of this research effort, addressing Questions 1 and 3.
Question 2 will be addressed during Phase 2.
Phase I research will determine whether or not the filter in the
initial filtration step of the TCLP acts as an exclusion filter for
certain constituents of oily wastes for which soil in landfills does
not. To make this determination, the fractional amounts of four
specific oily wastes that pass through the TCLP filter will be
compared with the fractional amounts of the same wastes that pass
through packed soil columns designed to represent landfill conditions.
In addition, TCLP extractions will be performed on the four wastes in
triplicate to evaluate the single—laboratory precision of the
procedure and to identify any procedural problems associated with
testing oily wastes in the TCLP test devices. Both TCLP test devices
(extraction bottles and the zero headspace extractor) will be included
in this evaluation.
4—49
-------
METALS AND
MISCELLANEOUS
ANALYTES
chairperson
Theador Martin
Research chemist
Environmental Monitoring
and Support Lab
U.S. EPA
26 W. St. Clair
Cincinnati, Ohio 45268
-------
EVALUATION OF MICROWAVE TECHNIQUES TO PREPARE
SOLID AND HAZARDOUS WASTE SAMPLES FOR ELEMENTAL ANALYSIS
David A. Binstock, Peter M. Grohse, Percy L. Swift, and Alvia
Gaskill, Jr., Research Triangle Institute, Research Triangle Park,
North Carolina; Thomas R. copeland, ERCO/A Division of ENSECO, Inc.,
Cambridge, Massachusetts; Paul H. Friedman, Office of Solid Waste,
U.S. Environmental Protection Agency, Washington, D.C.
ABSTRACT
The use of microwave energy to facilitate sample decomposition prior
to elemental analysis is now receiving considerable attention. Both
wet and dry digestions are achievable. When microwave energy is
used in combination with acid mixtures in closed vessels, the
combined pressure and rapid heating can reduce digestion times to a
few minutes from hours or days that may be required for open beaker
digestions. This savings in time and labor is significant and has
prompted the Office of Solid Waste to evaluate this technology as a
preparation tool for solid wastes. Of particular interest are used
oils and other fuels slated for incineration, incinerator ash and
particulates from these processes.
This study reports on the evaluation of a commercially available
microwave oven sample preparation system for this application. The
effect of sample preparation conditions, including the acid matrix,
heating time, and pressure were evaluated for fifteen toxic or
hazardous elements in particulates, ashes, oils, and oil fuels.
Analyses were carried out by inductively coupled plasma emission
spectroscopy.
INTRODUCTION
The techniques that are typically used to prepare Resource
Conservation and Recovery Act (RCRA) wastes for analysis for metals
and other elements are relatively time consuming, requiring several
hours to several days to complete. They also often involve the use
of acid digestions and thermal decomposition steps which may result
in analyte losses, incomplete recoveries, or sample contamination.
These limitations are well known to the analytical community and to
the end users of these data in EPA, states, and industry. The
resulting inefficiency of these techniques reduces laboratory sample
throughout, drives up the cost of analytical testing and impedes
decision—making. Given these concerns, the USEPA Office of Solid
Waste is interested in developing cost effective sample preparation
techniques for metals and other elements in environmental and
process waste samples. Once developed, these techniques can then be
written as methods for inclusion in “Test Methods for Evaluating
Solid Waste, SW—846 ” and made available to the user community.
5 -1
-------
One particularly attractive sample preparation technique that is now
receiving considerable attention is microwave assisted sample
dissolution. The use of microwave energy to facilitate sample
decomposition prior to elemental analysis has received considerable
attention in recent years. Both wet and dry digestions are
achievable. When microwave energy is used in combination with acid
mixtures in closed vessels, the combined pressure and rapid heating
can reduce digestion times to a few minutes, from hours or days that
may be required for open beaker digestions. This savings in time
and labor is significant and has prompted the Office of Solid Waste
to evaluate this technology as a preparation tool for solid wastes.
Of particular interest are used oils and other fuels slated for
incineration, incinerator ash, and particulates from these processes.
Previous evaluative work with those matrices has been carried out by
Nadkarni 1 -. Using a commercial microwave oven and an HF/aqua—regia
digest, National Bureau of Standards (NBS) Coal 1632a and NBS SRN
1633a Fly Ash were solubilized. Copeland 2 (1985) used a HNO 3 /H 2 0 2
microwave procedure to prepare waste oils for determination of As,
Be, Cd, Ni, and Pb.
This study reports on the evaluation of a commercially available
microwave oven sample preparation system. The effect of sample
preparation conditions, including the acid matrix, heating time, and
pressure were evaluated for toxic or hazardous elements in
particulates, ashes, oils and oil fuels. Analyses were carried out
by inductively coupled plasma emission spectroscopy.
EXPfltIMENTAL METHODS
Microwave Oven
The MDS—81D Microwave system (CEM Corporation, Indian Trail, NC) was
used for this study. The oven re8elflbles a standard microwave oven,
but is equipped with additional features to facilitate sample
preparation. For example, the Teflon—coated microwave cavity has a
variable speed corrosion resistant exhaust system. The main element
of the system couples a precise microwave variable power system with
a programmable micro—processor digital computer. Other elements
include a rotating turntable, Teflon vessels with caps and a
patented pressure relief valve, a capping system, and a cooling tank.
The Teflon sample vessels and caps are designed to withstand
pressures up to 100 psi and temperatures up to 2000C.
Inductively Coupled Plasma Emission Spectrometry (ICPES )
AU analytical measurements were performed using an instrumentation
Laboratory 200 ICAP.
5—2
-------
Reagents
All inorganic acids used were of “Ultrex” quality, from J. T. Baker
Chemical Co. Other chemicals were of analytical reagent grade
quality. Deionized water of 18 MA/cm specific resistivity was used.
Combustion Source Materials
The evaluation of microwave procedures was carried out using the
following materials:
o NBS coal 1632a
o NBS coal fly ash 1633a
o Municipal incinerator dust
o Oil—fired power plant fly ash
o Used motor oil
Microwave Preparation Procedures
Total Digestion Procedure A 300 mg sample is placed in a 60 mL
Teflon digestion vessel equipped with a relief valve and treated
with 1 niL concentrated HNO 3 , 3 mL HF, and 0.5 niL HC1O 4 . The vessel
is sealed and heated in the microwave oven at 15 percent power for 5
minutes, followed by 30 percent power for 15 minutes. The vessel is
cooled, the cap removed, 3 niL HF added, resealed and heated at 20
percent power for one hour. The cap is then removed and the
contents evaporated until fuming ceases. Six niL of 20 percent HNO 3
are added and evaporated to dryness; 15 niL 20 percent Hc1 added, the
vessel sealed and heated at 10% power until dissolution of the
residue is achieved. The vessel is uncapped, evaporated to dryness,
and 15 niL 5 percent HC1 is added and heated until the solution
clears. The vessel is cooled and diluted with deionized water to 50
niL. Total digestion time is two hours.
HF/Aqua—Regia Procedure A 300 mg sample is placed in a 60 niL Teflon
digestion vessel equipped with a relief valve and treated with 6 niL
aqua—regia (3HC1:1HNO 3 ) and 2 niL HF. The vessel is sealed and
heated in the microwave oven at 100 percent power for 3 minutes.
The vessel is cooled, the cap removed, the digestate filtered and
transferred to a 50 niL volumetric flask with deionized water.
Oily Waste HNO 3 Procedure A 300 mg sample is placed into a 120 niL
Teflon digestion vessel equipped with a relief valve and 12 mL
concentrated HNO 3 is added. The vessel is sealed, capped, and
digested according to the six steps below:
5—3
-------
Step 100% power 0% power Vent
1 6 minutes 2 minutes vent
2 2 minutes 3 minutes vent
3 6 minutes 3 minutes vent
4 7 minutes 4 minutes vent
5 8 minutes 4 minutes vent
6 10 minutes Cool to room temp. vent
Total digestion time is 40 minutes.
RESULTS
Total Digestion Procedure
A total digestion (Hc10 4 /HF/Hcl) microwave procedure was used to
prepare NBS SRI4 1633a fly Ash. With the exception of Mn and Pb,
recoveries were all within 25 percent of the NBS values (Table i).
In addition, results for most elements are comparable to those from
a more time—consuming conventional open—beaker digestion of the same
material.
A modified total digestion procedure was used to prepare NBS SRN
1632a coal. Due to the potential explosion hazard of the HC1O 4 /coal
mixture, the elements with the exception of Mn, Pb, and Co (Table
2). The low recovery for Pb is more likely a result of the
analytical procedure (ICP) rather than the microwave preparative
step.
Tables 3 and 4 provide results for the total digestion of an
oil—fired power plant fly ash and municipal incinerator dust.
Results for Al, Mn and V for the power plant ash, and Al, As, Mn and
V for the incinerator dust were corroborated by Neutron Activation
Analysis.
5-4
-------
Table 1. Analysis of NBS SRM 1633a Fly Ash
(Total Digestion Procedure)
(ug/g)
Element Mean ± S.D. (n = 3) NBS Values % Bias
Conventional 3
Digestion
Al 160,000 ± 28,000 (140,000) +14
148,000
Be 11.9 ± 0.6 (12) —1
15
Ca 9360 ± 1060 11,000 ± 100 -15
11,200
Co 49.5 ± 0.3 (46) +8
---
Cr 169 ± 59 196 ± 6 -14
160
Cu 90.3 ± 3 118 ± 3 -24
110
Mn 133 ± 2 (190) —30
175
Ni 130 ± 25 127 ± 4 +2
130
Pb 45.1 ± 20 72.4 ± 0.4 -38
82
V 313 ± 3 (300) +4
310
Zn 205 ± 8 220 ± 10 —7
210
n = number of replicates
( ) = uncertified value
% bias = % difference between certified and experimental
results
Table 2. Analysis of NBS SRM 1632a Coal
(Total Digestion Procedure)
(ug/g)
Mean ± S.D. (n= 2)
Element RTI NBS Values
% Bias
Al 30,200 ± 5900 (31,000)
Ca 2050 ± 340 2300k
-3
-11
Co 9.72 ± 0.19 (6.8)
+43
Cr 31.2 ± 2.8 34.4 ± 1.5
-3
Cu 15.6 ± 0.6 16.5 ± 1
-5
Fe 10,400 ± 600 11,100 ± 200
-6
Mg 1030 ± 180 -—
Mn 19.6 ± 0.6 28 ± 2
-30
Ni 17.3 ± 1.8 19.4 ± 1
—11
Pb ND 12.4 ± 0.6
Sc 5.45 ± 0.56 (6.3)
-14
V 44.2 ± 0.7 44 ± 3
+0.4
Zn 34.9 ± 0.3 28 ± 2
+25
n = number of replicates
( ) = uncertified value
% bias = % difference between certified and experimental results
5-5
-------
Table 3. Analysis of Oil-Fired Power Plant Fly Ash
(Total Digest Procedure)
(ug /g)
Element
Rh NM
Mean ± S.D. (n=3) Mean ± S.D. (n=2)
Al
44,700 ± 7120 43,600 + 200
Ca
12,700 ± 1860
Co
697±86
Cr
834 ± 196
Cu
634±85
Fe
77,000 ± 10,800
Mg
17,500 ± 2670 <14,300
Mn
438±67 504+40
Ni
15,800 ± 2010
Pb
1470 ± 214
V
24,700 ± 3380 23,600 ± 300
Zn
5,050 ± 681
NM
= Neutron
Activation Analysis
n
number of
replicates
Table 4. Analysis of Municipal Incinerator Dust
(Total Digestion Procedure)
(ug/g)
Element
RTI NM
Mean ± S.D. (n=3) Mean ± S.D. (n=2)
Al
12,500 ± 1700 11,500 400
As
87.5 ± 21.3 115 ± 6
Ca
Cd
27.8 ± 3.0%
75.8 ± 8.3 <154
Co
5.64 ± 1.4
Cr
32.4 + 4.5
Cu
134 ±15 <847
Fe
Mg
7170 1150
3810 ± 580 <13,000
Mn
209±27 169±5
NI
13.5 1.2
Pb
1970 ± 230
Sc
1.82 ± 0.24
V
33.9 ± 1.8 35.0 ± 3.0
Zn
9210 ± 1610
NAA = Neutron Activation Analysis
n = number of replicates
5-6
-------
Table 5. Analysis of NBS SRM 1632a Coal
(Modified HF/Aqua—Regia Procedures)
Mean ± S.D. (n = 3)
(ug /g)
Element
Aa
Bb
Cc
NBS Values
Al
24,600
22,700 ±
1600
19,600 ± 3500
31,000
Cd
<0.5
<0.5
<0.5
0.17 ± 0.02
Co
ND
7.37 ±
0.73
5.77 ± 0.74
(6.8)
Cr
20.1
24.6 ±
2.7
23.8 ± 5.3
34.4 ± 1.5
Cu
10.3
11.4 ±
1.0
10.5 ± 4.0
16.5 ± 1
Fe
7270
8940 ±
267
9180
11,100 ± 200
Mn
19.6
28.2 ±
1.9
20.7 ± 1.9
28 ± 2
Ni
18.3
12.9 ±
2.2
16.6 ± 0.4
19.4 ± 1
Pb
7.2
15.4 ±
3.2
9.68 ± 3.00
12.4 ± 0.6
V
33.0
47.6 ±
3.1
38.9 ± 2.2
44 ± 3
Zn
23.7
15.1 ±
2.3
13.7 ± 5.5
28 ± 2
aHF/Aqua...Regia, microwave 3 minutes @ 100% power, n = 1
bHF/Aqua_Regia, microwave 6 minutes @ 75% power
cAshed at 400°C 4 hours-HF/Aqua-Regia, microwave 3 minutes @ 100% power
n = number of replicates
( ) = uncertified value
Table 6. Analysis of Used Motor Oil
(Oily Waste HNO 3 Digestion)
Element
Microwave/ICP
Mean ± S.D. (n=2)
Spike Recovery (%)
NM
Mean ± S.D. (n=2)
Al
17.2 ± 3.6
83.9
20.6 ±
4.3
Ca
632±6
Cd
2.23 ± 0.21
6.64 ±
1.00
Cr
6.00 ± 0.59
71.9
Cu
14.8 ± 0.3
81.0
17.9 ±
3.0
Fe
148 ± 1
93.7
Mg
380 ± 6
98.5
418 ±
71
Mn
38.1 ± 0.3
25.8 ±
0.5
Ni
ND
74.5
Pb
679±4
102
Zn
1160 ± 10
NAA = Neutron Activation Analysis
n = number of replicates
5—7
-------
Table 7. Comparison of Digestion Time of
Microwave and Conventional Techniques
Acid Digestion
Microwave
Conventional
Nitric Acid
40 minutes
4
to 6 hours
Hydrofluoric/aqua—regia
3 to 10 minutes
6
to 8 hours
Total digestion
2 hours
16
hours
Keywords
Microwave oven, R RA wastes, acids, combustion, preparation,
digestion, metals, fuels, incineration, ashes
5-8
-------
HF/Agua—Regia Procedure
NBS SRM 1632a coal was digested using three modifications of the
HF/Aqua—Regia microwave procedure (Table 5). The modified
conditions were 3 minutes at 100% power (column A), 6 minutes at 75%
power (column B), and a dry ash of 4 hours, followed by 3 minutes at
100% power (column C).
Microwave digestion for 6 minutes at 75% power yielded higher
recoveries for Co, Cr, Mn, Pb, and V than the other variations,
whereas digestion for 3 minutes at 100% power gave highest
recoveries for Ni and Zn. Dry ashing followed by microwave
HF/aqua—regia gave the lowest recoveries, with the exception of Fe,
which was higher.
Oily Waste HNO 3 Procedure
The oily waste microwave procedure was used to digest a used motor
oil (Table 6). With the exception of Cd, there is good agreement
between microwave/ICP—AES results and those by Neutron Activation
Analysis. In addition, acceptable recoveries were obtained from
spiking the used oil with NBS 1084 —— wear metals in oil.
CONCLUS IONS
Our preliminary work indicates that substantial time/cost savings
can be achieved using microwave digestion —— particularly with
closed vessel procedures. Table 7 illustrates the considerable time
savings for the three procedures used in this study compared to
conventional techniques. There is also a potential reduced need for
reagents such as HC1O 4 and HF.
Future work will involve refining the acid digestion/microwave
conditions for combustion source samples and other RCRA wastes.
REFERENCES
1. R. A. Nadkarni, “Application of Microwave Oven Sample
Dissolution in Analysis,” Anal. them . 56:2233 (1984).
2. T. Copeland, Methods for Determining As, Be, Cd, Cr, Ni and Pb
in Waste Oils , EPA/OSW Final Report, Contract No. 68—01—7075,
WA No. 3 (1986).
3. University of Missouri (1985).
4. E. S. Gladney, Compilation of Elemental Concentration Data for
NBS Biological and Environmental Standard Reference Materials ,
Los Alamos Scientific Laboratory (1981).
5-9
-------
RESULTS OF AN INTERLABORATORY STUDY OF ICP METHOD 6010
COMBINED WITH DIGESTION METHOD 3050*
Thomas A. Hinners, Research Chemist, U.S. Environmental Protection
Agency, Environmental Monitoring Systems Laboratory, Las Vegas,
Nevada; and Clifton L. Jones, Vernon F. Hodge, Donald M. Schoengold,
Homigol Biesiada, Thomas H. Starks, and Joseph E. Campana,
Environmental Research Center, University of Nevada, Las Vegas,
Nevada
ABSTRACT
An Interlaboratory study of solid wastes using inductively coupled
plasma atomic emission spectroscopy (ICP—AES), which is included in
the EPA methods publication SW—846, was performed with nine
participating laboratories. The study focused on the application of
Method 6010 for the analysis of solid materials including dried
sludges, sediments, and fly ash. Spiked predigests of seven
different solid materials and several quality control (QC) materials
were used to characterize and evaluated Method 6010 when sample
digestion was excluded as a variable.
In practice, the digestion of solid samples is necessary to apply
Method 6010 to the analysis of solid wastes. Therefore, a parallel
study of Method 3050 (Acid Digestion of Sediments, Sludges, and
Soils) was included in the collaborative study. A statistically—
designed homogeneity study on the seven solids was performed by the
coordinating laboratory prior to the colloboration study in order to
measure the between—sample and within—sample variability of the
solid portions that were to be sent to the participating
laboratories. Portions of the seven homogeneous solids with
appropriate QC samples and spiking solutions were provided to the
participating laboratories to be digested by Method 3050 and
analyzed according to Method 6010. Because homogeneous solid
samples were used in this study, the typically large variation
contributed by sampling was minimized. This collaborative study,
based on the experimental design involving undigested and
predigested solids, provides data on the interlaboratory and
Intralaboratory variabilities of Method 3050 and Method 6010
together and independent of one another.
*Ajthough the research described in this article has been
supported by the United States Environmental Protection Agency
through Contract Number 68—01—7159 to the University of Nevada,
Las Vegas, Nevada, 89119 and Viar and Company, Alexandria,
Virginia, 22314, it has not been subjected to Agency review and
therefore does not necessarily reflect the views of the Agency,
and no official endorsement should be inferred. Mention of trade
names or commercial products does not constitute endorsement or
recommendation for use.
5-11
-------
One spiked predigest had a high level (1,000 ppm) of vanadium and
molybdenum In order to test the Impact on the measurement of the
other analytea. The analytical results on this spiked predigest
demonstrate that certain elements, in particular antimony, arsenic,
selenium and thallium, give analytical results with high uncertainty
when the interfering elements vanadium and molybdenum are present at
high levels.
The statistical data obtained on the homogeneity study of the solids
used in the study will be reported, and the variation in analytical
results as a function of the digestion method and the instrumental
method will be discussed. Results obtained on both sequential and
simultaneous ICP—AES as well as atomic absorption spectrophotometry
will be compared statistically. The use of the method of standard
additions (MSA), required In Method 6010, will be evaluated
critically, because the application of the NSA affects the
economics, turnaround time, and the practicality of the method, as
well as the data quality. On average the use of the method of
standard additions did not provide ICP—AES data that was more
accurate than that obtained by normal calibration when
atomic—absorption data were used as the reference values.
INTRODUCTION
An Interlaboratory study of solid wastes using the EPA analytical
Method 6010 entitled “Inductively Coupled Plasma Atomic Emission
Spectroscopy” (ICP—AES), which is included in the EPA methods
publication SW—846, was performed with nine participating
laboratories. This interlaboratory study concentrated on the
application of Method 6010 for the determination of 23 elements In
seven solid materials including dried sludges, sediments, and fly
ash. The 23 target elements are: Al, Sb, As, Ba, Be, Cd, Ca, Cr,
Co, Cu, Fe, K, Pb, Mg, Mn, Mo, Ni, Se, Ag, Na, Ti, V, and Zn. This
study followed a single—laboratory evaluation that investigated the
application of Method 6010 to a variety of aqueous and solid—waste
samples. The different waste matrices studied in the
single—laboratory evaluation required the utilization of several
different digestion procedures. In contrast, this interlaboratory
study examined Method 6010 for the analysis of solid wastes that
were digested using a single digestion procedure.
Since the digestion of solid samples is necessary to apply Method
6010 for the analysis of wastes, a thorough study of Method 6010
must also include digestion as a variable. Consequently, a parallel
study of Method 3050 (Acid Digestion of Sediments, Sludges, and
Soils) was included as an integral part of the Interlaboratory
study. The present study was designed to determine the performance
of Method 6010 both independent of and together with the Method 3050
digestion procedure.
5—12
-------
Seven solid materials, representative of solid wastes, were selected
as the method evaluation materials. Three of the materials (river
sediment, coal fly ash, and estuarine sediment) are Standard
Reference Materials from the National Bureau of Standards, and one
material (the mine tailing) is an EPA reference material. The
other three solids (a contaminated soil and two industrial sludges)
were obtained from the EPA. A detailed, homogeneity study was
performed by the coordinating laboratory before the solids were
distributed to the participating laboratories. The results
indicated that the solid samples were homogeneous.
Sixteen grams of these homogeneous solids were distributed to the
laboratories to be digested by Method 3050, both unspiked and
spiked. The spiking solution that was provided to the laboratories
contained 19 of the 23 target elements and was designed to be added
to the solids prior to digestion to bring the concentrations of the
19 elements In the laboratories’ digests to minimum levels of about
100 times the corresponding “Estimated Instrumental Detection
Limits” given in Method 6010. It was not necessary to spike Al, Ca,
Fe, and Mg into the bulk digests (or spiked solids) because of the
high endogenous concentrations of these metals in the 7 solid
samples. Having each laboratory spike portions of the solid samples
with the spiking solutions prior to digestion assured that each
laboratory used equally spiked aliquots of the solid (which
eliminated the need to create uniformly—spiked solids for
distribution). The resulting unspiked and spiked solids digests
were analyzed by Method 6010.
In order to remove sample—preparation variability from measurement
variability, bulk digests of the 7 solid samples were prepared by
the coordinating laboratory by digesting 40 grams of each of the
seven solid materials. These bulk digests were spiked with. standard
solutions in sufficient quantity to yield two liters of solution
with the concentrations of all 23 elements at about 100 times the
corresponding “Estimated Instrumental Detection Limits.” Thus, the
spiked bulk digests of the seven solid samples were very similar in
composition to the spiked solids digests that were prepared by the
laboratories. Data from the Method 6010 analysis of these spiked
bulk digests could be compared to data from the spiked solids in
order to estimate the variances due to the digestion and analysis
procedures. In order to test the effects of high levels of V and Mo
on the determination of the other analytes by Method 6010, the
spiked bulk digest from the fly ash solid was also spiked to contain
0.1 percent of these interfering elements.
In addition to the solid samples and the spiked bulk digests, two QC
solutions containing the target elements were provided to the
participating laboratories for analysis with and without digestion.
Because these solutions were carefully prepared and verified by the
coordinating laboratory, the results could be used to estimate the
5—13
-------
accuracy of the Methods. Other solutions were provided to the
participating laboratories to insure high ICP—AES data quality.
These were initial calibration verification solutions and an
interference check solution.
The results of this collaborative study yielded quantitative
information on the precision and accuracy of Method 6010,
independently and together with Method 3050. Data obtained on
sequential and simultaneous ICP—AES Instruments as well as by atomic
absorption spectroscopy (AAS) were compared statistically, and the
results are reported. The method of standard additions (NSA) is a
conditional requirement of Method 6010, 80 its effect on data
quality was Investigated.
RESULTS AND DISCUSSION
The wide range of instrumental detection limits (IDL’s) reported by
the participating laboratories are listed in Table 1 along with the
IDL’s required for the 6010 study and other lists. The UNLV
Recomaended Required Instrumental Detection Limits” (to ensure more
uniform levels of detection) given in Table 1 were obtained by
inspection of the reported mean IDL’s, the “Estimated Instrumental
Detection Limits listed in Method 6010,” and the minimum reported
IDL’s.
This multilaboratory evaluation of Method 6010 demonstrates that the
method, as described, is capable of achieving excellent precision as
estimated in terms of percent relative standard deviation (percent
RSD), for the determination of the 23 elements In quality control
(QC) solutions (Table 2). These QC solutions contained the 23
elements at concentrations of approximately 100 times the
instrumental detection limits, and the solutions were
interference—free in that no Interfering elements were present at
high concentrations. The percent RSD’s for the elements range from
3.1 percent to 9.1 percent for the QC solutions that were analyzed
by Method 6010 without digestion and from 2.6 percent to 52 percent
for the QC solutions that were analyzed after digestion by Method
3050. The median percent RSD’s for the 2 sets of QC solutions are
6.5 and 6.7 percent, respectively. This precision is considered
excellent for these solutions. Silver with a percent RSD of 52 Is
the lone outlier in the QC solution set that was digested before
analysis.
The interlaboratory precision for Method 6010, with digestion
eliminated as a variable, was determined for the 23 elements in the
spiked bulk digests of six representative solid complex matrices,
including river and estuarine sediments and industrial sludges
(Table 3). The analyte concentrations In these spiked bulk digests
were about 100 tImes the Instrumental detection limits. The median
percent RSD’s for the 6 sedIments across 23 elements range from 6.8
5-14
-------
TABLE 1. INSTRUMENTAL DETECTION LIMIT SUMMARY (ug/L)
UNLV
Range for
Participating
Mean for
Participating
6010 Study
Required
CLP#
Recommended
Required
Element Laboratories
Laboratories
IDL
CRDL
IDL
Al 17—192 101 200 200 100
Sb 25—116 70 200 60 70
As 22—400 139 200 10 100
Be 1—4 2.4 5 5 2
Cd 3—18 6.4 20 5 5
Ca 4—4500 660 5000 5000 50
Cr 4—31 11 30 10 10
Co 4—38 16 50 50 20
Cu 3—24 11 30 25 15
Fe 3—80 30 100 100 30
Pb 10—200 89 200 5 70
Mg 1—3600 550 5000 5000 30
Mn 1—12 4.3 20 15 4
Mo 4—60 22 30 ** 15
Ni 8—31 20 60 40 20
Se 41—590 200 300 5 150
Ag 3—20 10 30 10 10
Ti 49—400 180 200 10 150
V 3—50 23 50 50 20
Zn 2—18 9 20 20 10
Ba 1—60 13 200 200 15
Na 13—4940 930 5000 5000 150
K 189—4800 1230 5000 5000 500
Potassium determination is highly dependent upon operating
conditions and plasma position so a value was not included.
** Molybdenum is not assayed in the CLP.
* CLP is Contract Laboratory Program of the EPA and CRDL is the
Contract Required Detection Levels (where atomic absorption is
necessary for some elements).
+ UNLV is University of Nevada at Las Vegas.
-------
TABLE 2. PRECISION VALUES FOR THE MEASUREMENT OF THE 23
TARGET ELEMENTS BY METHOD 6010 IN THE QUALITY
CONTROL SOLUTIONS
Method
Element
6010 a
%RSDC
Method 6010 with
Method 3050 b
Element %RSDC
Ag
9.1
Ag
52
Ti
8.5
As
13
Zn
8.3
Cd
11
Cr
8.2
Se
10.1
Sb
7.7
Ti
9.5
Se
7.5
Mo
8.9
Ca
7.4
V
8.4
Cd
7.0
Sb
7.7
Mo
6.9
K
7.2
K
6.6
Zn
6.8
V
6.6
Ba
6.8
Mg
6.5
Ca
6.7
As
6.4
Ni
6.6
Al
6.3
Mg
6.2
Co
5.9
Na
5.8
Fe
5.9
Pb
5.6
Pb
5.9
Fe
5.3
Be
5.8
Cr
5.2
Ni
5.7
Mn
4.5
Cu
5.6
Co
4.3
Mn
4.3
A].
4.0
Na
4.2
Be
2.9
Ba
3.1
Cu
2.6
a The median percent RSD is 6.5 percent.
b The median percent RSD is 6.7 percent.
C The percent RSD data are presented in order of increasing
precision.
5-16
-------
TABLE 3.
01
PERCENT RSD’s FOR THE DETERMINATION OF THE 23 TARGET ELEMENTS
IN THE SPIKED BULK DIGESTS
ELEMENTS
HAZARDOUS
WASTE
1
RIVER
SEDIMENT
FLY
ASH
ESTUARINE
SEDIMENT
INDUSTRIAL
SLUDGE
ELECTRO-
PLATING
SLUDGE
MINE
TAILING
Al
11
19
16
1.9
11
13
7.6
Sb
5.6
52
73
8.7
3.2
24
4.4
As
13
11
83
22
25
8.6
5.3
Be
5.8
5.8
57
4.8
6.4
9.9
8.5
Cd
11
6.6
5.7
7.6
3.1
9.8
12
Ca
8.8
9.4
5.6
5.3
8.5
7.0
8.7
Cr
6.2
5.5
36
7.6
5.8
7.8
39
Co
11
14
21
6.8
6.7
11
13
Cu
4.4
4.3
9.7
6.0
11
7.8
12
Fe
6.6
8.3
8.8
6.0
6.9
8.4
8.4
Pb
15
7.2
22
4.7
3.9
5.6
8.0
Mg
8.8
8.1
15
9.4
8.0
20
10
Mn
10
13
14
10
10
9.6
5.5
Mo
20
33
19
31
36
36
21
Ni
9,4
8.9
24
3.4
5.1
9.2
12
Se
7.5
13
24
6.2
•
13
13
19
Ag
29
25
56
46
47
19
27
Ti
19
12
55
29
30
20
50
V
12
58
7.5
7.3
5.5
32
18
Zn
9.1
6.7
7.6
15
11
2.5
16
Ba
11
10
8.7
6.4
8.0
20
11
Na
17
38
49
4.7
5.8
9.8
7.9
K
8.8
7.4
4.2
4.8
13
5.8
7.9
MEDIAN
PERCENT
10
11
19
6.8
8.0
9.8
11
RSD
-------
percent to 11 percent. Thus, the precision for the measurement of
the target elements in these complex solutions is very good.
The seventh spiked bulk digest, from coal fly ash, was spiked with
very high levels of molybdenum and vanadium (0.1 percent). The
median percent RSD’s for the determination of the 23 elements in
this spiked digest range from 4.2 percent to 83 percent with a
median of 19 percent (Table 3). The 12 percent median RSD for fly
ash digests without added Mo and V (Table 4) suggests that these two
elements decreased the measurement precision of many of the target
elements.
When Method 6010 and Method 3050 are applied in combination for the
determination of the 23 elements in spiked solids, the apparent
measurement precision decreases (Table 4) when compared to the
corresponding spiked bulk digest. The median percent RSD’s for the
7 solids across the 23 elements range from 10—19 percent. The
spiked solid samples were spiked prior to digestion to ensure that
the concentrations of the analytes were approximately 100 times
greater than the instrumental detection limits.
The median percent RSD’s for the same 7 solids, unspiked, range from
17—30 percent (Table 5). This poorer precision when compared to the
spiked solids results because over 50 percent of the reported
concentration values are less than 100 times the average of the
instrumental detection limits. In other words, as the
concentrations approach the instrumental detection limits the
precision decreases as indicated by the higher percent RSD values.
Four elements among those with the highest median percent RSD’s are
antimony, molybdenum, silver and thallium. For those elements that
were present in the digests of the unspiked solids at concentrations
100 times greater than the IDL’s (due to their occurrence in high
concentrations in the unspiked solids), the precision Is comparable
to the precision for the spiked solid samples.
The Method 6010 variance and the Method 3050 variance can be
calculated from the data base resulting from the analyses of the
spiked bulk digests and the spiked solid samples (Table 6). A
statistical analysis of the data shows that in general, the
digestion procedure and the inductively coupled plasma atomic
emission spectroscopic analytical procedure contribute about equally
to the overall measurement uncertainty or precision (variance) for
the determinations of the 23 target elements in the 7 homogeneous
solids.
The accuracy of the determination of the 23 target elements by
Method 6010, estimated from the analyses of the quality control
solutions, is shown to be excellent, within 99± 3 percent of the
“true” values with the exception of silver. Similarly, the accuracy
of Method 6010 in combination with Method 3050 was also found to be
5-18
-------
TABLE 4. PERCENT RSD’s FOR THE DETERMINATION OF THE 23 TARGET ELEMENTS
IN THE SPIKED SOLIDS
ELEMENTS
HAZARDOU
WASTE
S
1
RIVER
SEDIMENT
FLY
ASH
ESTUARINE
SEDIMENT
INDUSTRIAL
SLUDGE
ELECTRO—
PLATING
SLUDGE
MINE
TAILING
Al
17
24
20
22
14
18
26
Sb
27
56
25
62
28
40
59
As
13
27
16
22
19
20
22
Be
16
13
7.6
8.6
18
7.0
16
Cd
13
8.4
9.5
13
20
18
19
Ca
7.3
9.0
12
10
13
14
12
Cr
7.9
22
9.7
7.1
18
10
26
Co
18
22
11
9.4
17
13
18
Cu
13
14
11
9.7
19
9.1
12
Fe
14
19
44
12
18
15
18
Pb
15
6.4
9.6
11
20
19
5.8
Mg
5.9
8.5
17
9.0
16
10
10
Mn
14
9.1
11
10
17
19
9.4
Mo
19
31
23
18
18
43
20
Ni
13
20
9.8
10
20
16
17
Se
13
15
10
10
15
18
13
Ag
19
31
50
51
46
52
49
Ti
15
30
40
29
28
39
45
V
18
19
12
10
17
41
24
Zn
14
12
11
13
20
8.2
19
Ba
8.4
9.7
7.2
10
16
30
7.2
Na
14
39
25
9.4
22
15
12
K
19
17
17
15
22
5.7
15
ME DI AN
PERCENT
14
19
12
10
18
18
18
RSD
-------
TABLE 5. PERCENT RSD’S FOR THE DETERMINATION OF THE 23 TARGET ELEMENTS
IN THE UNSPIKED SOLIDS
(in
N)
ELEMENTS
HAZARDOUS
WASTE
1.
RIVER
SEDIMENT
FLY
ASH
ESTUARINE
SEDIMENT
INDUSTRIAL
SLUDGE
ELECTRO-
PLATING
SLUDGE
MINE
TAILING
Al
19
32
19
23
15
23
17
Sb
38
78
0
0
47
68
57
As
53
48
32
74
83
44
28
Be
31
27
27
35
42
70
41
Cd
37
17
57
51
17
22
59
Ca
8.0
13
10
11
10
17
9.0
Cr
10
19
28
23
12
12
90
Co
35
60
23
12
21
47
30
Cu
24
9.0
16
17
17
12
20
Fe
13
24
52
10
15
12
18
Pb
8.0
12
33
37
16
17
17
Mg
6.0
11
20
10
17
14
9.0
Mn
9.0
17
20
10
18
21
11
Mo
30
42
20
58
57
49
26
Ni
14
39
34
21
16
20
40
Se
42
61
0
30
.
43
74
77
Ag
41
43
49
0
37
54
60
Ti.
31
30
0
0
38
45
130
V
21
72
15
17
28
35
47
Zn
14
12
20
9.0
12
9.0
20
Ba
7.0
11
4.1
14
23
38
9.0
Na
52
52
34
9.0
16
17
13
K
23
34
20
17
32
19
24
MEDIAN
PERCENT
23
30
20
17
18
22
26
RSD
-------
TABLE 6. ESTIMATED PERCENTAGE CONTRIBUTIONS OF METHOD 6010 ICP
VARIANCE AND METHOD 3050 DIGESTION VARIANCE TO TOTAL VARIANCE
Elements 6010 ICP 3050 Digestion
Al. 42 58
Cd 47 53
Ca 39 61
Co 52 48
Cu 87 13
Fe 13 87
Pb 57 43
Mg 83 17
Mn 57 43
Mo 95 5
Ni 30 70
Se 100 0
Ti 85 15
V 18 82
Zn 64 36
Ba 28 72
K 19 81
As 1 100 0
Be 1 25 75
Ag 1 17 83
Sb 2 2 98
Cr 2 14 86
Na 2 21 79
Median 3
52
48
1.
Method 6010
variance
estimates
pooled
over 6
of 7
spiked
bulk digests
2.
Method 6010
variance
is median
over 7
spiked
bulk
digest
variances.
3.
Medians based on first 17 elements listed above.
5-21
-------
excellent, within 91±3 percent of the “true” values with the
exception of silver. Silver showed poorer accuracies of 82 percent
and 53 percent respectively. The accuracies are for these solutions
that contain the analytes in concentrations of approxImately 100
times the IDL’s.
The accuracy of the ICP Method 6010 can be estimated for complex
matrices by comparing the average concentrations of the target
elements in the spiked bulk digest (as determined by Method 6010) to
the corresponding concentrations which were determined by AAS by one
of the participating laboratories. A null hypothesis approach that
is based on the mean and on the corresponding standard deviation was
used to determine if the ICP—AES and AAS values are significantly
different at the 95 percent confidence level. The results Indicate
that only two out of 184 elemental measurements by the two methods
are signficantly different. In some cases where the ICPIAAS ratios
are very different (less than 0.75 or greater than 1.25), the
standard deviations in the ICP measurements are very high, and
therefore, the differences in the means are not significant.
Overall, the agreement between ICP and AAS is excellent.
A comparison of the results obtained by using Method 3050 with
Method 6010 to analyze the EPA—CLP solid reference material,
mine—tailing powder, for which the “true” values were obtained
previously by multilaboratory ICP—AES and AAS analyses, gives an
indication of the combined method accuracy for a solid matrix. The
comparsion (Table 7) shows that 21 of the 23 elements are within the
control limits for the reference material set by the EPA, while the
results on the other two elements (selenium and thallim which are
near the LOD) are outside of the limits. It should be noted that in
the EPA—CLP, these two elements are normally determined by graphite
furnace atomic absorption spectrometry, which is more accuracte than
ICP—AES for the determination of these elements at low levels.
These results indicate that Method 6010 can achieve good accuracy In
complex matrices for the analytes studied.
The method of standard additions was reauired for less than 10
percent of the total analyses. Results by ICP—AES using the method
of standard additions were compared with non—MSA data and atomic
absorption spectrometry results for the spiked bulk digest samples.
The comparison of this limited data set (Table 8) indicates that on
the average there is no significant Improvement In the data quality
when the method of standard additions Is used with Method 6010 for
the analysis of the solid matrices that were used in this study.
A comparison between data obtained on simultaneous and sequential
inductively coupled plasma spectrometers Indicated that the data
were statistically indistinguishable.
5-22
-------
TABLE 7. RESULTS FOR THE ANALYSIS OF THE UNSPIKED MINE TAILING SAMPLE
COMPARED TO THE EPA CLP CONTROL LIMITS*
DUPLICATE
ELEMENT
No.1
MEAN
STD
DEY
RSD
(%)
CLP
TRUE
CONTROL
LOWER
LIMITS
UPPER
MEAN
TRUE
WITHIN
LIMITS
Al
13900
1580
11
15200
7500
22900
0.91
YES
Sb
12
19
158
<20
0
44
YES
As
618
161
26
680
380
980
0.91
YES
Be
0.5
0.2
49
<1
0
1.6
YES
Cd
1.7
1.4
83
<1
0
7.8
YES
Ca
9850
718
7
10520
7850
13200
0.94
YES
Cr
12
12
104
17
0
46
0.71
YES
Co
7.2
2.1
29
6.9
0
19
1.05
YES
Cu
215
63
30
265
220
310
0.81
NO
Fe
10300
1580
15
11200
5910
16500
0.92
YES
Pb
5660
1050
19
5830
4310
7340
0.97
YES
Mg
14200
1040
7
14730
10910
18560
0.96
YES
Mn
Moa
92800
56
926
16
1
29
91735
68600
114900
1.01
YES
Ni
seb
21
43
8.9
51
43
117
22
<1
5.0
0
39
11
0.93
YES
NO
Ag
Tic
8.0
73
6.8
107
85
146
<2
3.8
0
0
26
9.1
19.23
NO
NO
V
13
7.2
54
19
0
46
0.72
YES
Zn
362
71
19
425
317
535
0.85
YES
Ba
397
43
11
430
360
510
0.92
YES
Na
3390
374
11
K
8130
2620
32
8150
4540
11770
1.00
YES
“
Loncentratlons in mg/Kg.
a Mo Is not a CLP element.
b The average LOD for Se is
C The average LOD for Ti Is
( n
N)
C A)
10 mg/kg.
9 mg/kg.
-------
( 1
TABLE
7.
(concluded)
DUPLICATE
No.2
STO
RSD
CIP
CONTROL
LIMITS
MEAN
WITHIN
ELEMENT
MEAN
DEY
(%)
TRUE
LOWER
UPPER
TKU
LIMITS
Al
14300
3170
22
15200
7500
22900
0.94
YES
Sb
5.8
4.2
74
<20
0
44
YES
As
636
185
29
680
380
980
0.94
YES
Be
0.5
0.2
39
<1
0
1.6
YES
Cd
1.6
1.4
87
<1
0
7.8
YES
Ca
9670
962
10
10520
7850
13200
0.92
YES
Cr
15
15
100
17
0
46
0.89
YES
Co
7.8
2.6
33
6.9
0
19
1.13
YES
Cu
235
25
11
265
220
310
0.89
YES
Fe
9810
2070
21
11200
5910
16500
0.88
YES
Pb
5550
846
15
5830
4310
7340
0.95
YES
Mg
14300
1540
11
14730
10910
18560
0.97
YES
Mn
Moa
92300
55
11000
14
12
26
91735
68600
114900
1.01
YES
NI
seb
21
43
8
49
37
112
22
<1
5.0
0
39
11
0.93
YES
NO
Ag
11 c
7.5
74
6.3
105
84
141
<2
3.8
0
0
26
9.1
19.54
YES
NO
V
13
8.1
62
19
0
46
0.71
YES
Zn
358
71
20
425
317
535
0.84
YES
Ba
414
27
7
430
360
510
0.96
YES
Na
3340
477
14
K
7350
1112
15
8150
4540
11770
0.90
YES
-------
TABLE 8. COMPARISON OF NSA AND NOP1-MSA RESULTSa
SAMPLE NAME
ELEMENT
N
NON-
MEAN
CONC)’
MSA
SD
N
MSA
MEAN
CONC)’
SD
%RATIO
SIG.
DIF.c
HAZARDOUS WASTE.
Cd
5
894
117
3
940
84
95
NO
HAZARDOUS WASTE
Ti
5
4410
788
3
4510
1130
98
NO
HAZARDOUS WASTE
Zn
5
4310
426
3
4560
250
95
NO
RIVER SEDIMENT
Ti
7
3160
2210
3
5050
675
63
NO
FLY ASH
Cd
5
754
422
3
897
219
84
NO
FLY ASH
Cr
5
1480
885
3
2390
1090
62
NO
FLY ASH
Pb
4
4100
634
4
6770
3300
61
NO
FLY ASH
Mn
4
1910
233
3
1750
304
109
NO
FLY ASH
Ni
3
1530
154
4
1350
500
113
NO
FLY ASH
Ti
4
5530
3730
3
1950
2470
284
NO
ESTUARINE SEDIMENT
Ti
5
3870
1290
3
3340
2850
116
NO
INDUSTRIAL SLUDGE
Ti
5
4470
872
3
4620
2230
97
P lO
ELECTROPLATING SLUDGE
Ti
3
4600
740
4
5350
1120
86
NO
MINE TAILING
Cd
5
850
69
3
985
112
86
P40
a -
- -
t
hree or
uniy etements tnat requirea tne application 01 tne M R D
more laboratories are included as statistically significant.
b Concentration for liquids in ugh; concentration for solids in mg/kg.
C Result of a null hypothesis approach used to indicate whether MSA and non-MSA
results are significantly different.
N - Number of cases.
% Ratio - non-MSA to MSA mean concentrations.
SPIKED BULK DIGESTS
c:i
(ii
(continued)
-------
TABLE 8. (continued)
UNSPIKED SOLIDS
NON-MSA
NSA
SAMPLE NAME
ELEMENT
N
MEAN
C0NC) SD
N
MEAN
CONC)’
SD
%RATIO
5 1G.
DIF.C
HAZARDOUS WASTE.
Be
4
0.8 0.1
3
0.7
0.2
93
NO
HAZARDOUS WASTE
HAZARDOUS WASTE
HAZARDOUS WASTE COUP.)
Cr
Co
Ni
6
6
5
95 8.4
8.0 2.4
17 1.3
3
3
4
111
9.1
13
10
1.5
8.9
86
88
128
YES
NO
NO
RIVER SEDIMENT
Sb
6
325 266
3
169
246
192
NO
RIVER SEDIMENT
Cd
6
11 2.5
3
11
3.5
103
NO
RIVER SEDIMENT
Co
5
21 16
4
21
19
99
NO
RIVER SEDIMENT
Ni
6
44 20
3
27
7.0
161
NO
RIVER SEDIMENT (DUP.)
Cd
6
10 1.6
3
10
0.7
107
NO
RiVER SEDIMENT (DUP.)
NI
6
39 13
3
38
19
105
NO
FLY ASH
Be
6
3.0 0.8
3
2.6
1.2
114
NO
MINE TAILING
Cd
4
2.3 1.6
3
1.9
1.1
122
NO
MINE TAILING
Zn
6
372 44
3
340
119
109
NO
MINE TAILING (DUP.)
Cd
4
2.4 1.6
3
1.5
0.8
158
NO
MINE TAILING (DUP.)
Co
6
7.3 2.5
3
8.8
3.1
83
NO
MINE TAILING COUP.)
Ni
5
21 5.6
4
21
11
100
NO
MINE TAILING COUP.)
Zn
6
365 43
3
345
122
106
NO
ELECTROPLATING SLUDGE
Cd
6
113 24
3
96
41
118
NO
ELECTROPLATING SLUDGE
Mn
6
226 31
3
254
126
89
NO
ELECTROPLATING SLUDGE
COUP.) As
6
33 20
3
41
20
80
NO
ELECTROPLATING SLUDGE
(DUP.) Mo
5
14 11
3
21
7.3
68
NO
INDUSTRIAL SLUDGE
As
4
11 6.6
3
26
11
41
YES
(Conti nueclj
-------
TABLE 8. (concluded)
SPIKED SOLIDS
NON-MSA MSA
MEAN MEAN SIG.
SAMPLE NAME ELEMENT N CONC) SD N CONC)’ SD %RATIO DIF.C
HAZARDOUS WASTE Co 6 45 8.2 3 30 2.2 149 YES
HAZARDOUS WASTE Pb 6 340 104 3 238 14 143 NO
HAZARDOUS WASTE Mo 6 39 20 3 29 2.8 134 NO
HAZARDOUS WASTE Ni 6 57 10 3 37 2.9 152 YES
HAZARDOUS WASTE (DUP.) Co 6 48 4.8 3 56 11 85 NO
HAZARDOUS WASTE (DLJP.) Pb 6 390 29 3 338 112 115 NO
HAZARDOUS WASTE (DUP.) Ni 6 61 3.5 3 58 14 106 NO
ESTUARINE SEDIMENT Cd 6 46 4.7 3 53 2.2 87 NO
ESTUARINE SEDIMENT Mo 6 37 19 3 47 2.5 79 NO
ESTUARINE SEDIMENT Ni 6 65 6.7 3 73 1.3 89 NO
ESTUARINE SEDIMENT 11 6 180 65 3 239 24 75 NO
ESTUARINE SEDIMENT (DUP.) Ni 6 63 6.9 3 74 3.3 86 YES
MINE TAILING Ni 6 64 7.9 3 60 15 108 NO
MINE TAILING (DUP.) Ni 6 63 6.9 3 64 19 99 NO
ELECTROPLATING SLUDGE (DUPI) Ti 6 160 46 3 304 104 53 YES
-------
RECOMMENDATIONS
The experimental design used in this multilaboratory study has
resulted in several excellent sets of multidimensional analytical
data that deserve consideration beyond the intended scope of this
report. Further analysis and interpretation of this data base is
suggested.
The presence of high concentrations (0.1 percent) of added vanadium
and molybdenum in the fly ash spiked bulk digest could account for
the apparent decrease in the precision of Method 6010 for the
determination of many of the 23 target elements In this matrix
compared to the 6 other solid digests. The interfering effects in
this matrix should be studied further.
The poor precision, accuracy, and spike recoveries for silver
demonstrated in this study, should be noted in both Method 3050 and
Method 6010. The possibility of precipitation in the
nitric/hydrochloric acid digestion matrix as well as
phototransformation should be discussed in Method 3050.
The poor spike recovery of antimony, observed in this study, should
be noted in Method 3050. In particular, the possibility of the
formation of oxide and oxo—chioride precipitates of antimony in the
nitric/hydrochloric acid digestion matrix should be discussed.
The application of the method of standard additions (MSA), a
conditional requirement of Method 6010, affects the economics, the
turnaround time of analysis, the practicality of the Method, as well
as the data quality. Although this report indicates that, on the
average, NSA data was not significantly different from non—NSA data,
(based on a selected set of statistically significant data), the
requirement for the application of the MSA should be investigated
further.
When soil—containing matrices are being analyzed by Method 6010, the
authors are of the opinion that the method of standard additions
should not be required for those elements that are endogenous to
soils in high concentrations. The high— concentration endogenous
elements in soils Include Al, Ca, Fe, Mg, K, and Na.
Required detection limits should be included in Method 6010 to
ensure uniform limits of detection. Recommended required detection
limits, based on this interlaboratory study, are presented in Table
1 of this report.
Methods 6010 and 3050 should be reviewed carefully for technical
completeness and comprehensiveness in view of the findings and
experience obtained in this multilaboratory evaluation.
5-28
-------
PRELIMINARY STUDIES OF THE SEPARATION AND DETERMINATION OF
CR (VI) AND CR (III) IN WATER BY
SOLID ADSORBENT E RACTION AND GFAA. ANALYSIS
R.A. Stockton, S.R.C. Priebe, E.L.S. McClendon, N.J. Friederich,
Midwest Research Institute, Kansas City, Missouri
INTRODUCTION
The importance of the speciation of Cr(III) and Cr(VI) has been widely
recognized. Cr(III) is an essential trace element in mammalian
systems. However, due to the toxic effects of Cr(VI) it is considered
an environmental and industrial hazard. The Environmental Protection
Agency (EPA) has supported the development of three ari lytical methods
that are approved for the determination of Cr(VI). The approved
methods for the analysis of Cr(VI) are based on coprecipitation
(Method 7195), colorimetric techniques (Method 7196), and
chelation/extraction (Method 7197).
Unfortunately, Cr(VI) is a relatively labile species and will convert
to Cr(III) under normal preservation procedures before analysis can be
accomplished. Cr(VI) lability is emphasized by the EPA SW—846 methods
which require a holding time limitation of only 24 hours prior to
analysis. This holding time is extremely difficult to meet under
normal circumstances of sampling, shipping, and analysis.
An additional analytical approach based on paired ion chromatography
has been used to reliably separate 3 a d quantify Cr( III) and Cr(VI)
using element specific detection. ‘ ‘ These methods are based on
separation of the Cr species in a reverse phase HPLC eluent. The
counter ion modifies the surface of a reverse phase HPLC column such
that the retention and subsequent separation of Cr(III) and Cr(VI) is
possible.
The application of the reverse phase adsorbent combined with the
paired ion chromatography phenomena may provide an alternative to the
collection and separation of chromium species while extending the
sample holding time. This paper will present a preliminary procedure
of rapidly separating the Cr(VI) from Cr(III) in aqueous samples
during sample collection. The isolation procedure for Cr(VI) and
Cr(III) requires comparable time to existing methods and may be
performed by field sampling personnel without specialized training.
EXPERIMENTAL
A PRP—l (styrene divinylbenzene co—polymer, Ailtech Associates) sample
cleanup cartridge is used to retain the chromium compounds. The
sample is prepared by the addition of tetrabutylammonium hydroxide
(TBAOH) solution to a known volume of 4 filtered aqueous sample so that
the TBAOH concentration is 5 x 10 molar. The PRP—1 sorbent is
prepared by wetting the sorbent with 2 mL of methanol. The methanol
rinse is followed by 1 niL of sodium hydroxide solution (pH 11.5).
The sample/rBAOH mixture is then passed through the PRP—1 reverse
5—29
-------
phase sample cleanup cartridge. The Cr(VI) is retained by the paired
ion on the PBP—1 column. The TBAOH/sample solution has a pH of 10
which forms the insoluble Cr(III) hydroxide. This precipitate is
filtered from the sample by the column. The Cr(VI) is selectively
removed from the column by eluting with two milliliters of sodium
hydroxide (pH 11.5). This procedure also allows preconcentration
since the Cr(III) and Cr(VI) are totally retained on the column.
After eluting the Cr(VI) from the column with sodium hydroxide, the
total chromium determined in the column eluent will represent Cr(VI)
only. Cr(tII) can be subsequently removed with 2 mL of 10% (v/v)
nitric acid and determined as total chromium so that a mass balance
may be performed. The sodium hydroxide/nitric acid elution sequence
must be followed or Cr(VI) will contaminate the Cr(III). The eluents
may be preserved and transported as aqueous samples because the
chromium species is no longer a consideration, since only total
chromium determination is required for the final analysis.
RESULTS
Six ion exchange resins, one C silica based, and one styrene
divinylbenzene based resins were a1uated in this study. With each
adsorbent various rinse, sample retention, and sample elution
solutions were used.
The PRP—1 resin provided the best retention and percent recovery of
the resins evaluated. When the procedure as described above was
followed, 92.1—102% of the CrP1I) was recovered. In addition,
92.0—114% of the Cr(III) was recovered. Table 1 provides a svn’nary of
the retention, elution, and recovery characteristics of Cr(III) and
Cr(VI) using the PRP—1 resin.
Instrumental, detection limits for chromium is 12 nqjmL. The percent
relative standard deviation in 7.3%. These results are determined
from 14 determinations of 5 ng4nL standard. Method detection limits
have not been determined. However, estimated detection limits from
the percent recovery should allow similar detection limits. Trace
enrichment may also be used to realistically enhance the detection
limits by two orders of magnitude.
The results obtained in this paper are from laboratory water
standards. As such, the result may not represent what occurs in a
field sampling situation. Therefore, to validate this procedure for
environmental samples, a wide range of environmental samples should be
spiked and analyzed by current EPA approved methods and this
procedure. From this spiking data comparison the applicability of
this procedure may be determined.
REFER ES
1. Environmental Protection Agency Manual SW—486, “Test Methods for
Evaluating Solid Waste,” 3rd edition, 1986.
2. A. syty, R. G. christensen, T. C. Rains, At. Spectrosc. 7(4) , p.
5-30
-------
89, 1986.
3. F. E. Brinckmann, K. L. Jewett, W. P. Iverson, K. J. Irgolic, 3. C.
Enrhardt, R. A. Stockton, 3. Chromatogr . 191, p. 31, 1980.
4. R. A. Stockton, N. J. Friederich, “Proceedings of the International
Symposium on Metal Speciation, Separation, and Recovery,” 1166—1167,
Chicago, Illinois (1986).
5—31
-------
TABLE 1
RETENTION, ELUTION, AND RECOVERY CHARACTERISTICS
OF Cr(JII) ANt) Cr(IV) USING PRP1
Rinse Retention % Retention Eluent % Elution % Recovery
Solution Solution Cr(III) Cr(VI) Cr(III) Cr(VIj Cr(IiI) Cr(VJ) (III) Cr(VI )
a b p}(IO 92 98 c a 100 94 92.0 92.1
d b pH lO 99 99 c e 93 88 921 87.1
f b pHlO 96 99 c e 102 85 97.9 84.2
a b pHlO 99 99 c a 115 103 114 102
a NaOH TO pHll.5 (<0.003 N)
b 0.0005 N TBAOH
c 10% }1N0 3
d Deionized water
e Hethanol
f 0.0001 N NaOH
-------
EVALUATION OF SW—846 COLD—VAPOR MERCURY METHODS 7470 AND 7471
Werner F. Beckert, Environmental Monitoring Systems Laboratory, U.S.
Environmental Protection Agency, Las Vegas, Nevada; Jerry D.
Messman, Mark E. Churchwell, Robert L. Livingston, Donald L. Sgontz
and Gordon F. Wallace, Battelle Columbus Division, Columbus, Ohio
ABSTRACT
The protocols for Methods 7470 and 7471 in the SW—846 methods manual
are designed for the cold—vapor atomic absorption spectrometric
(CV—AAS) determination of total mercury in aqueous (extracts,
wastewater, ground water, etc.) and solid (soils, sediments,
sludges, etc.) materials, respectively. Aqueous samples are
digested with a combination of nitric and sulfuric acid plus
permanganate and persulfate solutions at elevated temperature; solid
samples are heated either in a water bath with aqua regia and
permanganate, or in an autoclave with sulfuric acid, nitric acid and
perman,ganate. The mercuric ions in the digests are then reduced
with hydroxylamine, and the elemental mercury is aerated into the
AAS cell, either in a closed (recirculating) or an open system,
where the absorption at 253.7 nm is determined.
The topic of this presentation are the results of a performance
evaluation study of these methods conducted without and with changes
by the Battelle Columbus Division Laboratory. The methods were
evaluated using aqueous and solid environmental samples of
homogeneous and known compositions in order to assess the accuracy
and precision of the methods without introducing uncertainties due
to sample inhomogeneities. The methods, as originally written, were
satisfactory for the analyses of samples containing relatively high
concentrations of mercury. However, the method quantification
limits for the closed and open systems were inadequate for the
determination of mercury in environmental samples such as ground-
water samples, when low concentrations were encountered. Spectral
interferences which are caused by nonspecific absorption of the
analytical radiation degraded the accuracy of the recirculating
CV—AAS iethod.
The protocols in Methods 7470 and 7471 were modified to improve
mercury detectability and to minimize the additive effect of
nonspecific background absorption. The use of an amalgamation
CV—AAS system operated in an open configuration resulted in an
improvement of one order of magnitude in instrumental detection
limit for mercury over the recirculating method. The results of the
analyses of four reference sediment materials and a simulated
aqueous waste sample by the amalgamation CV—AAS method indicated
acceptable accuracy and precision for most of the environmental
samples tested. As expected, the amalgamation CV—AAS method
revealed no spectral interferences resulting form nonspecific
absorption of the analytical radiation by organic vapors.
5—33
-------
The modified protocols are at present the subject of a multi—
laboratory evaluation with the Battelle Columbus Division laboratory
as the coordinating laboratory. Seven laboratories are
participating in this study.
INTRODUCTION
SW—846 Method 7470, entitled “Mercury (Manual Cold—Vapor Technique)”
is designed for the determination of total mercury in
mobility—procedure extracts, aqueous wastes, ground water, and other
aqueous samples. The samples are digested on a steam bath with
sulfuric acid and nitric acid, followed by the addition of
permanganate to oxidize suif ides and of persuif ate to oxidize
organ.ics. SW—846 Method 7471, entitled “Mercury in Solid or
Semisolid Waste (Manual Cold—Vapor Technique)” is designed for the
determination of total mercury in soils, sediments, bottom deposits,
and sludge—type materials. The samples are digested on a steam bath
with aqua regia; any sulf ides remaining are oxidized with
perman,ganate. An alternative sample preparation procedure for
Method 7471 (which will be referred to as Method 7471A) specifies
digestion with a mixture of sulfuric acid, nitric acid, and
permanganate under heat and pressure in an autoclave. In all three
methods, excess permanganate, persulfate and any free chlorine
present in the digests are reduced with hydroxylamine as the
hydrochloride or sulfate, and mercuric ions are reduced with
stannous sulfate or chloride. The elemental mercury is then aerated
from the digests through the absorption cell of an atomic absorption
spectrophotometer and is continuously recirculated in a closed
system until a steady—state absorption signal is attained for
quantification at 253.7 nm. This technique is based largely on the
system, where the mercury vapor is passed through the absorption
cell only once, may be used instead of the closed system. Potential
interferences listed in the methods are certain unspecified volatile
organics, copper, excess chloride, and sulfide.
PHASE 1 — SINGLE—LABORATORY EVALUATION
The objective of the Phase 1 effort was to conduct a single—
laboratory evaluation of the current EPA protocols for SW—846
Methods 7470 and 7471 and to determine whether or not any revisions
of the current protocols would be warranted to improve analytical
performance. The methods were evaluated using aqueous and solid
environmental samples of homogeneous and known compositions in order
to assess accuracies and precisions of the methods without
Introducing uncertainties such as those which would be due to sample
inhomogeneitles. The research was conducted in three phases: (1)
evaluation of the EPA protocols, as currently written, (2)
Implementation and evaluation of selected modifications in the
current EPA protocols, and (3) testIng of the proposed revisions of
5-34
-------
the EPA protocols on representative environmental samples. This
single—laboratory evaluation was conducted by, and at, the Battelle
Columbus Laboratory.
INSTRUMENTATION AND EQUIPMENT
A Model 603 atomic absorption spectrophotometer (The Perkin—Elmer
Corporation) was used for all analytical measurements. The
cylindrical glass absorption cell with quartz end windows was
mounted on a conventional burner head for convenient positioning and
for two—dimensional alignment in the optical path of the
instrument. A portable lamp was used to heat the absorption cell to
slightly above ambient temperature, thus preventing condensation of
water vapor on the end windows. The recirculating CV—AAS system
(Figure 1) required in Methods 7470 and 7471 was assembled from
components and materials as specified in the current protocols. The
recirculating system was configured, as necessary, for evaluation of
the open (“once—through”) CV—AAS system referenced in the current
protocols.
All glassware was washed with hot detergent solution, rinsed with
deionized water, and filled with a 1:4 (v/v) reagent—grade nitric
acid/deionized water mixture. After standing for at least 4 hours
at ambient temperature, the acid solution was removed and the
ion Cell
Bubbler
Sample Solution
in BOO Bottle
Scrubber
Containing
a Mercury-
Absorbing
Medium
Figure 1. Apparatus for the recirculating CV-AAS method.
Air Pump
5-35
-------
glassware rinsed extensively with deionized water. Watch—glass
covers and plastic pipet tips were soaked in beakers with 1:4 nitric
acid/deionized water, and were then rinsed extensively with
deionized water.
REAGENTS AND STANDARDS
All chemicals were of ACS grade or better. Approximately 5 ug of
mercury were observed in the digested reagent blanks throughout
this study. Inorganic—mercury working standards were prepared fresh
daily by serial dilutions with deionized water/nitric acid
(100:0.15, vlv) from a commercial l000—ing/L mercury stock solution
(Fisher Scientific Company). Methyl mercuric chloride standards
were prepared by diluting, with deionized water, a stock solution
previously prepared from the solid material (Pfalz and Bauer), and
diphenyl mercury standards were prepared by diluting, with
chloroform, a stock solution previously prepared from the solid
material (Aldrich Qiemical Company).
SAMPLES
One simulated aqueous sample and four solid homogeneous reference
samples of known compositions were used to evaluate the methods.
The aqueous waste sample was prepared by decanting and combining the
aqueous portions from three sludge—type waste samples provided by
the EP QA Materials Bank. The pooled aqueous sample was filtered,
spiked with 200 ug of mercury as mercuric chloride, acidified with
nitric acid to pH 2, and diluted with tap water to 2 liters. The
endogenous mercury and the spike added up to a final mercury
concentration of approximately 110 ug/L for the simulated aqueous
waste.
The solid reference materials were two Standard Reference Materials
from the National Bureau of Standards — River Sediment NBS—SRM 1645,
1.1 ug HgIg, and Estuarine Sediment NBS—SRN 1646, 0.063 ug Hg/g —
and two reference materials from the National Research Council of
Canada, Ottawa, Ontario, Canada — Marine Sediment NRCC BCSS—1, 0.129
ug Hg/g.
EXPiltIMENTAL PROCEDURES
The samples were digested according to the procedures described In
the current protocols. However, all sample digestions following
Methods 7470 and 7471 were performed In 125—niL Erleninayer flasks
rather than in the specified BOD bottles. Sample calibration
digests could then eaaiiy be transferred to volumetric flasks and
diluted to calibrated volume for subsequent analysis. The reason
for this change was the concern that sample heterogeneity could
become a major problem in the analysis of samples with a relatively
high mercury content, and that therefore the digestion of small
5-36
-------
sample sizes as a means of adjusting the final mercury concentration
within the linear dynamic range of the method would be less
desirable than the dilution of relatively concentrated sample
digests. The NBS Standard Reference Materials, in recognition of
the dependence of homogeneity on sample size, are certified only for
potions above a specified weight.
To samples subjected to the autoclave digestion in Method 7471A,
5—mL portions of deionized water were added before the addition of
the acids. This reduced the possibility of losses form the samples
caused by spattering during the addition of the concentrated acids.
To all sample digests, hydroxylamine reagent was added while the
digests were still in the Erlenxneyer flasks in order to reduce
excess permanganate and manganese dioxide particles adhering to the
flask walls. Only then were the digests transferred to the 100—niL
volumetric flasks or directly to the reduction—aeration vessel, as
appropriate, for analysis. The digests or digest aliquots, after
transfer to the BOD bottle, were diluted with deionized water to 100
niL, stannous chloride was added to reduce the mercuric ions to the
element, and the BOD bottle was connected to the cold—vapor
generator apparatus for the mercury determination. The atomic
absorption was measured for 120 seconds; the steady—state absorbance
peak was reached after approximately 60 seconds.
RESULTS OF THE EVALUATION OF THE CURRENT PROTOCOLS
An evaluation of the current protocols with the minor changes
described above gave an Instrumental detection limit for the
recirculating method of 0.01 ug Hg, based on seven replicate
measurements, and of about 0.02 ug Hg for the open method. This
corresponds to 0.1 ug/mL and 0.2 ug/mL for aqueous samples, based on
100—niL sample portions. The linear dynamic range for both methods
extends from the instrumental detection limit to approximately 1.0
ug Hg. This represents 2 orders of magnitude of linear response.
The precision of the mercury measurements, expressed as percent
relative standard deviation, was approximately 1 percent for
duplicate analyses of the simulated aqueous waste samples, when
Method 7470 was used. When the duplicate set of samples was spiked
with mercuric chloride at 1, 2 and 5 times the endogenous mercury
levels, the spike recoveries ranged from 101 to 112 percent over the
spike range. The precision for triplicate measurements was 4.7
percent for digests of SRN—1645 samples prepared according to Method
7471 and 6.1 percent for Method 7471A, and the recoveries were 0.93
and 1.11 ug/g, respectively, as compared to the certified value of
1.1 ± 0.5 ug/g. When SRM 1646 was digested using Method 7471A,
however, the RSD was 26Z and the recovery was 0.18 ug/g, as compared
to the certified value of 0.063 ugfg. As can be seen from these
results, the recirculating CV—AAS method gives good precision and
accuracy for samples of high mercury concentrations. However, when
5—37
-------
the mercury levels are low as in SRM 1646, high imprecision and poor
accuracy result from inadequate detectability and possibly an
unidentified background interference.
In order to evaluate potential spectral interferences caused by
nonspecific background absorption of primary mercury radiation by
volatile organic compounds, benzene and methyl ethyl ketone (NEK)
were added to NBS—SRN 1645 digest prior to the reduction and
aeration step. The amounts of benzene (3 uL) and MEK (1 mL) were
the levels required to produce nonspecific background absorbances
equivalent to the mercury absorbance for a clean NRS—SRN 1656
digest. When analyzed, an average mercury concentration of 2.41
ug/g (RSD 6.6%) was found for triplicate measurements in the
presence of beazene, and 2.19 ug/g (RSD 2.9%) in the presence of
MEK. These data represent an approximate doubling of the apparent
mercury concentration when compared to the certified mercury value
for SRi’! 1645 of 1.1 ± 0.5 ug/g. It is anticipated that nonspecific
absorption caused by volatile organics may be automatically
compensated for by using background correction.
METHOD MODIFICATION AND RUGGEDNESS STUDY
The evaluation of the current protocols revealed that sample
preservation and digestion procedures could be used without
significant modifications. However, it also showed that the CV—AAS
apparatus specified in the protocols — which is based on the Hatch
and Ott design — does not represent a state—of—the—art approach when
compared to other modern CV—AAS designs described in the
literature. The recirculating CV—AAS system in the current
protocols was therefore changed to an open CV—AAS system which
included an on—line amalgamation/thermal desorption process between
the reduction—aeration step and the atomic absorption measurement.
This modification is based on an amalgamation method previously
described by Long et al. 2 for the determination of mercury in
ambient air. This change offers (1) enhanced detectability for
trace mercury determinations and (2) eliminatIon of nonspecific
absorption interferences which are due to volatile organics in the
sample.
A CV—AAS system using an amalgamation process for on—line mercury
preconceutration is not a unique approach but rather represents an
established and modern state-of—the—art method. In addition to
numerous research groups describing customized variations of mercury
preconceutration by amalgamation in the literature 26 , one major
manufacturer of atomic absorptioninstrumentationlncorporates an
amalgamation system for mercury preconcentration into a commercial
CV—AAS apparatus.
In the modified methods, the ambient air train and the BOD bottle
were replaced as a dedicated reduction/aeration vessel. The mercury
5-38
-------
aerated front the solution is now concentrated onto silver wool and,
after water vapor and other volatile species have been purged from
the system, thermally desorbed from the silver wool and swept into
the absorption cell of the AAS. A detailed schematic diagram of the
amalgamation CV—AAS apparatus is presented in Figure 2. The
replacement of the recirculating ambient air train with a
flow—through nitrogen purge eliminated the need for the air pump and
the mercury scrubber medium. The incorporation of the gas sparging
bottle improved the efficiency of the aeration of the mercury vapor
from the solution and thus reduced the time required for the
amalgamation step. The absorption step allowed purging of the gas
train between absorption and thermal desorption. This eliminated
any volatile organics and water vapor which could contribute to
nonspecific background absorption so that desiccants and automatic
Instrumental background correction were not required.
Charcoal
• Trap
Pt. 1 r9ing Cylinder
(Reduction-Aeration
Sample Cell)
Figure 2. Apparatus for the amalgamation CV-AAS method.
Needle Valve
y
Glass Stopcock
Absorption
Cell
Tygon to Glass
Connections
Nit gen\ Flow Meter
Cylinder
Silver Wool
Chrom-Alumel
Resistance Heating
Winding
5-39
-------
Several operating parameters of the amalgamation CV—AAS system were
studied and optimized. The nitrogen gas flow which affects the rate
of mercury aeration from the solution in the sparging bottle and, in
conjunction with the desorption temperature, determines the rate of
mercury desorption, was optimized at 450 mL/min. With this flow
rate, more than 90 percent of the mercury vapor of both high and low
mercury concentrations was released from the solution within two
minutes. A Varlac voltage setting of 20 volt output was found to be
optimal, rugged and safe at the 540 mL/inin nitrogen flow rate, and a
0.7—g portion of silver wool, with a total loading capacity of 3 to
4 ug mercury, was deemed to be adequate for mercury concentrations
analyzed within the linear dynamic range of this system.
RESULTS OF TilE AMALGAMATION METHOD EVALUATION
An instrumental detection limit of approximately 1 ng mercury was
determined for the amalgamation method, using the 3—sigma
criterion. This represents an approximately 10—fold improvement
over the recirculating CV—AAS system. The linear dynamic range of
the amalgamation system extends from the absolute instrumental
detection limit to 100 ng. The linear dynamic ranges of both the
recirculating CV—AAS method (10 to 1000 ng) and of the amalgamation
method span approximately 2 orders of magnitude of mercury
concentration. The linear range of the recirculating method is
shifted to higher mercury concentrations because of lower
sensitivity.
The higher sen8itivity of the amalgamation system will require
dilution of environmental sample digests containing relatively high
mercury concentrations. Multiple dilutions of digests were
therefore analyzed to examine whether or not such dilutions would
affect the accuracy and precision of the measurements. The results
of triplicate measurements on 5—fold and 10—fold dilutions of a
NBS—SRN 1645 digest indicate that dilutions of up to 10—fold have no
adverse effects on the accuracy and precision of the amalgamation
CV’—AAS measurements.
The addition of benzene (3 uL) and MEK (1 mL) to NBS—SRN 1645
digests prior to the reduction and aeration step did not increase
the apparent mercury recoveries, as was the case for the
recirculating method. The average mercury concentrations measured
by amalgamation CV—AAS in the sample digests spiked with the organic
solvents agreed within the uncertainty limits with the mercury
concentrations measured in the sample digests in the absence of
these volatile organics. Because background correction with a
deuterium arc lamp (or by the Zeeman effect) is not required in the
presence of volatile organics when the amalgamation method is used,
the spectrophotometer could be operated in the double—beam optical
configuration which minimized source—related noise.
5-40
-------
The sample digestion procedures described in Methods 7470 and 7471
were tested for mercury recovery from organomercurials. Portions of
the simulated aqueous waste and of NBS—SRM 1646 were spiked in
triplicate with phenyl mercury and methyl mercuric chloride,
respectively, and, for comparison, with mercuric chloride. The
spiked samples were digested according to Methods 7470, 7471 and
7471A, and the digests were analyzed by using the amalgamation
method. The results listed in Table 1 show that mercury was
completely recovered from the inorganic and organic mercury
compounds *
To determine potential interferences by sulfide, copper and chloride
ions, simulated aqueous waste samples were spiked with these
potential interferants, digested according to Method 7470, and
analyzed by using the amalgamation method. The results listed in
Table 2 show that none of these ions, at the concentrations used,
interferes significantly with the mercury determinations.
Table 1.
Mercury Spike Recoveries By Amalgamation CV—kAS Using
Digestion Procedures from Methods 7470, 7471 and 7471A
Average Percent Recovery
of Hg Spike
Methyl Mercuric
Sample
Mercuric
Diphenyl
(Digestion Procedure)
chloride
Mercury
chloride
Simulated Aqueous Wastea
(109)
125
113
(Method 7470)
NBS—SRN 1646 b
86
100
93
(Method 7471 Steam Bath)
NBS—SRN 1646 b
96
107
97
(Method 7471k Autoclave)
a 50-inL samples spiked prior to digestion with 4.7 ug Hg as
diphenyl mercury and 4.3 ug Hg as methyl mercuric chloride,
resp. The endogenous mercury content was 110 ug/L; no HgC1 2 was
added.
b 0.2—g samples spiked prior to digestion with 0.02 ug Hg as
mercuric chloride, 0.95 ug Hg as diphenyl mercury, and 0.86 ug
}Ig as methyl mercuric chloride, resp.
5-41
-------
Table 2.
Effects of Selected Inorganic Matrix Ions on Mercury
Determinations in Simulated Aqueous Waste
Digested by Method 7470 and Analyzed by
Amalgamation CV—AAS
Samplea Concentration
Measured, ug/L
Simulated Aqueous Waste 111 ± 3.0
Simulated Aqueous Waste 108 + 1.5
+ Sulfide (4Omg/L) —
Simulated Aqueous Waste 101 ± 0.6
+ Copper (20 mgIL)
Simulated Aqueous Waste 108 ± 2.5
+ Chloride (68 g/L NaC1)
a The eudogenous Hg concentration of the simulated aqueous waste is
approximately 110 ug/L.
The four solid reference materials listed earlier were digested in
triplicate according to the Method 7471 procedure, and were analyzed
for mercury by using the amalgamation CV—AAS system with aqueous
calibration standard. The digests were also analyzed by the method
of multiple standard additions; they were spiked with mercury as
mercuric chloride at lx, 2x, and 4x or 5x the endogenous mercury
concentrations in the respective sediment digests. From the results
presented in Table 3 it can be seen that when aqueous calibration
standards were used the mercury recoveries ranged from 91 percent to
119 percent for three of the materials; the relative precision of
the analyses ranged from approximately 1 percent to 7 percent. The
recovery for the NRCC BCSS—1 material was marginally high with 125
percent; the same material also showed a similar positive bias when
the method of multiple standard additions was used. The results of
the analyses using the method of multiple standard additions
revealed that multiplicative interferences were not present in the
CV—AAS measurement step for the determination of mercury in the
reference materials. The relative precislons of the analyses by the
method of multiple standard additions ranged from 4 percent to 14
percent.
5-42
-------
CONCLUSIONS
The steam bath and autoclave digestion procedures in Methods 7470
and 7471 were generally adequate for the determination of mercury in
the spiked and unspiked samples tested, including organornercurial
spikes. A potential problem is the reagent blank concentration
which was approximately 5 ng mercury per digest throughout the
study. The greater inherent sensitivity of the amalgamation CV—AAS
system over the recirculating and the open CV—AAS systems extends
the present instrumental detection limit to absolute mercury
concentrations of 1 n,g, provided that the reagent blank level is
sufficiently low. Volatile organic species and water vapor which
produce unspecific background absorption interferences with the
recirculating CV—AAS system can be eliminated when the amalgamation
CV—ASS system is used, and correction with a deuterium arc or Zeeman
background corrector is not required.
Table 3.
Mercury Recoveries From References Sediment Materials
Using Method 7471 Steam Bath Digestion and
Amalgamation CV—AAS
Concentration (ug/g)
Certified
Sample Reference Value
Standard
Method of Multiple
Calibration Standard Additions
NBS—SRN 1645 1.1 ±
0.5
1.9 ± 0.056 1.07 ± 0.15
NBS—SRN 1646 0.063
± 0.012
0.075 ± 0.0053 0.063 ± 0.0069
NRCC MESS —i 0.171
± 0.014
0.185 ± 0.0035 0.193 ± 0.0080
NRCC BCSS—i 0.129
± 0.012
0.161 ± 0.0020 0.165 ± 0.0089
The amalgamation CV—AAS method requires approximately five minutes
per sample for analysis, and longer times may be necessary for
samples requiring dilutions because of relatively high mercury
concentrations. However, the advantages of the amalgamation CV—AAS
methods more than outweigh this disadvantage.
5-43
-------
PHASE 2 - MULTI-LABORATORY EVALUATION
The protocols for Methods 7470 and 7471, as modified and tested in
iase 1 of this study, are being evaluated in a multi—laboratory
study. The Battelle Columbus Division Laboratory is acting as the
coordinating laboratory that Is responsible for the design of the
study and its execution, including selection of the participating
laboratories; sample procurement, analysis and shipping; sample
stability studies; standard preparation and shipping; preparation,
testing and shipping of the amalgamation component; instruction
preparation and data evaluation; and QA/QC.
Batteile Columbus Division staff members have assembled 23 sIlver
amalgamation cells and tested them for uniformity of response.
Calibration curves were constructed for each cell using a reagent
blank and standards containing 10, 50, and 100 ng of mercury. The
slopes of the calibration curves for these cells averaged 0.0055
absorbance/nanogram of mercury with a standard deviation among cells
of 0.0006, or a relative standard deviation of 11 percent. These
results Indicate good consistency of performance among amalgamation
ceUs with respect to amalgamation efficiency and thermal desorption
characteristics.
Samples for the study have been collected and characterized, as
necessary. These samples have been further evaluated for mercury
content, homogeneity, and recovery of predigestion spikes of methyl
mercuric chloride and mercuric chloride. The results indicate good
overall recoveries. In order to avoid potential problems with
low—level spikes added to samples containing chemicals which may
prematurely reduce mercury to the elemental state with resultant
loss of the spike, the samples have been shipped unspiked. Organic
and inorganic mercury standards to be used as predigestion spikes
have been sent to the participants with Instructions as to how much
spike to add.
Form letters were sent to 90 laboratories that were assumed to have
the necessary skiils, equipment, and interest required to
participate in this multi—laboratory evaluation study. Of these
laboratories, 22 responded with proposals which included equipment
descriptions, qualification statements of personnel, previous
relevant experience, and cost bids. Nineteen of these laboratories
received Battelle—tested mercury amalgamation cells, a l000—mg/L
inorganic mercury standard, a solid waste sample (NBS—SRN 1633a,
Coal Fly Ash), a water sample of low mercury content, standards for
spiking the samples, and instructions. The laboratories analyzed
the samples according to the instructions and reported the results
to the Battelle Columbus Division Laboratory. The final selection
of the 7 participants was based on the results of these analyses
and, of course, on the available funds.
5-44
-------
The selected laboratories received the following samples which will
be analyzed in duplicate by three laboratories and in triplicate by
four laboratories:
o Canadian sediment MESS—i
o Incinerator fly ash
o Incinerator fly ash spiked with an inorganic mercury compound
o Columbus municipal sewage sludge
o Columbus municipal sewage sludge spiked with an organic
mercury compound
o Ground water sample from a farm well
o Wastewater from an animal facility spiked with art inorganic
mercury compound
o ASTM Type II water spiked with an inorganic mercury compound.
The participating laboratories are at present analyzing these
samples. The results obtained from the laboratories will be
subjected to statistical evaluation designed to provide estimates of
the accuracy and the inter— and intralaboratory precision of Methods
7470 and 7471. In addition, the data will be examined to determine
the effects of mercury concentrations and matrix components on the
performance of these methods.
The results of the multi—laboratory evaluation study will be used to
make any necessary revisions to the Method 7470 and 7471 protocols.
It is anticipated that the revised protocols will be available by
the end of this fiscal year.
NOTICE
Although the research described in this article has been supported
by the United States Environmental Protection Agency under Contract
Number 68—03—3226 to the Battelle Memorial Institute, Battelle
Columbus Division, through the Acurex Corporation, it has not been
subjected to Agency review and therefore does not necessarily
reflect the views of the Agency, and no official endorsement should
be inferred. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
REFERENCES
Hatch, W. R., and W. L. Ott. Determination of Sub—Microgram
Quantities of Mercury by Atomic Absorption Spectrophotometry.
Anal. Chew. 40:2085—2087 (1968).
Long, S. J., D. R. Scott, and R. J. Thompson. Atomic Absorption
Determination of Elemental Mercury Collected from Ambient Air
on Silver Wool. Anal. Chem. 45(13):2227—2233 (1973).
5-45
-------
Murphy, J. Determination of Mercury in Coals by Peroxide Digestion
and Cold Vapor Atomic Absorption Spectrophotometry. At.
Absorpt. Newslett. 14(6):151—152 (1975).
Dogan, S., and W. Baerdi. Preconcentration on Silver Wool of
Volatile Organo— Mercury Compounds in Natural Waters and Air
and the Determination of Mercury by Flameless Atomic Absorption
Spectrometry. Intern. J. Environ. Anal. (hem. 5:157—162 (1978).
Welz, B., N. Neicher, II. W. Sinemus, and D. Maler. Picotrace
Determination of Mercury Using the Amalgamation Technique. At.
Spectroac. 5(2):37—42 (1984).
Welz, B, and K. Meicher. Improved Sensitivity for the
Determination of Lowest Mercury Concentrations Using Larger
Sample Volumes. At. Spectroac. 5(2):59—61 (1984).
5-46
-------
FACTORS AFFECTING EP TOXICITY METALS RESULTS
Andrea Jirka, Chief, Inorganic Laboratory Section, Marilyn Shannon,
Inorganic Chemist, John Morris, Metals Team Leader, Pankaj
Parikh, General Inorganics Team Leader, U.S. Environmental
Protection Agency, Region V, Central Regional Laboratory, Chicago,
Illinois
AR S TRACT
The EP TOX test, as described in SW—846 method 1310, can produce
highly variable results, especially for Pb. This variability is
more pronounced in some samples than in others. The magnitude of
the variability is greater than would be expected simply as a result
of sample inhomogeneity. This phenomenon can be a serious problem
since duplicate determinations can differ by orders of magnitude.
Thus it may be impossible to determine whether a waste exceeds
regulatory action levels.
Using RCRA Performance Sample VI as a model, CRL investigated
several possible causes for the problem. The investigation
indicated that the problem was not a function of the laboratory
water, the acid used for extraction, the type of extractor used, or
the sample size. The significant factor seemed to be the technique
that was used to adjust the pH. When the pH was adjusted using
small aliquots of acid with constant stirring, the results for Pb
were relatively low. If larger aliquots of acid were added without
constant stirring, the results were significantly higher. The
results were most dramatic when the pH of the sample was
inadvertently allowed to fall below 4.8. When this occurred, the
concentrations of Pb, Cr, Ni, and Zn were significantly higher (and
within the “acceptable” range for the PE study).
These observations indicate that the sample pH is a critical factor
affecting the EP TOX results. If the pH ever falls below a critical
value (presumably 4.8) in any portion of the sample, the results can
be significantly elevated.
Based upon these observations, a synthetic sample containing lead
nitrate was prepared. It was extracted under varying pH adjustment
conditions. The results of these experiments will be reported.
BACKGROUND
Under the Resource Conservation and Recovery Act (RCRA) there are
four characteristics attributable to a waste that can classify the
waste as hazardous. One of these characteristics is the extraction
procedure toxicity (EP TOX). The EP TOX method defines a leaching
procedure used to test waste materials. The procedure is intended
to simulate the natural leaching that would occur if the waste were
5-47
-------
discarded in a leaking landfill. The RCRA regulations set maximuni
allowable concentration levels for a group of organic chemicals and
metals that are extracted from the waste using the EP TOX
procedure. This study is concerned with factors which can affect
the concentration levels of metals that are extracted using the EP
TOX procedure.
In the EP TOX test a waste sample of known weight is separated by
filtration Into liquid and solid portions. The solid portion is
transferred to an extraction vessel. A dilute solution of acetic
acid is added to the waste until the pH is lowered to 5.0 ± 0.2.
The sample and acid are mixed for 24 hours using a stirring or
tumbling action. During the mixing procedure the pH of the
sample/acid slurry is checked frequently and acid Is added as
necessary to maintain the pH at 5.0 ± 0.2. After 24 hours of mixing
the solid and liquid are separated. The solid residue is
discarded. The volume of the liquid is adjusted and the liquid is
recombined with the liquid portion of the original sample. This
combined liquid is the EP extract. The EP extract is analyzed for
the EP TOX metals, As, Se, Pb, Ba, Cd, Cr, Hg, and Ag. Other metals
may also be of interest.
As laboratories began large scale testing of wastes for EP TOX
metals, CRL Reg. V observed an unexpectedly large degree of
variability in the results. This variability seemed more serious
for lead than for the other metals. Also, certain samples seemed
more sensitive to the variability than others. Table 1 describe8
lead results for a sample that was analyzed by three laboratories.
Laboratory #1, whIch represented the waste facility produced results
which showed the facility to be In compliance with the regulations.
Laboratory 12, which represented the regulating agency showed the
facility to be In violation. The third laboratory, a referee
laboratory, produced results which were just below the action level.
Table 2 descrIbes a similar situation for lead data produced by a
single laboratory. For the sample in question results ranged from
1.26 to 49 mg/i (lead). Within this laboratory duplicate analyses
for water samples normally agree to within 10%. Results for solid
samples normally agree to within 20%. Clearly there are factors
unique to the EP TOX procedure which are causing widely scattered
results.
EXPERIMENTS AND RESULTS
The Reg. V CRL conducted a series of experiments in order to
identify the cause of the variability in the EP TOX metals results.
The RCRA Study Vu performance evaluation (PE) sample was chosen for
study because reference values are available and because CRL had
originally reported low results for several metals. Only the
affected metals (Cr, Pb, Zn, and Ni) were studied.
5-48
-------
Initially CRL hypothesized that the variable results could be caused
by using sample weights other than the 100 g. specified in the
method. In theory smaller samples would be extracted more
thoroughly because of better contact with the acid. However,
discussions with other laboratories indicated that the sample size
was not a factor: Several labs used small sample weights and
produced results near the reference value. Historical data from CRL
supports this observation. Also, when CRL reanalyzed the PE sample
the results were similar although different sample weights had been
used.
RL investigated the laboratory water as a possible cause for the
low results due to CO 2 content of pH effects. CLR extracted the PE
sample using distilled, deionized laboratory water as well as
distilled water from an outside laboratory. Table 3 shows no
significant difference between the results for the two sources of
water. Moreover, neither source produced low results.
RL also investigated the possibility that the nature of the
extractor could be causing the difference in results. A stirrer and
a tumbler extractor were compared. Table 4 shows slightly higher
results for the tumbler. However, the difference is not
sufficiently large to explain the variability in results. Again,
both mixers produced results which were higher than the originally
reported results.
At this point CR1 had six sets of data for Pb, Cr, Ni, and Zn. All
had been produced by following the specifications of the official
method. Three sets of data described low results and three sets
described higher results especially for lead. Also, the low results
were internally equivalent, as were the higher results. CR1
reviewed the analytical technique and observations of the analysts
involved in the study. It became evident that lower results were
obtained when the technique used to adjust the pH included addition
of small aliquots of acid along with constant miring. When larger
aliquots were used and the samples were not mixed continuously the
lead results were considerably higher. Also, on one occasion the
analyst allowed the pH of the sample extract to fall to 4.4. In
this case the results for all four metals were considerably higher
(Table 5). It is interesting to note that the results are closest
to the reference values in the extract where the pH was allowed to
drop below the required level of 5.0 ± 0.2.
Discussion with other labs supported our observations. In fact some
labs intentionally adjust the pH to 4.8 to save time in the pH
adjustment step and to assure higher recoveries.
Review of limited historical data indicates that those samples which
require frequent p11 adjustment are most sensitive to variability in
results.
5-49
-------
NOTE: Additional experiments could not be conducted because the
PE sample had been exhausted. Attempts to synthesize a
similar sample were unsuccessful.
DISCUSSION AND CONCLUSIONS
Original sample weight, laboratory water, and type of mixing
apparatus did not significantly affect the recovery of EP TOX metals.
Certain ‘sensitive’ sample types are prone to highly variable
recoveries for EP TOX metals. These include samples which require
frequent pH adjustment in order to maintain the pH at 5.0 ± 0.2.
There Is some evidence that samples with high iron contents may
contribute to this phenomenon.
For lead, in sensitive samples lower results can be obtained when
the pH is adjusted by adding small ( 1 ml) aliquots of acid with
constant stirring. The results are higher when the aliquots of acid
are larger and stirring Is not constant. This Is probably because
in the second case the pH is actually falling below 4.8 in some
areas of the sample. pH 4.8 is known to be a critical point for the
solubility of lead In aqueous solutions.
If the p1! of the extract is allowed to fall to 4.4, significantly
higher recoveries are observed for Cr, Ni, and Zn in sensitive
samples.
RECOMMENDATIONS
If the EP TOX test is to remain in use, EPA should specify the
technique for adjusting the PH. We recommend that acid should be
added In small aliquots using a buret and that the samples should be
stirred constantly. We also recommend that the method Include a
cautionary statement describing the possible effects of exceeding
the allowable pH range of 5.0 ± 0.2 at any time or point In the
extraction. We also recommend that the method specify mixing times,
rather than chronological time for carrying out the extraction.
TABLE 1. EP Pb Data (3 labs)
Sample matrix — oil
LAB I Pb (mg/l) RCRA action level (mg/l )
1 0.2 5.0
2 12.0 5.0
3 4.7 5.0
5-50
-------
TABLE 2. EP TOX Pb Data (Single lab)
Sample matrix — industrial solid waste
Pb (mg/i) Action Level (mg/i )
(1.26) 5.0
11.3 5.0
49.0 5.0
40.2 5.0
19.5 5.0
34.0 5.0
X 30.9
S 15.3
( ) Outlier — rejected in calculation of statistics
TABLE 3. The effect of the source of water on the EP TOX results
for Cr, Pb, Zn, and Ni
Sample matrix — RCRA PE Sample
Reference value EP TOX results (mg/i)
Metal ( mg/i) CRL water (n2) Reference Water
CR 1.68 0.2 0.2
Pb 219.0 186.0 184.0
Zn 0.34 0.2 0.14
Ni 65.8 33.0 24.2
TABLE 4. The effect of the extraction vessel on EP TOX results for
Cr, Pb, Zn, and Ni
Sample matrix — RCRA PE Sample
Reference value EP TOX results (mg/i)
Metal ( mg/l) Stirrer Tumbler (n=2 )
Cr 1.68 0.2 0.2
Pb 219.0 162.0 198.0
Zn 0.34 0.15 0.2
Ni 65.8 20.3 24.2
5—51
-------
TABLE 5. The effect of pH control on EP TOX results for Cr, Pb,
Zn, and Ni
Sample matrix — RCRA PE Sample
EP TOX results (mg/i )
Reference value narrow pH normal pH pH below
Metal ( mg/I) control (n=3) control (n3) 4.8 (n=l )
Cr 168 0.2 0.2 2.7
Pb 219.0 14.0 190.0 250.0
Zn 0.34 1.0 0.2 0.5
Ni 65.8 31.0 23.0 78.0
ACKNOWLEDGEMENTS
The authors thank Kanchan Patel of the Illinois Environmental
Protection Agency for providing reference water. The authors also
thank Florence Williams of the US EPA Office of Solid Wastes for
providing reference values and information related to the PE saniple.
REFERENCES
Test Methods for Evaluating Solid Wastes , Second Edition, US
Environmental Protection Agency Publication SW—846, July, 1984.
5-52
-------
ION CHROM(YIOGRl PHY FOR THE Z½NP LYSIS OF ANIONS
IN HP ZM DOUS W1 STE MATRICES
Roger Kell, Group Leader of Wet Chemistry, E. Scott Thcker, Manager of
Analytical Chemistry, Sherry Cogan, Analytical Chemist, Chemical Waste
Management, Inc. Technical Center, Riverdale, Illinois; Robert J.
iJoyce, Dionex Corporation, Sunn ’vale, California
ABSTRACT
Accurate quantitation of inorganic anions is of interest to those
involved in the analysis and disposal of hazardous wastes, and those
needing information necessary for incineration of the wastes. The
information also serves a confirmatory role in cross—checking other
parameters such as acid content. Formerly, quantitation of these
anions required a unique procedure for each ion necessitating large
expenditures of time and materials. Ion Chromatography has the
capability of performing single, simultaneous determinations of
several anions with a minimal number of interferences and materials.
It can also be readily automated.
Ion Chromatography using a carbonate/bicarbonate eluant and chemical
suppression is compared to colorimetric and selective ion electrode
techniques for the identification and quantification of several anions
in hazardous waste zuatricies. Steps taken to ensure on—line readiness
as well as the acceptability of analytical results produced using Ion
Chromatography are outlined.
Samples prepared by oxygen bomb digestion were split into equal
portions and analyzed using both Ion Chromatography and the previously
utilized procedures to quantitate chloride, sulfate and fluoride in
typical samples. Evaluation of the results confirmed that under the
Ion Chromatographic conditions tested, this method was an acceptable
alternative to chlorimetric analysis for sulfate and chloride.
However, it was not a suitable substitute for the selective Ion
Electrode analysis for fluoride. It was found that several organic
anions interfere with the fluoride measurement when using the typical
carbonate/bicarbonate eluant. Alternate eluants could possibly
eliminate these interferences but were not investigated.
A parallel analysis of hazardous waste samples for incineration, in
which two different Ion Chromatographs were used to quantitate
chloride, bromide and sulfate gave an overall discrepancy in results
of less than 20% in most cases. This was considered acceptable in
light of the complex nature of the sample matrices tested.
INTRODUCTION
Knowledge of the type and level of anions present in hazardous waste
streams or generated during treatment of these streams is important
and necessary for proper hazardous waste management in areas such as
solvent—derived fuels programs and incineration units. The anions of
interest are Fluoride (F ), Chloride (Cl ), Bromide (Br ), and
Sulphate (SO ), with other anions such as Nitrite (NO 2 ), Nitrate
(NO 3 ), and hosphate (P0 4 3) of interest but lesser importance arid
frequency in commonly encountered samples.
5-53
-------
Prior to the investigation of Ion Chromatography for the measurement
of these anions, colorimetric and selective ion electrode methods
have been employed. This required separation of the prepared sample
into single test portions——one for each anion of interest. Applying
calibration standards, blanks, duplications, and fortifications to
each test portion, as many as 35 individual determinations were
necessary per sample received.
Ion Chromatography has emerged in recent years as an attractive
option for the simultaneous separation, identification, and
quantification of several anions present in a single sample. The
following method of analysis applies specifically to the Dionex 2000i
Series, Ion Chromatograph with Autoion 100 Controller, ASM—l
Autosampler and Spectraphysics 4270 Integrator, available from Dionex
Corporation, 1228 Titan Way, Sunnyvale, CA 94088—3603. ‘Where
possible, attempts have been made to generalize the procedures.
Information provided may or may not be applicable to other systems.
SCOPE AND APPLICATION
This method addresses the quantification of commonly encountered
inorganic anions in liquid and solid, domestic or industrial waste.
Detection at the ppm level is achievable in most matrices.
SUMMARY OF METHOD
Theory: The mechanism by which anions are separated in Ion
chromatography is based on the relative affinities of the various
ions for the column resin. The columns are packed with thousands of
large sulfonated (negatively charged) resin beads surrounded by a
greater number of smaller, aminated (positively charged) resin beads.
The positive charges on these aminated beads attract the negatively
charged eluant anions——carbonate and bicarbonate. When a sample
containing analyte ions Is injected into the stream of eluant flowing
through the column, they compete with the carbonate and bicarbonate
ions for available sites. Monovalent ions, such as chloride,
displace the similarly valenced bicarbonate ions; divalently charged
ions, such as sulfate, displace carbonate ions. The analyte ions are
able to temporarily displace the eluant ions because they have a
greater affinity for the resin. The degree of this affinity varies
among different species of anions, as evidenced by the chloride peak
in a chromatogram emerging before the sulfate peak. Eventually, all
of the analyte anions will be displaced by the large volume of eluant
ions competing for sites on the resin and pushing analyte ions back
into the eluant stream to be flushed onward through the column.
After the eluant, containing the recently separated analyte anions,
emerges from the column, It enters the suppressor column. The
suppressor enhances the conductivity of the analyte anions and
suppresses the conductivity of the eluant ions. In the suppressor,
Ions are exchanged with dilute acid, which protonates them to
create the weakly dissociated, low—conductance carbonic acid from
the eluant (to reduce background noise in the chromatograni) and the
strongly dissociated, highly conductive analyte acids, such as
5-54
-------
hydrochloric and sulfuric acids, which will give strong signals
when passed through the conductivity detector and produce peaks in
the chromatogram.
Organic compounds may attack the polystyrene—divinylbenzene resin
used to pack separator columns. Any sample which contains organic
material must be rendered suitable for the ion chromatograph. One
method for this involves combustion of the sample in oxygen via a
bomb calorimeter. After digestion, dilution, and filtration, samples
are injected onto the column for subsequent separation and
quantification.
Aqueous samples require only dilution and filtration; solids must
first be put into solution. After filtration, samples are injected
onto the column for subsequent separation and quantification.
INTERFERENCES
Either one of two species which elute with similar retention
times/volumes can interfere with identification and quantification of
the other. For example, nitrate can interfere with bromide. It is
recommended that the ratio of their concentrations be no greater than
10:1 if they are to be simultaneously quantitated. Similarly,
several organic acids of low molecular weight (such as forinate and
acetate) interfere with the analysis of fluoride since they co—elute.
Since detection of ions is based on the specific conductance of the
eluting solution, another interference occurs as a negative peak due
to the water used in dilution and sample injection. A negative peak
(“dip”) results from the lower conductance of water relative to that
of the carbonate/bicarbonate eluant. This can interfere with the
integration of early—emerging peaks which must be sufficiently
resolved from the dip to allow successful quantitation. This
interference can be eliminated by using the carbonate/bicarbonate
eluant to dilute the sample.
In samples containing a component which is present at a high level of
concentration, the peak caused by that particular ion can interfere
with the analysis of subsequently eluting ions. When the column is
overloaded with an extremely high concentration of anions, the eluant
ions cannot replace the analyte ions at a high enough rate within the
column to achieve the sharp, narrow peaks obtained at lower levels of
concentration. The peak broadens, causing not only a shift to longer
retention times, but, possibly, a great enough effect to “bury” the
next peak by preventing proper resolution. Dilution of the sample to
the appropriate degree may minimize the interference.
There are many classes of compounds which can interfere with the
accurate quantitation of inorganic anions by occupying available
sites on the column resin, thereby making the sites unavailable to
analyte ions and reducing column capacity. Some of these contam-
inants bind irreversibly. They include dO 4 , Fe (CN) 6 3 , other
metal—cyanide complexes, aromatic organics, large aliphatic
organics (i.e., surfactants), and humic acids. Al , high
concentrations of Fe, Ni, Cu, Zn, hydroxide ion, peroxides,
5-55
-------
dissolved gases, and microbial growth have the potential to
interfere with analyses. Because these sources of contamination
usually cause irreversible damage, prevention may be the best mode
of avoiding the interferences they create. Filtration and
degasification of eluants can prevent particulate accumulation, as
well as the formation of bubbles within lines or columns.
Microbial growth, encouraged by the hospitable carbonate!
bicarbonate environment provided by the eluant used in this method,
can be discouraged by frequent flushing and proper storage of
columns. More importantly, fresh eluant should be prepared
frequently in small batches to be used immediately. Storage of
diluted eluant can allow microbial growth and cause column
blockage. Routine column clean—ups with iN }IC1 can reduce the
concentration of some metals and inorganic species retained by
columns.
APPARATUS AND MATERIALS
Ion Chromatograph (see Figure 1) consisting of a pump, an injection
valve, a conductivity detector, a suppression device or column, and a
set of columns, including a separator suitable for the separation of
inorganic anions, a guard column of the same material, and another
guard column specifically for trapping out metals that may be found
in sample matrices.
Integrator or chart recorder.
Volumetric glassware.
Pipets (5 microlitera to 10 milliliters).
Syringes and Luer Lock Adapters.
REAGENTS
Nitrogen gas 99.998% purity (for valve actuation).
18 Megaohm deionized water, filtered through a 020 micro membrane
filter.
Na 2 CO 3 , reagent grade.
NaHCO 3 , reagent grade.
H2S04 concentrated, reagent grade.
HC1, concentrated, reagent grade.
Sodium Tartrate, reagent grade.
Acetonitrile, HPLC grade.
5-56
-------
Figure 1
Ion Chromatograph Components
METAL OR
ORGANIC
TRAP
GUARD COLUMN
5-57
-------
Standards Preparation: 1000 ppm stock solutions should be prepared
monthly, as follows, by dissolving oven—dried (at 105°C for 1/2 hour
then cooled) reagent chemicals in deionized water. Refrigeration of
the resulting solution is recommended:
F : 2.2100g NaF/liter
Cl: l.6484g NaC1/liter
NO 2 : l.4998g NaNO 2 /liter
Br: 1.4894g KBr/liter
NO 3 : l.3707g NaNO 3 /liter
P0 4 3 : 1.4330g KH 2 PO 4 /liter
S0 4 2 : l.8141g K 2 S0 4 /liter
Calibration Standard Preparation: A mixed anionic solution, suitable
for use as a calibration standard, is prepared from 1000 ppm stocks
in a 1 liter volumetric flask and diluted to volume with deionized
water:
5 ppm = 5 mL. each: F/Cl/Br
15 ppm = 15 mL. each: N0 2 /N0 3 /S0 4 2 /P0 4 3
A free (inorganic) anion Quality Control Check solution can be
prepared from the calibration standard solution by diluting it 1:1.
It will have theoretical anion concentrations of 2.5 ppm fluoride,
2.5 ppm chloride, 2.5 ppm bromide, 7.5 ppm nitrite, 7.5 ppm nitrate,
7.5 ppm phosphate, and 7.5 ppm sulfate.
A total anion Quality Control Check solution, which must be oxygen
bomb digested to liberate organically bound anions, can be prepared
by placing 25 grams each of 2—Chlorobenzothiozole and
Bromo—4—fluorobeflZefle in a 100—ml volumetric flask and making up to
volume with Butyl Acetate. This solution will have an approximate
heating value of 10,700 BTU/lb., and after digestion, total anionic
concentrations of 4.8% Chloride, 12.1% Sulfate, and 7.3% Bromide.
Eluant Preparation: The optimal eluant concentration for separations
is dependent on the column set to be used. As a rule, anion
separations require a carbonate/bicarbonate eluant which can be
prepared in quantity at lOOx the recommended strength and diluted
with deionized water as needed. Eluant should be filtered and
degassed prior to usage. (Bubbling N 2 gas through the eluant or
sonication for 20 minutes is sufficient to remove dissolved gases
that could cause the pump to lose prime. Allowing the deionized
water to reach room temperature before dissolving the salts is also
sufficient.) Additionally, an in—line filter for the eluant which
includes 35— micron and 5—micron filters enclosed in a Teflon
housing, is highly recommended.
5-58
-------
Regenerant Preparation: An anion suppression system is used to
enhance the signal to noise ratio of analyte relative to eluant
anions by converting them to their strongly dissociated,
highly—conductive acidic form while also reducing the interfering
conductance background of the eluant anions via a similar mechanism.
This is achieved by forcing a dilute acid solution to flow through
the suppression device in a direction opposite to the direction of
eluant flow and providing a membrane interface through which exchange
can occur.
SAMPLE COLLECTION, PRESERVATION, AND HANDLING
Free Anion Determinations: Liquid samples should be diluted and
filtered. Solids must be put into solution with deionized water
before dilution and filtered through a 10 micron filter, followed by
0.20 micron filter. Solutions with organic or metal content should
be swirled with an appropriate ion—exchange resin or run through a
preparatory extraction column, unless an organic or metal guard
column is to be used prior to the separator.
Total Anion Determinations: Samples should be oxygen—bomb digested
in a 5% (or greater) (wiw) sodium bicarbonate solution. Bomb
washings should be collected in a volumetric flask and diluted to a
known volume with subsequent rinsings. The diluted solution obtained
from the bomb contents and washings should be filtered in two stages
through a 10 micron filter, followed by a 0.20 micron filter.
Further dilution may be required to produce on—scale peaks.
PROC EDURE
Instrument Initialization: Ensure eluant and regenerant reservoirs
have been filled with the appropriate solutions. Set pump flow rate
to 2 tnL/minute and begin regenerant flow through the suppression
device. Allow system to equilibrate approximately 15 minutes or
until pressure and conductivity readings stabilize. The conductivity
reading for eluant only should be approximately 15 microsiemens.
Flush and load the sample loop by injecting approximately 1 niL of
deionized water (from 10 cc luer lock syringe). “Autozero” the
conductivity cell.. Ensure a flat baseline is observed. If baseline
is not flat, check eluant flow—rate, check for leaks, and recheck
regenerant regulator pressure. (A high—drifting baseline indicates
that pressure is too low, and a low—drifting baseline indicates
pressure is too high.) (See Figure 2.) If this does not produce the
desired baseline, a longer equilibration period may be required.
Instrument Calibration: If a flat baseline is observed with
deionized water, inject the calibration standard to establish
retention times for the analyte peaks. Once retention times have
been noted, they can be entered, along with the correct concentration
values into the integrator and the calibration standard reinjected to
serve as a calibration and Instrument Performance Check solution.
(See Figure 3.) (If the recording device used does not have this
capacity, a series of calibration standards can be used to set up
linear calibration curves from which relative sample concentration
5-59
-------
values can be calculated using peak heights or areas.)
When using an integrator with the capacity to automatically
recalibrate, the calibration standard should be run once for every
ten sample runs to ensure accuracy in integrator output. To monitor
column degradation, the Instrument Performance Check solution can be
used to calculate the theoretical number of plates. (See
CALCULATIONS section.)
The Quality Control Check solutions should be injected and
concentration values determined to be within acceptable limits prior
to sample analyses.
CHANNEl. A
6.93
CHANNEL A
INJECT 08/06/86 06:17:09
INJECT 07/26/86 08:23:03
.76
INJECT 09/10/86 11:45:58
FIGURE 2
BASELINE ADJUSTMENT
a. High—Drifting Baseline b. Low—Drifting Baseline c. Flat Baseline
CHANNEL A
.66
2.56
5-60
-------
CHANNEL A
INJECT 07/31/86 13:41:28
L 89 •
Bromide
Ni trate
Phosphate
7.76 Sulfate
Fluoride
Chloride
2.38 Nitrite
Separator Column: Dionex AS4A
Flow Rate: 2 mi/mm.
Eluant: Sodium Carbonate/Bicarbonate
FIGURE 3
TYPICAL CALIBRATION SOLUTION RUN
3.85
4.66
5—61
-------
Sample Analyses: When injecting samples manually or using an
autosampler, deionized water should be injected frequently (after
every 7th sample) to aid in preventing cross—contamination of samples
and to provide a continual baseline check.
In the case of a complex chromatogram or insufficient resolution of
peaks, a spike solution of known concentration should be added to
allow positive Identification of ambiguous peak(s). Additional
dilution(s) of samples may be necessary to obtain on—scale peaks.
Column Clean—ups: Constant care of columns will help to ensure
chromatograms are clear and data reliable. Obvious signs of column
degradation include:
Marked Increase In system backpressure.
Poor resolution between peaks (i.e., “valleys” less than 50% of
peak heights).
Shift In retention times of 10% (i.e., chloride peak shifts from
1.60 to 1.79 or more),
NOTE: Changes in eluant concentration can cause a shift in
retention times. When changing eluant, ensure column has been
equilibrated and calibration standard analyzed for new retention
times.
When column degradation occurs to the extent that the above signs are
noted, the guard and/or separator columns should be cleaned.
Separator and Guard Column Clean—up:
Disconnect the suppression device.
Remove organic trap guard column and reverse the order of the
separator and guard columns. (Frequent clean—ups on the guard
column alone should reduce the necessity for cleaning the
separator column.)
Pump deionized water through the column at a flow rate of 2.0
rnL/min. for 10 mInutes.
Pump iN RC1 through the column at a rate of 2.0 mi/mm. for 20
minutes.
Pump deionized water through the column at a flow rate of 2.0
mi/mm. for 5 minutes.
Remove the used bed supports and replace with new at column
ends.
5-62
-------
Organic Trap Guard Column Clean—ups:
Pump a 90% solution of acetonitrile through the column (with all
other columns and the suppression device disconnected) at a flow
rate of 0.5 mi/mm. overnight (or, if necessary, at a rate of
2.0 mi/mm. for 1—2 hours).
Pump deionized water through the column at a flow rate of 2.0
mi/mm. for 15 minutes.
Replace the bed supports.
After clean—up is complete, replace columns in their proper sequence.
Re—equilibrate by pumping eluant through the system.
NOTE: A mild clean—up for guards and separators which is not as
harsh but also not as effective as the clean—up described above,
consists of pumping 0.10 M Na 2 CO. over the reverse—ordered (separator
followed by guard) columns, with trap and suppressor disconnected,
for 20—30 minutes, remove used bed supports and replace with new,
resequencing the columns, and, finally, re—equilibrating the system
with the proper eluant.
CALCULAT IONS
Chromatographic peaks are quantitated on the basis of height or area
of peaks. Peak heights are best utilized in instances where peaks
are sharp and narrow. Peak areas are more appropriately used when
peaks are broad and symmetrical. This depends on several factors,
including the type and efficiency of the column used as well as its
age and degree of degradation. For integrators which identify and
quantify peaks using the method of external standards:
Concentration of Anion (ppm) =
(Peak Height or
Area of Sample ) (Dilution (Concentration of
(Peak Height or x Factor) x Anion in Calibration
Area of Calibra— Standard)
tion Standard)
Column Efficiency/Degradation: As a guide to assist the analyst in
determining column efficiency and degradation, calculating the
theoretical number of plates and the resolution of two separate peaks
can be employed:
Theoretical Number of Plates =
t = retention time of peak
N = 5.545 — where
W W 1 width at half height
i — ofpeak
2 2
5-63
-------
Resolution of Peaks (R)
t difference in retention times of
2 t the two peaks
R W +W where W
2 1 1 width of first peak
W 2 width of second peak
Method Detection Limits (MDL): To determine method detection limits,
solutions containing a single anionic specie at a concentration near
the expected detection limit should be analyzed a minimum of seven
times. The mean and standard deviation for these values is then
calculated. The method detection limit is equal to three times (3X)
the standard deviation. (See Figure 4 for typical values.)
Limits of Quantification (LOQ): In order to ensure a margin of
confidence between the method detection limit and reported values,
the method detection limit can be multiplied by an additional factor
(usually 10) to obtain a lower limit of quantification. (See Figure
4 for typical values.)
In cases where peaks occur below the level of quantification (MDL x
10), use the following calculation to obtain the threshold value for
the analysis:
(Log) x (Dilution Factor) Conc. of Anion (ppm)
with this result reported as “less than.”
LOQ (MDL x 10)
ANION (ppm ) s MDL (ppm) ( ppm )
Chloride 0.217 0.016 0.048 0.48
Nitrate 0.935 0.040 0.119 1.19
Sulfate 1.155 0.030 0.090 0.90
Phosphate 0.750 0.049 0.146 1.46
Bromide 0.955 0.0095 0.029 0.29
FIGURE 4
TYPICAL VALUES FOR MET}IOD DETECTION LIMITS ANT) LIMITS OF
QUANTIFICATION
5-64
-------
QUALITY CONTROL
Analyze the Instrument Performance Check solution once per batch of
samples. From this, calculate the theoretical number of plates for
phosphate, checking to ensure the value is within control limits
before proceeding. Analyze the Quality Control Check solution before
each batch of samples, checking to ensure values are within control
limits for each anion to be analyzed before proceeding. Duplicate
and fortify every tenth sample. Duplication error should not exceed
20%. Fortification recovery should be between 80 — 120%. Normally,
free anion samples can be fortified with an appropriate amount of the
free anion Quality Control Check solution. Total anion
(bomb—combusted) samples should be fortified prior to digestion with
an appropriate amount of the total anion Quality Control Check
solution.
Duplications (10% of Samples)
Mean X = ÷ X 1 where X 1 lower result
X 2 higher result
2
Percent Error = ( X 2 — X ) 100
x
Fortifications (10% of Samples)
Theoretical Spike ( Conc. of Std.) x (Wt. of Std. )
(Wt. of Sample)
Percent Recovery = ( Spike Result) — (Avg. ÷ Dup. + Orig. ) x 100
Theoretical Spike
5-65
-------
SAMPLE HANDLING FOR THE ANALYSIS OF
CYANIDES IN SOLID AND HAZARDOUS WASTES
C. John Ritzert, Manager—Technical Operations, Lancy Environment
Services, Zelienople, Pennsylvania
ABSTRACT
Cyanide bearing wastes are among the most common industrial wastes
disposed of In this country. Wastes generated from electro—plating,
petroleum, metal finishing, pharmaceutical, mining, coking and other
manufacturing operations each present unique handling problems due
to their complex matrices. EPA has expressed a great deal of
concern regarding the disposal of these cyanide wastes and has
proposed regulations based on the form of cyanide present in the
waste. EPA ’s Intent Is to control those cyanide complexes which are
toxic to humans and animals and which have potential for formation
of hydrogen cyanide gas when mismanaged.
The challenge facing EPA and the regulated community Is to develop
analytical methodology which will determine those cyanide forms
which are a detriment to the environment. The currently approved
method for evaluating 8olid waste (Method 9010 for total and
amenable cyanides) Is found In SW846. Although this method is
sufficient for analyzing these cyanide forms in water and
wastewater, it does not deal adequately with complex waste matrices
such as tars, oils, sludges or soils.
This paper discusses cyanide forms of environmental significance
with regard to decision making criteria. Problems associated with
preparing and handling of waste matrices are discussed, and
procedures are proposed for solving those problems. Data gathered
from the application of proposed methods are presented.
INTRODUCTION
One of the most challenging compounds to analyze, for both
Industrial and environmental control, is cyanide. Simply
understanding definitions and terms associated with the complex
forms of cyanides, Its analytical chemistry and interferences is a
challenge. It is apparent that we need to focus efforts on
protocols that not only meet a regulatory need, but also provide
users of analytical methods with appropriate information to make
decisions.
For environmental decision making, EPA’s intent is to control those
cyanide complexes which are toxic to humans and animals and which
have potential for formation of hydrogen cyanide gas when
mismanaged. The challenge, therefore, is to define those cyanide
5-67
-------
forms which are toxic under these conditions and provide adequate
methodology to accurately teat for these forms.
The currently approved method (SW846, Method 9010) is executed
wholly from water methods published by EPA. This is an excellent
place to begin method development because of the enormous amount of
method study conducted by ASTM, Standard Methods Committee, EPA and
Industry. However, compared to the varied aolid, slurried and oily
matrices of wastes, analysis of waters is probably a cake walk.
METHOD S LZCTI0M
In order to develop appropriate methodology for handling these
varied matrIces, we must define and evaluate several variables which
influence our ability to predict the impact of cyanide forms on
human health and the environment. First, we must determine cyanide
forms of interest based on the utility of each form to predict
environmental impact. Forms of cyanide identified below were chosen
because they are environmentally significant and their analytical
methodology La well establIshed, with documented accuracy.
o Total Cyanide
o Amenable Cyanide
o Microdif fusible Cyanide
o Weak Acid Dissociable Cyanide
The analytical methods for each of these forms have the ability to
measure hydrogen cyanide gas which could be evolved when wastes are
mismanaged in an acid environment. The total cyanide procedure,
however, also measures those cyanides which do not easily dissociate
In this management scenario. Method 9010 defines amenable cyanide
by the difference between two measurements, total cyanide before and
after chlorination.
Experienced analysts are well aware of difficulties which arise from
the use of this technique. The microdif fusible cyanide method is a
procedure which has recently proven Itself as a valuable technique
for estimating dissoclable cyanides. However, this is a time
consuming procedure, and the accuracy and reproducability of the
data are highly dependent on the analyst’s experience and skill.
The weak acid dissoclable cyanide method Is a distillation procedure
which measures cyanide evolved from a mildly acidic (pH 4—5)
solution.
Of the procedures described, total cyanide and weak acid dissociable
cyanide procedures remove cyanides from the sample matrix into an
easily analyzed matrix of sodium hydroxide in water. These methods
are also direct methods, requiring minimal operator manipulation,
and both forms can be determined using the same sample handling and
preparation procedures.
5-68
-------
Total cyanide is typically viewed as a required parameter for the
determination of disposal options. This methodology must,
therefore, be defined for complex matrices of soils, sludges, oils
and tars. Because of the decided advantage of having only one set
of sample handling and preparation procedures for both these
methods, weak acid dissociable cyanide methodology is the preferred
method for evaluating the environmental impact of wastes for the
management scenario mentioned earlier. This method provides
accurate, reproducible, and rapid results necessary for decision
making.
SAMPLE HANDLING
Over the past thirty pius years, water chemists have refined cyanide
analyses to provide reliable data of defined accuracy The basic
approach to developing methods for complex solid and hazardous waste
matrices is to develop sample preparation procedures that take
advantage of these tested methods by removing cyanides into a water
matrix. It has been long established that distillation is an
appropriate technique to accomplish this transfer.
Directly applying distillation to the matrices of soils, sludges,
tars, and oils presents several problems that were not dealt with In
previous water method development.
1 — How does particle size of a sample effect recovery of
cyanides?
2 — When solids accumulate in the bottom of the flask, is
recovery reduced?
3 — When oils or tars are added directly to the flask containing
water, is recovery reduced because of lack of misibility?
These questions and others are being dealt within a joint effort by
EPA/OSW and ASTM Committee D—34 on Waste Disposal. The established
water methods are being re—evaluated and revised to provide suitable
methodology for the wide variety of hazardous and solid wastes.
The first change made was to add magnetic stirring to the classical
cyanide distillation apparatus. Data from replicate determinations
of solid samples are typically non—reproducible, with high standard
deviations, when not agitated. One such set for results is shown
below. These poor results are caused by the solids laying
undisturbed on the flask bottom, limiting solution contact.
Determinations on a similar sample with adequate agitation produced
much more acceptable results.
5-69
-------
Replicate
1 86
Not agitated 2 34 Standard Deviation — 253
3 33
4 62
Replicate
1 20
Agitated 2 18 Standard Deviation — 26
3 21
4 24
This approach will not overcome problems with sample non—homogeneity
but will significantly improve the chances of obtaining reproducible
results.
A second addition to the method is a proposed requirement for
particle size reduction to less than 60 mesh for monolithic and
granular samples. This proposal is not yet founded in scientific
evidence, but is the product of some logic. It is important to
handle a sample in a manner that maximizes the method’s ability to
analyze any cyanide that may be absorbed into a porous sample.
After some discussion, 60 mesh was selected as small enough to
provide accurate results while still being practical and attainable
by almost any laboratory.
Oil and tar present problems of their own because they are iminisible
In water. An extraction procedure based on the technique for fatty
acid removal prior to cyanide analysis is proposed. The sample is
quantitatively added to a separatory funnel containing a buffer
solution. A suitable solvent (iso—octane, hexane) is added, and the
sample Is extracted. The buffer solution is then transferred to the
distillation flask along with any solids, and analysis is begun.
This procedure is still in the development stage, and there are few
data available on accuracy and reproducibility.
CASE DATA
Unfortunately, we are just now at the beginning stages of developing
new methods, and there are few data available on all the
modifications discussed. There are examples of data available,
however, that compare results of various methods. One such example
is illustrated below.
5-70
-------
Cyanide Method Duplicates Duplicates
(mg/Kg) (rag/Kg) (mg/Kg) (mg/Kg)
Total 4700 4800 1.4 1.6
Weak Acid Dissociable 185 180 0.19 0.22
Amenable * * 0.2 0.5
Microdiffusion 99 103 0.100 0.100
*Matrix interference
Total and weak acid dissociable data were generated using the
magnetic agitation modification mentioned. These data represent
excellent reproducibility. The asterisk notes that, due to matrix
interference, reliable results could not be obtained, pointing out
the necessity of applying an alternate technique f or this
determination.
Cyanide recovery utilizing sample agitation on distilled spike
samples Is shown in the table below. It is Important to note that
both the total and weak acid dissociable cyanide procedures
demonstrate acceptable recovery. Data for amenable cyanide by
difference shows both poor recovery and inadequate detection
capability due to interferences.
Percent
Cyanide Method Sample Recovery
(mgi L)
Total 312 93%
Weak Acid Dissociable 1.2 86%
Amenable <40 70%
Microdif fusion 0.308 83%
SUMMARY
A major effort has been Initiated by EPA and ASTM to address the
short—comings of published cyanide methods for solid and hazardous
wastes. The number and variety of waste matrices requiring testing
challenges the analyst’s ability to generate reliable data for
today’s environmental decision making. Progress is being made on
developing sample handling procedures and methods should be
available in the near future.
5-71
-------
DETERNINA TION OF TOTAL SULFIDE IN SOLID WASTE
Nirtha Umana, Jeff Keever, James Beach, and Linda Sheldon, Analytical
and Chemical Sciences, Research Triangle Institute, Research Triangle
Park, North Carolina
ABSTRACT
The purpose of this research was to improve, modify and revise Method
9030, in order to make it applicable to a wide variety of waste
matrices. AS part of this effort, a literature review and experiments
were performed to identify potential method interferences and
pretreatments to remove them, to identify and appropriate techniques
for extracting suif ides from solids and oils, and to develop a
procedure to determine acid—insoluble sulfides.
From the literature review, reported data for Method 9030 suggest that
recovery is better than 90%. The precision for sulfide determination
is relatively good (CV 6%). The method detection limit was found to
be approximately 0.2 mg/L.
Experiments performed in our laboratory demonstrated that sulfite and
bisulfite are significant interferences to Method 9030, but they can
be removed with formaldehyde. A distillation procedure was developed
for this purpose. The recovery obtained with the distillation
procedure was approximately 100%, the precision was good (CV 7.5%),
and the method detection limit was approximately O.2mg/L.
An extraction procedure was developed for solid samples that cannot be
distilled. Results of extraction experiments demonstrated that
sulfides can be extracted from solids and oils with an alkaline (pH
12) solution. Modifications to the extraction procedure may be
required depending upon the physical and/or chemical properties of the
matrix.
An acid—insoluble sulfide procedure was also developed.
Acid—insoluble sulfide can be determined using a concentrated HC1
distillation.
The three developed procedures were tested with industrial wastes to
ascertain their range of applicability. A Modified Method 9030 has
been prepared incorporating the three developed procedures.
5-73
-------
TOX - A SCREENING PARAMETER FOR ENVIRONMENTAL SAMPLES
John L. Snyder, Chemist III, Lancaster Laboratories, Inc.,
Lancaster, Pennsylvania; Peter N. Keliher, Professor of Analytical
Chemistry, Villanova University, Villanova, Pennsylvania
ABSTRACT
The EPA procedure for determining Total Organic Halogen in water
samples is described In Method 9020, SW—846. In this method,
organic halogens present in water samples are adsorbed onto an
activated carbon column which is then pyrolyzed in a high
temperature oxygen furnace. These organic halogens are converted to
hydrogen halide gas, trapped in an acetic acid buffer and
microcoulometrically titrated. However, the method as written only
applies to TOX determinations in water.
Although not as specific as more complex analytical techniques such
as Gas Chromatography or GC/MS, TOX is relatively a rapid and
Inexpensive technique. Many organic pollutants contain organic
halogens. For example, of the 113 organic priority pollutants, 71
are halogenated, with chlorine being the predominant species. TOX,
therefore, can be used as a screening tool to determine the extent
of pollution in many environmental samples and to assess if more
specific chemical analysis is necessary.
In this paper a Mitsubishi TOX—lO Analyzer was used to determine TOX
on a variety of environmental samples. These included groundwaters
and wastewaters, transformer oils, waste solvents, sludges, soils,
and TCLP extracts. The sample preparation and analytical techniques
necessary to determine the TOX on each of these sample matrices will
be discussed. Brief mention will be made concerning interferences
to the TOX determinations. Finally, comparisons will be made
between TOX data and data from other analytical methods such as Gas
Chromatography and GC/MS.
INTRODUCTION
When using regulatory language, the term TOX (Total Organic Halogen)
usually refers to an analytical technique outlined in the EPA Method
#9020 found in SW846)- The method developed in Germany during the
early 1970s from a method to check granular activated carbon filters
used in water treatment plants for breakthrough of halogenated
organic contaminants. The method was further refined to its present
state by the Drinking Water Research Division of the Municipal
Environmental Research Laboratory, U.S. EPA, Cincinnati, Ohio. 2
Instrumentation to semlautomate the method was developed by private
industry under contract to the EPA. Presently two commercial
5-75
-------
instruments are marketed In the United States. 3,4 ( t more
extensive history of TOX is given in reference 2.)
In Method #9020 a water sample is passed slowly through two
activated carbon columns held in series to adsorb the halogenated
organic compounds. These carbon columns are then washed with a KNO 3
solution to remove Inorganic halides and are combusted in a high
temperature furnace. Organic halogens are converted to the hydrogen
halide gas and carried by the gas stream to a titration cell where
the gas is trapped In a acetic acid/acetate buffer. The hydrogen
halide precipitates silver ions from solution and lowers the preset
end point potential, which is monitored by a silver/silver/chloride
electrode system. A constant pulsed current generates silver ions
and coulometrically titratea the Boltition to the end point
potential. By the application of Faraday’s law, the concentration
of the organic halogen in the sample can be calculated as Cl.
The term TOX requires some qualification. For our purpose, any
compound with a C—X bond will be considered an organic halogen. The
common halogens, chloride, bromide, and iodide are detected, though
not individually, by the method. Fluoride, because of the
solubility of AgF in aqueous solution is not detected as TOX.
The term Is further qualified by the fact that only those organic
halogens which are adsorbed onto the activated carbon and not
removed by the washing step will be detected as TOX. It has been
shown that some organic halogens, or example chioroacetic acid,
weakly adsorb onto the activated carbon and can be removed during
the washing step. 5 In addition, other nonhalogenated organic
compounds can compete with halogenated organic halogens for
adsorption on the activated carbon column. 6
One of the main advantages of TOX Is that it can serve as a fairly
rapid screening technique for pollutants. Many of the contaminants
which man has introduced into the environment are halogenated. This
halogenatlon not only makes these compounds more resistant to
biodegradation, but also increases their toxicity to aquatic life,
animals, and man. consider the priority pollutant compounds.
Seventy—one of the 113 organic compounds on the list are halogetiated
with chlorine being the predominate halogen. These reasons,
combined with the fact that carbon containing compounds are
ubiquitous in nature make TOX superior to TOC (Total Organic
Carbon), another commonly used analytical tool, as a pollution
indicator.
TOX has most often been applied to aqueous samples. In this paper,
lOX was determined on a number of different kinds of environmental
samples. These Included ground waters, TCLP extracts, waste oils,
solvents, soils and sludges. Spike recovery data was obtained for a
variety of halogenated compounds in these matrices. The precision
5-76
-------
of the TOX method was also determined. Comparisons of TOX to other
analytical methods were made.
EXPERIMENTAL
Two Mitsubishi TX—b analyzers were used to determine TOX. These
were kept in a positive pressure room free of halogenated solvent
vapors.
Aqueous Samples : Samples were loaded through two 40 mg activated
carbon columns arranged in series at a rate of 3.3 mi/minute to
adsorb organic halogens. Following loading, these columns were
washed with 4 ml of 0.8 potassium nitrate solution and transferred
to the TOX furnace. The following instrument settings were used for
burning the carbon columns:
02 flow 300 cc/mm
AR/0 2 flow 150 cc/mm
Furnace temperature (max) 850° C.
Combustion time 2.0 mm. before start of
titration
End point potential 280—300 my.
Oils and Solvents : Oil samples were diluted in pesticide grade
hexane and injected directly onto a prewashed 40 mg carbon column
and immediately placed into the sample entry part of the furnace.
An extra delay of one minute was set on the TOX Analyzer. During
this period Argon gas flowed over the sample and the sample was
slowly Inserted Into the center of the furnace. This was done to
facilitate the evaporation of the hexane solvent and to prevent
explosion of the carbon plug.
Solvent samples not miscible with hexane were either injected
directly or diluted with methanol or ethyl acetate prior to
injection. If these solvent samples were miscible with water, the
samples were Injected directly onto the top of a carbon column and
washed with potassium nitrate solution before the combustiou step.
Solid Samples : Two methods were used to extract sludge and soil
samples. In the first method the solid sample was mixed with
anhydrous sodium sulfate and extracted using 250 ml of a 1:1 (vlv)
mixture of acetone/hexane in a soxhlet extractor f or four hours.
After washing three times with deionized water to remove the
acetone, the extract was concentrated to 10 ml on a steam bath using
a Kuderna—Danish apparatus and Snyder column 7 . This hexane extract
was injected directly onto a washed carbon column and combusted in
the instrument.
5—77
-------
The second extraction procedure was performed using an Ultransonics
Heat Systems sonic probe fitted with a microtip. The sonicator was
set at 50% duty cycle and the output control at 5 to accommodate the
inicrotip. After mixing the solid sample with sodium sulfate, adding
20 ml of hexane, and positioning the microtip in the hexane just
above the solid, the sonicator was pulsed for two minutes.
After the sonication, the hexane was decanted and gently suctioned
through a Buchner funnel fitted with a glass fiber filter and
containing a smail amount of sodium sulfate. Solid samples were
extracted with three successive 20 ml portions of hexane. After the
last sonication, the entire sample was filtered. This residue was
then washed with hexane and the extract was made up to a 100 ml
using bexane. A portion of this extract was then Injected onto a
carbon column and Introduced into the TOX—lO furnace.
Below are other methods referenced in this paper:
1. PUBS in Oils : Oil samples were dissolved in hexane and
eluted through a florisil column with 200 ml of 6%
diethylether in hexane. Five ul were injected into a CC
under the following conditions:
Injector and detector temperatures — 250°C
Oven temperature — 200°C (isothermal)
Carrier — N 2 (60 mi/minute)
Column — 2 a x 4 mm II) with 1.5% SP—2250/l. 95% SP—240l packing
Detector — electron capture
2. Solvent Identification : Samples were dissolved in carbon
disulfide or methanol and injected into a CC under the
following conditions:
Injector and detector temperatures — 250° C
Primary column 2 a r 4 ma ID glass column with 1%
SP—l000 on 60/80 mesh carbopak B
Oven temperature — initial temperature 80°C wIth two minute
hold, ramp at 8°/win, to 220°C, hold for 15 minutes
Carrier — N 2 (80 mi/win)
Detector — flame Ionization detector (FID)
Samples were confirmed on a secondary column.
Secondary column — 60 a x 0.32 mm ID capillary with 1.0 urn
DB—5 film
Injector and detector temperatures — 250°/320°C
Split injection 16:1
Oven temperature — Initial temperature 35°C with 10 minute
hold, ramp at 5°/win, to 300°C
Carrier — He
Detector — FID
5-78
-------
3. Total Halogen in Oil : Halogens in oil samples were
determined gravimetrically as described in ASTM Method D808)- 3 -
4. Combustion/Ion Chromatography Method : Oil samples were
combusted as above in ASTM D808. The combustion gases were
trapped in 10 ml of eluent from the ion chromatograph.
The bomb washings were brought to a final volume of 100 ml.
One hundred microliters of these washings were then
automatically injected into a Dionex 20201 Ion chromatograph
under the following conditions:
Flow: 2.0 mi/minute
Eluent: 2.8 mM NaHCO 3 /2.2mM Na 2 CO 3
Column: ASC—3
Suppressor: Anion membrane suppressor (.025 M H 2 S0 4 )
Chloride was quantitated by peak height using a three level
calibration curve.
5. Volatile Organics Compounds were determined as described in
EPA Method 624.12
6. Acid and Base/Neutral Compounds were determined as described
in EPA Method 625.’
7. Pesticide / PCBs were determined as described in EPA Method
608.14
RESULTS AND DISCUSSION
TOX recovery in aqueous matrices . Table 1 shows data on several TOX
water standards. The 2,4,6 Trichiorophenol standard was a quality
control standard which was determined repeatedly over a period of
about six months. This standard was made fresh daily according to
Method #9020 in EPA SW—846. An average recovery of 96% and a
relative standard deviation (RSD) of 6.7% was obtained. The high
recovery indicates that the 2,4,6 TCP was readily adsorbed onto the
activated carbon and nearly completely converted to the titratable
hydrogen chloride species during the combustion step.
The other standards listed in Table 1 are mixtures of priority
pollutant compounds. Table 2 illustrates the composition of the
mixture of volatile organic halogens. Twenty ul of this standard
were injected into one liter of deionized water to produce the final
TOX concentration of 144 ug Cl/liter. Tables 3 and 4 show the
composition of the base/neutral and pesticide/PCB standards.
The recovery demonstrated on each one of these standards ranged from
60—70%. Loss of recovery might be attributed to any of the
5—79
-------
following: (1) experimental error in making of the standard, (2)
weak adsorption onto the activated carbon and breakthrough of
organic halogens, (3) loss of volatile organic halogens during the
transfer of the carbon column from the loader to the furnace, (4)
chemical decomposition of the organic halogen into a form that was
not adsorbed or eluted during washing, (5) incomplete combustion of
the organic halogen in the furnace to produce a titratable hydrogen
halide.
The latter reason seems to be ruled out because of the higher
recoveries (nearly 90%) shown on each of these three standards when
injected directly into the TOX—lO furnace. (See Table 2 and Table
8.)
A variety of groundwater and waste water spikes are shown in Table
5. In all cases recoveries were very good ( 85%). None of the
sample matrices interfered with the TOX determination.
Application of TOX to TCLP extraction . The Toxicity tharacteristic
Leaching Procedure (TCLP) has been promulgated and finalized by the
EPA. 10 MatLy of the target compounds which are regulated by the TCLP
are halogenated. This is especially true of the volatile organic
compounds which are to be extracted using the zero head space
extractor (ZHE). Table 6 shows spike data on TCLP extraction fluids
spiked with trichioroethene. As seen from the results, the presence
of acetic acid in the extracts (0.1 M) did not adversely effect the
spike recoveries indicating good carbon adsorption of the TCE.
Even though the TCLP requires that only extraction fluid one be used
in the ZHE, extraction fluid two was also spiked. The lower pH of
this fluid did not significantly alter recovery of the TOX.
TOX determinations were used to evaluate the extraction efficiency
and precision of the TCLP using the ZHE on soil spiked with
trichloroethene. The same soil, which had been oven dried and
passed through a fine mesh sieve, was used in each of the TCLP
extractions shown in Table 7. The TCLP extract of the soil showed a
TOX background level of 40 ug Cl/liter. This background was
subtracted from the TOX of each TCLP extract. The extraction
efficiency was obtained by dividing the total TOX in 500 nil of the
TCLP extract (ug c i) by the total TCE as ug Cl added to the 25 g of
soil. Extraction efficiencies of 40—70% were obtained. The soils
spiked at the highest concentration showed consistent recovery of
about 70%. An RSD at the 95% confidence level of 7.1% was estimated
using small number statistics on the TOX of the (replicate) TCLP
leachates.
5—80
-------
SOLVENTS AND WASTE OILS
Oils and solvents which will be burned as fuel or incinerated pose a
potential threat to the environment. Waste oils with a TOX greater
than 1000 ppm cannot be sold as a used fuel.’ 2
TOX recovery in solvents . Table 8 shows TOX recovery data on a
variety of halogenated solvent standards made in methanol and
hexane. The volatile, pesticide, and base/neutral compounds are the
same as those listed in Tables 2, 3, and 4. The volatile standard,
however, was made at a slightly higher concentration (8420 ug
Cl/mi). Good to excellent recoveries were observed on each
standard. The relative standard deviations, which averaged 3.6%,
showed good precision for the technique.
TOX in transformer oils . Table 9 and 10 demonstrate that TOX can be
used to screen for the presence of PCBs in transformer oils. It
must be realized, however, that other organic halogens, such as
chlorinated benzenes often found in transformer oils, will also be
detected as TOX. In Table 9 EPA certified standards of three
different arochlors are shown. Two levels of standards are shown
for the Arochlor 1242 and 1260; three levels are shown for 1254. In
all cases except for the two lowest levels of Arochior 1254, the TOX
is within 5Z of the certified value (expressed as ug Cl).
The detection limit of this TOX method for solvents was established
at 10 ug Cl/mi when 50 ul of hexane were injected directly into the
instrument. Lower detection limits are hindered by the volatile and
explosive nature of hexane. Fifty ul was found to be the maximum
safe injection size.
TOX and PCB determinations of real—world transformer oil samples are
compared in Table 10. A best fit line was calculated after
assigning the total P B value as ug Cl/g to X and the TOX value to
Y. In slope intercept form, the equation of the “best fit” line
was: y = 1.17 x + 18.2. A correlation coefficient of 0.99997 was
also obtained. If the results of these methods were identical, a
slope of one and intercept of zero would be obtained. However, in
this case the slope was greater than one and the intercept was
positive, showing that the TOX method has a positive bias. This
bias might be explained by the presence of other halogenated
organ.ics in the oil or the background from the oil itself. However,
a very good linear correlation exists between the TOX and the total
PCB content of the samples.
TOX vs GC (FID) solvent analysis . Another comparison between TOX
and gas chromatography is shown in Table 11. This table shows data
on eight waste solvent and oil samples which were chromatographed to
identify and quantitate unknown solvents. The conditions for this
chromatography are given above in the experimental section. TOX and
5-81
-------
PCBs were also determined on each of the samples. Generally, a good
correlation exists between the TOX and CC data. Naturally, the
advantage of the GC method is the qualitative identification of the
individual halogenated solvents. The CC method, however, is not as
sensitive as TOX for most samples.
TOX vs total chlorine in waste oils . A final comparison between TOX
and two other combustion methods for determining halogens In fuels
and oils is given in Table 12. The ASTN D808 method is a wet
chemical method and is very tedious. Because it is a gravimetric
method using Ag NO 3 , it is specific to the same halogens (Cl, Br and
I) as TOX. The method is incorrectly titled as total chlorine. The
ASTh method, theoretically, should also determine Inorganic halides
present in oil samples. TOX, on the other hand, does not quantitate
inorganic halides. In fact, if they are present in waste solvent
samples, they will deposit on the walls of the furnace during the
combustion step because of their high boiling point. This is often
evidenced by poor repeatability, lingering coulometric titrations
with slow end points, and erroneously high results. Samples
analyzed after the injection of inorganic halides are often affected.
Inorganic halides are usually not a problem with water insoluble
oils and solvents because they do not appreciably partition into
organic phases. Dilution of oils in hexane also serves to reduce
the inorganic halogens in waste oils and solvents. Inorganic
halides will strongly interfere with TOX if they are present in
water miscible solvents such as some alcohols and acetone and must
be removed by washing with KNO 3 prior to combustion.
The combustion/ion chromatography (IC) method listed In Table 12
only quantitates total chlorine. The oxygen bomb combustion was
identical to the ASTM D808 method except the eluent from the
chromatograph was used to trap the combusted hydrogen chloride.
Of the three methods, the TOX method was the easiest and most rapid
to perform. The IC method and TOX method had comparable
sensitivities of about 100 ug dig. ASTM D808 was about five times
less sensitive.
SOLID SAMPLES
To determine TOX on solid samples such as sludges and soils requires
an extraction step. Inorganic halides often tied up in the solid
matrix make direct combustion of these samples impossible.
Recovery of TOX in solid samples . Two existing extraction
procedures were adapted in this study. Initially, the soxhiet
extraction procedure (Method 13540 SW—846) was used. As shown in
Table 14, good spike recoveries of nonvolatile halogenated compounds
were obtained. The precision of the method was also determined (see
5-82
-------
Table 15) and a RSD of 3.7% was found at the 200 ppm level. The
sensitivity of this method was 10 ppm and could be pushed as low as
one ppm by concentrating the hexane extract to one ml using the KD
apparatus.
This extraction procedure, however, precluded volatile halogens
because of the concentration step and also the washing step used to
remove acetone (see above).
In the second procedure (and the one currently used at Lancaster
Laboratories, Inc.) solid samples were extracted with a sonic probe
(Method 355O SW—846). Spike data on a variety of solid samples is
given in Table 16. Spike recovery ranged from 56—112% with an
average recovery of nearly 84%. Soil I was oven dried and sieved.
This soil was stored in our laboratory and used for spiking as part
of our quality assurance program. Five different soil samples (sent
to our laboratory) were spiked at the same level 2030 ug dig using
1,2,4 trichlorobenzene. An average spike recovery of 88% was
obtained and a RSD of 4.7% was found, showing good repeatability.
The other samples shown in the table were also real—world samples
sent to our laboratory. These were all spiked with 2,4,6
trichiorophenol in methanol at about 2000 ug dig. Only sample
1ll62542 showed poor recovery. This sample was sent to our
laboratory as an unknown solid and resembled charcoal. This
material may have retained the 2,4,6 TCP spike preventing recovery.
As might be expected, the TCE soil spike showed lower than average
recovery. Some of the volatile TCE was most likely lost during the
extraction procedure and when the carbon column was injected with
TCE prior to being pushed into the furnace.
TOX vs. priority pollutant compounds (GC/MS) . To evaluate further
the effectiveness of TOX as a screening parameter, a comparison
between TOX and organic halogens from the priority pollutant list
(expressed as c i) was made (see Table 17). The volatile, acid and
base/neutral compounds were determined by GC/NS. The pesticide/PCBs
were determined by gas chromatography.
One must realize that only a limited number of samples are shown on
this table and erroneous conclusions can be drawn from a small set
of data. More comparison data needs to be accumulated to concretely
prove any relation between TOX and priority pollutant data.
Nonetheless, a trend is shown in Table 17. TOX in most
environmental samples is generally higher than accounted for by
GC/MS priority pollutant data. This trend is also supported by the
author’s experience with TOX. In very few instances have
halogenated priority compounds been found when TOX is not detected.
When TOX has been found, however, it has not always been accounted
for by priority pollutant data. This is especially true of complex
solid samples such as the municipal sewage sludges CS1, MB1, and S22
found in Table 17. Note that TOX levels of 230, 63, and 35 ug dig
5-83
-------
were found in these samples, respectively. Yet, the CC/MS priority
pollutant data could not account for any significant portion of this
organic halogen.
SUMMIJRY AND CONCLUSIONS
Although TOX cannot be used to specifically identify halogenated
organic compounds, TOX has been shown to be an effective analytical
tool in screening environmental samples. The method is less time
consuming and, therefore, less expensive than more complex
analytical techniques such as CC and CC/MS. It has been
successfully applied to aqueous, organic and solid matrices. Good
spike recovery for a variety of halogenated organic compounds was
found in all of these sample matrices. Relative standard deviations
of about 5% were determined on each matrix. Interferences to TOX
were minimal and could be eliminated with proper precautions. In
most aqueous samples the detection limit of TOX is in the low ug
Cl/liter range when volumes of 100 ml or greater are loaded. This
advantage of concentrating the organic halogen in the carbon column,
however, is lost with organic and solid samples. In solids and
oils, a detection limit of 10 ppm was established. This detection
limit was limited by the amount of hexane which could be safely
injected into the combustion furnace.
Good correlation was seen between TOX and a variety of gas
chromatography methods. These included methods for the
determination of PcBs in oils with an electron capture detector and
the determination of organic solvents in waste oils and solvents
using a flame Ionization detector. Comparison was also made to two
methods for determining total chlorine in oils using a Parr oxygen
bomb combustion. The ASTh D808 method was found to be more
laborious and less sensitive than TOX. The combustion/IC method was
found to be about as sensitive s TOX and showed equal precision
(RSD about 3%). It, however, was more time consuming. The TOX
method showed a higher mean than the combustion/IC method on
multiple determinations of a waste oil sample.
A general trend showing that TOX was generally higher than accounted
for by GC/NS priority pollutant data was also reported. TOX is
nonspecific In determining organic halogens and, therefore, would
not be limited to detecting only priority pollutant compounds and
compounds with chemical and physical properties making them suitable
for gas chromatography.
5-84
-------
Table 1.
Tox Water Standards
Prepared RSD
Standard conc.* s* ___ Recovery
2,4,6 TCP 28 100 96 64 6.7 96
Volatiles 5 144 99 10.5 11.0 69
Base/neutrals 4 115 74 4.4 5.9 64
Pesticides 4 104 73 9.4 12.9 70
All standards were prepared in deionized reagent water sample
Volume — 100 ml
Loading rate — 3.3 mi/minute
* ug Cl/liter
5-85
-------
Table 2.
Volatile Organic Cocktail
% Halogen
Compound ug/ml as Cl as ug Cl/ml
methylene chloride 664 83.5 554.
1,1 dichioroethene 520 73.1 380.
1,1 dichioroethane 588 71.6 421.
trans 1,2 dichioroethene 620 73.1 453.
chloroform 628 89.1 559.
1,2 dich]oroethane 744 71.6 533.
1,1,1 trichioroethane 628 79.7 500.
carbon tetrachioride 596 92.2 549.
dichlorobromomethane 672 64.8 436.
1,2 dichioropropane 740 62.7 464.
trichioroethene 652 80.9 528.
dibromochloromethane 776 51.0 395.
2 chlorovinyl ether 668 33.6 224.
bromoform 392 42.0 374.
tetrachioroethane 692 84.5 584.
ch lorobenzene 768 31.5 242 .
Total 7196.
Standard made in methanol
Direct Injection :
Injection Conc. of Std. inount as Average
SizeJuij ( ug Cl/ui) Cl (ug) X (ug) % Recovery
5. 7200. 36. 2. 31. 86.
5—86
-------
Table 3.
PESTICIDE/PCB COCRTAIL
Compound
ug/m l
ug C1/inJ
PCB 1248
500.
240.
PCB 1254
500.
270.
Heptachior
500.
332.
BHC
250.
183.
PCB 1221
500.
105.
4,4’ DDD
500.
240.
Heptaciflor
Epoxide
500.
319.
Toxaphene
500.
340.
Endrin
500.
344.
4,4’ DDT
500.
250.
TOTAL
2620.
-------
Table 4.
BASE/NEUTRAL COCI TAIL
Compound ug/mi ug Cl/rn ]
4 bromophenyl phenyl ether 833. 118.
3,3’ dichlorobenzidifle 833. 231.
Hexac hloropefltadiefle 833. 648.
1,2,4 trichlorobeflZefle 833. 488.
Hexachiorobutadiefle 833.
TOTAL 2310.
-------
Table 5.
TOX Aqueous Matrix Spikes
Sample TOX-unspiked Spike TOX-spiked Recovery
No. Samp le* Added* Salnp le*
1072944 25 100 128 103
1075051 254 100 349 95
1074785 5 100 98 98
1084426 22 50 65 86
1045111 7 100 104 97
1029675 5 100 98 98
Spiked with 2,4,6 TCP
* ug Cl/i
Table 6.
TCLP Extract Matrix Spikes
Spike Conc. Ave. Spike Ave. %
( ugh) Recovered Recovery
Extraction Fluid #1 4.98 3 150 139 93
Extraction Fluid #2 2.88 2 150 146 97
Spiked with trichioroethene in methanol.
5-89
-------
Table 7.
TCLP of Soil Spiked with TCE
Conc, of TCE TOX of
Added to Soil TCLP Extract % TCE
( ug dig) ug Cl/liter Extracted
Soil Sample 0. 40
(background)
Soil Sample Ia 99. 3540 72.
Soil Sample lb 99. 3620 74.
Soil Sample lc 99. 3200 65.
Soil Sample 4 1.0 71 62
Soil Sample 5 1.0 62 44
25 g soil spiked and extracted
Final volume of ! CLP extract - 500 al
5-90
-------
Table 8.
Cjfl
TOX Solvent Standards by Direct Injection
Hexane was the solvent for chloroform
Methanol was the solvent in all other
and bromoform.
cases.
Compound(s)
N
Injected
(ug Cl)
Recovered
(ug Cl)
RSD
%
Ave. %
Rec.
2,4,6—TCP
15
5.00
5.04
3.6
101.
Vrolatiles
6
25.3
24.2
3.4
96.
chloroform
6
22.5
22.0
6.1
98.
Bromoform
6
9.80
9.76
2.0
100.
Pesticides
3
13.1
12.2
93.
3ase/Neutra ls
2
23.1
20.6
——
89.
CB 1242
7
21.1
21.1
3.0
100.
-------
Table 9.
0
TOX On EPA Transformer Oil PCB Standards
Aroclor
Calculated
Measured
Percent
Recoyer
1242
24.0
239.
22.
249.
92.
104.
1254
6.5
27.0
270.
11.
30.
256.
170.
lii.
95.
1260
30.0
341.
31.
331.
103.
99.
TOX values are given in ug dig.
-------
Table 10.
PCBs In Transformer Oils
EC/GC vs. TOX
Sample
G.C.
TOX
373305
128.
145.
373306
125.
143.
373307
122.
153.
377852
4.1
11.
378095
6.0
9.
379703
527.
637.
381633
8200.
9600.
381778
41.2
99.
381779
36.3
104.
( fl
( J
Values are given in ug C1/g.
-------
Table 11.
to
TOX vs GAS CEROKATOGRAPHY
(Waste Solvents and Oils)
Sample
TOX
Compounds
Identified
GC/ECD (?CBs
-
Solvent
2.0
2.0
Carbon
Tetrachioride
——
Solvent
10.8
11.
Xethy lene
Chloride
1,1,1 Tn—
chioroethane
—-
Solvent
27.0
30.1
Tetrachioro—
ethylene
ND
Solvent
<0.1
ND
——
ND
011
0.68
0.8
1,1,1 Tn—
chloroethane
ND
Oil
1.3
1.4
1,1,1 Tn—
chioroethane
ND
011
0.03
ND
——
.019
Oil
<0.01
ND
——
.0019
All values are given as percent halogen by weight.
-------
Table 12.
Comparison of Methods - Halogens in Waste Oils
Total Chlorine
Sample TOX Comb./IC ASTM D808
Standard I 1560 1620 1200
1145917 7950 a 6440 b 6700
1159392 1040 960 600
1159614 570 860 <500
1159615 4970 5250 5200
1160952 240 230 <500
1160953 1150 1120 960
All results as ug dIg
Standard I (1680 ug C1/g) - TCE in mineral oil
a Average 6 trials; s=190
b Average 5 trials; s=220
5—95
-------
Table 13.
TOX SPIKE RECOVKRY IN SLUDGE BY SOYRT.RT EXTRACTION
Sazple
Spike
Waterial
Spike Added
(ug C1/ci
Spike
Recovered
(ug ci/gi
% Recover
CS1
CS].
CS].
S31
2,4,6 T P
2,4,6 TCP
PCB 1242
2,4,.6 TCP
241
240
194
61
214
258
190
61
89
108
98
100
0
-------
‘C
Table 14..
PRECISION DATA FOR TOX IN SLUDGES
Priil
TOX (trnm
CS1
238
CS2
224
=
233 pp
CS3
245
S =
8.5
CS4
229
RSD
= 3.7%
CS5
228
-------
Table 15.
TOX Spikes of Solid Samples Sonic Probe Extractions
Ave.
Sample Spiking Spike Spike Recovery
Matrix N Compound Added* Recovered* _______
Soil I 3 TCE 510 380 75
Soil I 1 PCB—1260 290 240 83
Soil I 2 1,2 DCB 500 560 112
Natural Sousa 5 1,2,4 TCB 2030 1780 88
Carbonaceous 1 2,4,6 TCP 1650 920 56
Sludge (1162542)
Paint Sludge 1 2,4,6 TCP 2050 1920 94
(1159794)
Industrial 1 2,4,6 TCP 1940 1550 80
Sludge (1139812)
Industrial 1 2,4,6 TCP 1940 1620 84
Sludge (1134427)
* Units — ug Cl/g
a Five different soil samples — each extracted once
TCE - trichloroethene
DCB - dichlorobenzene
TCB — trichlorobenzene
TCP - trichiorophenol
N — number of determinations
5-98
-------
Table 16.
TOX vs. GC/MS Priority Pollutants
Priority Halogenated
Sample Pollutant Compounds Compounds
Type TOX Halogen Tested Detected
GW i 2100 a 500 a V,A,B,P Vinyl chloride
trans 1,2 dichioroethene
trichioroethene
2,4, 6 trichiorophenol
GW 2 767 a V,A,B,P Ch lorobenzene
trans 1,2 dichloroethene
trichioroetherie
WW 3 290 a V,A,B,P Chloroform
TCLP 3200 a 4190 a V Chloroform
1,2 dichioroethane
methylene chloride
1,1,1 trichloroethane
1,1,2 trichioroethane
tn chioroethylene
tetra chioroethene
GW 3 V methylene chloride
1, 1,1 trichioroethane
SW-CS1 230 b NDb V,A,B,P ND
SW-MB1 63k’ ND ’ V,A,B,P ND
SW-S22 35 b NDb A,B,P (some P; Tot
-------
REFERENCES
1. Test Methods for Evaluating Solid Waste , Method 9020, EPA
SW—846, 1986.
2. Stevens, A. A., Dressinan, R. C., Sorrell, R. K. and Brass, H. J.
Journal of American Water Works Assoc. , 1985, 77, Pp. 146—154.
3. TOX—lO Mitsubishi Chem. Lid. Ltd., Cosa Instruments, Norwood, NJ.
4. DX—20 Xertex Dobrmann, Santa Clara, CA.
5. Takahashi, Y., Moore, T. T., and Joyce, R. J., “Measurement of
Total Organic Halides (TOX) and Purgeable Halides in Water
Using Carbon Adsorption and Microcoulometric Determination,”
Chemistry in Water Reuse , Vol. 2, Ann Arbor Science, 1981.
6. Berger, D. L., Anal. Chem . 1984, 56, 2271—2.
7. Method 3540, EPA SW—846.
8. Method 3550, EPA SW—846.
9. Total Organic Halogens in Waste Materials, New York State
Procedure, January, 1985.
10. Federal Register, June 13, 1986, Vol. 51, PP. 21648—21693.
11. ASTh D808—81, American Society for Testing and Materials , 5.01,
p. 357, 1983.
12. US EPA 40 CFR. Part 136, Method 624, October 26, 1984.
13. US EPA 40 CFR Part 136, Method 625, October 26, 1984.
14. US EPA 40 CFR Part 136, Method 608, October 26, 1984.
15. Federal Register, June 13, 1986, Vol. 51, pp. 21648—26193.
16. Federal Register, Nov. 7, 1986, Vol. 51, pp. 40638—40654.
17. Federal Register, Nov. 29, 1985, Vol. 50, p. 49164.
5-100
-------
PYROLYS IS / M ICROCOULOMETRIC DETERM I NAT ION OF
TOTAL ORGANIC HALIDES IN SOLIDS AND OILY WASTES
Van T. Lieu and Van H. Woo
California State University Brown and Caidwell Laboratories
Long Beach, CA 90840 Pasadena, CA 91105
ABSTRACT
Two pyrolysis/microcoulometric methods for the determination of total
organic halides in solids and oily waste are described. The first method
involves the washing of solid smaple in a column packed with granular
activated carbon with a dilute nitrate solution to remove interfering
inorganic halides. The organic halides are subsequently determined by
pyrolysis/microcoulometric measurement of the sample solid-activated
carbon mixture. The second method involves the extraction of solid or oily
waste samples with a mixture of octanol and 1 M sulfuric acid. The organic
halides are extracted into the octanol phase and then analyzed by
pyrolysis/microcoulometry. The inorganic halides are removed by
extraction into the aqueous phase. The second method has the advantage of
extending the technique to the analysis of a broader spectrum of organic
halides including polar compounds such as chioroacetic acid.
5—101
-------
iNTRODUCTiON
Organic halides represent one of the most important groups of
compounds. Not only these compounds are generally toxic and/or
carcinogenic, but thy are also the most commonly occuring enviromental
pollutants. Pesticides and chlorinated organic solvents were once thought to
be the primary source of organic halide pollution. However, studies have
shown that chlorination of drinking water and wastewater produces
significant amount of these compounds( 1).
Because many different halides with widely varying characterics can
occur in a sample, it is not practical, if not impossible, to analyze all these
• compounds individually, it has been estimated that only 10 to 25% of these
compounds are amenable to individual analysis(2-3). An effective method
for measuring the amount of total organic halides could provide valuable
information on the occurance of organic halide contaminants. Several
workers(4-8) have employed pyrolysis/microcoulometry as a means of
measuring of organic halides in water. The technique involves isolation,
conversion of organic halides into inorganic halides and their subsequent
quantitative determination by microcoulometric measurement. The methods
used for the isolation of organic halides includes solvent extraction, gas
purging and carbon adsorption; the latter is the method presently in use in
the U.S. EPA Methods 450.1 and 9020 (9) for determining total organic
halides in waters. A semiautomatic system for carbon adsorption, pyrolysis
and microcoulometric dtermination has been developed and is commercially
available( 10). The primary advantages of this technique are ease of
operation, good sensitivity and low background levels. In this study, two
pyrolysis/microcoulometric methods for the determination of tatal organic
5-102
-------
halides in solids and oily wastes are described. The first method involves
the washing of solid samples in a column packed with granular activated
carbon with a dilute nitrate, solution to remove interfering inorganic halides.
The second method involves the extraction of solid or oily waste samples
with a mixture of octanol and 1 M sulfuric acid.
EXPERIMENTAL
Apparatus and Material
The Dohrman DX-20 Total Organic Halide Analyzer was used for the
analysis of total organic Halides. Two modes of operation of the analyzer
were employed. The boat loading mode was used for analysis of solid
samples by Nitrate Wash-Carbon Adsorption Method and the direct injection
mode was use for analysis of solid and oily waste samples by Solvent
Extraction Method.
A schematic of the Nitrate Wash-Carbon Adsorption apparatus is shown
in Figure 1. The apparatus was used for washing of solid samples to remove
interfering inorganic halides prior to pyrolysis/coulometric measurement.
All chemical reagents used were reagent grade obtained from Baker
Chemical or Aldrich Chemical and were used without further purification.
High purity deionized water was obtained form Sparkletts Water Company.
The granular activated carbons was Filtrasorb 400(Calgon Corp.. Pittsburgh.
PA) screened to a 100/200 mesh range. The columns for containing the
granular activated carbon and solid sample were 5 cm section of 6 mm, O.D.
Pyrex tubing with a 2 mm I.D. Since activated carbon adsorbs organic vapor
from air quite readily, the carbon was kept in a clean glass bottle with a air
5—103
-------
tight cap. Precautions were taken to minimize the time of exposure of the
carbon to air during the packing of columns with granular activated carbon.
Standard solutions of different halogenated organic compounds were
prepared by dissolving the compound of interest in ethyl acetate. Standard
soil samples of the different halogenated organic compounds were prepared.
Each soil sample was prepared by adding known volume of the standard
solution of interest to an accurately weighed out amount of garden soil which
had been dried and passed through a 30-mesh sieve. The soil sample was
then shaken vigorously and was allowed to equilibrate at least overnight
before use.
Solid Samole Analysis by Nitrate Wash-Carbon Adsorption Method
The glass columns ‘were cleaned with isopropyl alcohol, rinsed with
deionized water and dried. A Cerafelt(Johns-Manville, NY) plug was
inserted in one end of the glass column. The first column was half filled with
granular activated carbon and weighed. The remaining space was filled with
sample solid and reweighed to obtain the weight of sample used. A second
plug was used to cap the end. The second column was completely filled with
granular activated carbon and capped. The two columns were connected in
series as depicted in Figure 1. A 500 ppm nitrate wash solution under 65
psig nitrogen was used to wash the columns-in-series. After the desired
volume of the nitrate wash was eluted, typically 2 ml, the columns were
disconnected for the determination of halogen content in each column.
Several samples could be processed simultaneously.
In the pyrolysis-determination step, the solid content from each column
was extruded with the use of a plunger and placed directly onto a sampling
boat at the inlet of the Dohr man DX-20 analyzer and the analysis cycle was
5-104
-------
initiated. The sampling boat first moved into a 2500 C vaporization zone,
then through a 8000C combustion zone where all organics were oxidized and
halides converted to hydrogen halides. The gases then flowed into the
microcoulometric titration cell and determined quantitatively by titration
with silver ions generated coulonietrically.
Solid and Oily Waste Sample Analysis by Solvent Extraction
Accurately weighed out approximately one gram solid sample or
delivered from a microsyringe a suitable volume, typically 25 to 100 p1 of
oily waste into a 14 nil glass vial containing 1 ml of I M sulfuric acid and 5
ml of octanol.. The vial was shaken vigorously for one minute to extract the
organic halides into the organic phase and then centrifuged for 10 minutes at
1000 G. A 25 to 50 il aliquot of the octanol phase was withdrawn into a
microsyringe and injected directly into the Analyzer set up for direct
injection mode for determination of the halide content.
RESULTS AND DISCUSSION
Spud Sample Analysis by Nitrate Wash-Carbon Adsorption Method
The first method developed for the measurement of total organic halides
in solids was a modification of the EPA Method 9020 for determining total
organic halides in water samples. In the method developed, the solid sample
was washed with a 500 ppm nitrate solution to remove any interfering
inorganic halides that might be present. Any organic halides eluted from the
solid sample were trapped by the granular activated carbon. The use of two
columns in series ensured the complete adsorption of the organic halides
eluted. The results are given in Table 1. The recoveries for relatively non-
5-105
-------
volatile and non-polar compounds ranges from 72.3% to 93.0%. The low
recoveries of chloroform and chioroacetone may be ascribed to loss by
vaporization during sample preparation. Poor recoveries were also obtained
for polar compounds. Study of recovery of organic halides from water using
the EPA method showed similar poor results for polar compounds(l0). That
study showed that chioroethanol had about 20% recovery while chioroacetic
acid showed 0% recovery. The low recoveries may be ascribed to the poor
adsoprtion characteristic of the polar compounds on granular carbon. The
result of this study show a 52.1% and 58.3% recoveries for chioroethanol and
chioroacetic acid respectively. The higher results were probably due to the
incomplete desorption of the organic halides from the soil during the nitrate
wash.
Solid Sample Analysis by Solvent Extraction
A method for the determination of organic halides in solid has been
developed by Riggin et. al.(7) and involved the use ethyl acetate-water
mixture which extrtacted the organic halides into ethyl acetate phase for
pyrolysis/ inicrocoulometric analysis. Any inorganic halides that may be
present were removed into the aqueous phase. To evaluate the applicability
of the method, a broad spectrum of organic halides of different known
concentrations were analyzed using ethyl acetate-water mixture for
extraction. To 1 gram of standard soil sample in a 14 ml glass vial were
added 5 ml of ethyl acetate and I ml of distilled water. The sample was
shaken vigorously for 1 minute and then centrifuged for 10 minutes at 1,000
G. A 25 to 50 )JI aliquot of the ethyl acetate was analyzed by the Analyzer
set up for direct injection mode. The results are shown in Table 2. With the
exception of broinoform and 1 -iodooctane, the generally non-polar
5-106
-------
compounds had good recoveries ranging from 86.2 to 129.5%. Bromoform
and 1 -iodooctane showed slightly lower recoveries and were probably
because of the formation of bromine and other undetectable bromine and
iodine compounds(S). Polar compounds such as chioroacetic acid shows
very poor recoveries. In general, the more polar compounds had a lower
overall recoveries. The low recoveries may be ascribed to the loss of the
organic halides during the extraction step of the analysis, because the more
polar organic compounds were more soluble in water and was partially
extracted into the aqueous phase. Equilibrium calculations indicate that the
majority of the chioroacetic acid in water dissociated into chioroacetate and
hydrogen ions.
C ICH 2 COOH CICH 2 COO +
The acid dissociation constant for this reaction is 1.38 x 10-3. This being the
case, the chioroacetate ion is more soluble in the aqueous phase than in the
organic phase, therefore giving low recoveries for this compound.
One possible means of lowering the solubility of chloroacetic acid and
other polar compounds in water is to shift the dissociation equilibrium by
adding an excess of hydrogen ions to the aqueous phase. However, ethyl
acetate was found to be misible with 1 M sulfuric acid due to an acid
catalyzed hydrolysis of the ethyl acetate(1l). Instead of having two
irnmisible layers, there was only one layer of solution.
It was decided to try another slightly polar solvent which is immisible
with water and do not undergo hydrolysis under acidic conditions. A
mixture of 1 -octanol and 1 M sulfuric acid was chosen for use in extraction.
Unlike the ethyl acetate-sulfuric acid mixture, the octanol and sulfuric acid
existed in two separate layers. The results are shown in Table 3. As
expected, the non-polar compounds had good recoveries, such as 1-
5-107
-------
chloronaphthalene which had 97.9% recovery. The recovery of chioroacetic
acid improved substantially from the ethyl acetate-water method from 1.2%
to 83.9%. All other compounds less polar than chioroacetic acid showed
similar improvements in recoveries.
Oily Waste Analysis
For the purpose of this study, oily waste is qualitativelty defined as an
organic liquid or an emulsion of organic liquid and water with varying
degrees of volitilities and viscosities. Since oily waste sample is in the liquid
state, it was thought that a direct analysis of the oil by syringe injectioon
into the pyrolysis tube would be possible. This was found to be impractical
because sample with relatively high viscosity such as oils could not be
quantitatively injected into the pyrolysis tube. More important, oily waste
samples frequently contain water in the form of emulsion which usually
contains significant amount of interfering inorganic halides. In view of of
the difficulties with direct analysis, it was decided to analyze oily waste
sample using the octanol- I M sulfuric acid extraction procedure described
earlier for solid. This procedure should eliminate the interference due to
inorganic halides by rejecting them into the aqueous phase. Dilution of the
sample liquid with octanol during the extraction procedure should lower the
viscosity and should result in a solution being relatively easy to inject.
Standard oily waste samples containing various organic halide compounds
were prepared. Each sample was prepared by pipetting known volume of
standard organic halide solution of interest into known volume of toluene. A
portion of this oily waste was weighed and was added to a vial which
contained 5 ml of octanol and 1 ml 1 M sulfuric acid. The vial was shaken
vigorously for one minute and then centrifuged at approximately 1,000 G for
5-108
-------
10 minutes. An aliquot of the octanol phase was then analyzed by direct
injection with the analyzer set at injection mode. The results obtained are
shown in Table 4 and are similar to those from solids by extraction with
octanol-l M sulfuric acid. For non-polar compounds, the average recoveries
for I -chloronaphthalene and 1 ,2,4-trichlorobenzene are 83.4 and 111.2%
respectively. The more polar chioroacetic acid also showed good recovery,
giving a value of 81.6%. The recovery for chloroform is low and is probably
due to loss by volatilization.
To study the possible interference due to the presence of inorganic
halides, a standard oily waste sample containing inorganic chloride was
prepared by adding 6.276g of sodium chloride to a mixture made up of 30
ml water, 10 ml methanol, 200 ml toluence and 1.0 ml of trition X- 100, an
emulsifier. Three layers were obtained from the mixture: octanol, emulsion
and water layers. Portions of each layer were weighed and extracted with
octanol- 1 M sulfuric acid mixture in the same manner as described above.
The halide contents of each of the layers were then analyzed by direct
injection of the octanol phase. No halide was found in the octanol phase of
any of the three layers. The study showed that the presence of inorganic
halides in oily waste sample does not interfere with analysis by octanol- 1 M
sulfuric acid extraction procedure.
CONCLUSIONS
Two methods involving pyrolysis/microcoulonietric determination for the
measurement of organic halides in solids and oily waste were developed.
The first method involving nitrate wash-carbon adsorption is a modification
of the U.S. EPA method for the anlysis of total organic halides in water and is
5-109
-------
developed for the analysis of solids. This method offers an alternative
means for the analysis of organic halides in solids The second method
involving extraction with octanol- I M sulfuric acid mixture and is applicable
to solid and oily waste samples. For both types of samples, the organic
halides are separated from the inorganic halides and are subsequently
pyrolyzed and microcoulometrically analyzed. The second method extends
the analysis to a broader spectrum of organic halides including polar
compounds such as chloroacetic acid.
The pyrolysis/microcoulometric technique using a Dofirman DX-20
analyzer has been shown to provide an effective means for screening and
overall determination of organic halides. It is sensitive, relatively fast and
easy to use. However, this technique has its limitations. Since a sample can
contain a wide variety of compounds or mixture of compounds with different
efficiencies of recovery, no suitable standard can be used for calibration. As
a result, the accuracy of analysis is expected to be poorer than when a
spiked sample for a single known compound is analyzed.
5—110
-------
REFERENCES
1. Water Chlorination Envirotnental I tnpact and Health Effects, in
Proceedings of the Conference on the Enviromental Impact of Water
Chlorination; Jolley, R. L., Ed., Ann Arbor Science, Ann Arbor, MI, 1978, Vol. 1.
2. Dress man, R. C. and Stevens, A. A., The Analysis of Organo-ha [ ides in
Water- An Evaluation Update, J. Am. Water Works Assoc., 74:431 ,(l 983).
3. Kuhn, W. and Clifford, D., Experience with Specific Organic Analysis for
Water Quality Control in West Germany, in Proceedings of Water Quality
Technology Conference, Houston, TX, 1985.
4. Wegman, R. C. C. and Greve, P. A., The Microcoulometric Determination of
Extractable Organic Halogen in Surface Water, Application to Surface Waters
of the Netherlands., Sci. Total Environ. 7:235, (1977).
5. Glaze, W. H., Peyton, R. P. and Rawley, R.,Total Organic Halogen as Water
Quality Parameter: Adsorption/Microcoulometric Method, Environ. Sci.
Techol., 11:685, (1977).
6. Dressman, R. C., McFarren, E. F. and Symons, J. M., An Evaluation of the
Determination of Total Organic Chlorine (TOCL) in Water by Adsorption onto
Ground Granular Activated Carbon, Pyrohydrolysis and Chloride-Ion
Measurement, J. Am. Water Works Assoc. Technology Conference
Proceedings ,1977.
7. Riggin, R. M., Lucas, S. V., Lathouse, J., Jungclaus, G. A. and Wensky, A. K.,
Development and Evaluation of Methods for Total Organic Halide and
Purgeable Organic Halide in Wastewater, US EPA document 600/4-84-
008,1984.
8. Takahashi, Y., Moore, R. 1. and Joyce, R. J., Measurement of Total Organic
Halides (lOX) and Purgeable Organic Halides (POX) in Water Using Carbon
5-111
-------
Adsorption and Microcoulometric Determination, in Chemistry in Water
Reuse. Vol. 2, Cooper, W. J., Ed., Ann Arbor Science Publishers, Inc., 1981.
9. Total Organic Halides, in Test Methods for Evaluating Solid Wastes, US
EPA SW-846, Second Ed., 1982.
10. Dohrmann DX-20 Total Organic Halide Analyzer, Xertex Corp., Dohrmann
Division, Santa Clara, Ca.
11. Streitweiser, A. Jr., Heatticock, C. IL, Introduction to Organic Chemistry.
Second Ed., Macmillan Publishing, New York, 1981. p.539.
5—112
-------
Figure 1. Apparatus for the Recovery of Halogen Compounds in Solid
Samples by Nitrate Wash—Carbon Adsorption
M trogen
Pressure
Sarr Ie Reservoir
lifted with Nitrate
Wash Solution
First
Coltwnn
Second
Column
-------
Table 1. Recovery of Halogen Compounds in Soils by Nitrate Wash—
Carbon Adsorption Total Organic Halide Determination
COMPOUND
1—chloronaphthalene
1,2, 4—trichlorobenzene
1 ,2,4—trlchlorobenzene
bromocyclohexane
chloroform
2,4, 6_trichioropheflOl
2,4, 6 -trichioropheflOl
chioroace tone
1 chloropropaflOiC acid
1_chloroethanol
1 chloroacetiC acid
Amount Present (iig/g)
as Cl
Calculated Found*
786 731
810 688
766 572
814 589
785 341
795 575
786 618
775 121
792 17
689 359
807 470
Number
of Runs
3
3
5
3
1
4
4
3
3
3
3
Average
Recovery
( % )
93.0
84.9
74.7
72.3
43.5
72.3
78.5
15.6
2.1
52.1
58.3
Standard
Deviation
( % )
11.0
13.7
6.7
12.8
7.4
5.8
9.5
4.4
17.7
14.5
-4
*Cor,ICted for the 42pgCt/$SdI Odgfn yPrUSI t
-------
Table 2. Recovery of Halogen Compounds in Garden Soils by
Extraction with Ethyl Acetate and Water
COMPOUND
chloronaphthalene
chloronaphthalene
Amount Present
as Cl
(ug/g)
Number
of Runs
3
2
Average
Recovery
(Z)
104.0
129.5
Standard
Deviation
(%)
6.0
4.8
Calculated
Found*
10,464
972
10,570
815
p—dichlorobenzene
p—dichlorobenzene
13,600
833
13,816
993
2
2
101.5
119.2
0.3
1.0
1,2,4—trichlorobenzene
1,2,4—trichlorobenzene
10,270
1,842
8,975
1,588
2
2
87.3
86.2
1.3
11.4
bromocyclohexane
814
734
2
90.1
5.2
1—iodooctane
812
643
2
79.2
31
bromoform
bromoform
5,976
863
4,326
771
2
2
72.4
89.3
2.6
0.3
1—chioropropanoic
acid 792
462
2
58.2
6.3
1 —chioroacetic acid 13,300
184
6
1.2
0.6
LC)
Corrected for the 42pgCf/g Soil Orlgtn iy Present
-------
Table 3. Recovery of Halogen Compounds in Soils by
Extraction with Octanol—1.OOM H 2 S0 4
COMPOUND
1—chloronaphthalene
1—lodooctane
bromoc yclohexane
2,4, 6—trichioro phenol
1—chioropropanoic acid
1—chioroacetic acid
Amount Present (Alg/g)
88 Cl Number
Calculated Found* of Runs
786 770 3
812 635 3
814 707 4
811 749 3
792 574 4
807 677 5
Average Standard
Recovery Deviation
( %) ( % )
97.9 9.7
78.2 4.8
86.8 5.2
92.4 6.8
79.5 5.3
83.9 2.6
0
L
Corr.ctsd for the 42aiiCf/g Soil Odgtn ty Present
-------
Table • LI..
Recovery of Halogen Compounds in Oily Waste by
Extraction wtih Octanol—1M 11 2 S0 4
COMPOUND
1—chloronaphthalene
1,2, 4—trichlorobenzene
chloroform
1—chioroacetic acid
Amount Present ( ig/ml)
as Cl Number
Calculated Found of Runs
31,200 26,021 3
614 683 3
11,700 7,122 2
36,000 29,376 3
Average
Recovery
( % )
83.4
111.2
60.9
81.6
Standard
Deviation
( 7 )
16.1
18.9
6.2
11.1
-------
DEVELOPMENT AND EVALUATION OF TEST METHODS FOR ‘IOTAL CHLORINE
IN USED OILS AND OIL FUELS
Alvia Gaskill, Eva D. Estes, David L. Hardison, Research Triangle
Institute, Researrch Triangle Park, North Carolina; Paul H. Friedman,
Office of Solid Waste, U.S. Environmental Protection Agency,
Washington, D.C.
ABSTRACT
A current EPA regulation prohibits the sale for burning in
non—industrial boilers of used oils and oil fuels contaminated above
specified levels with certain metals and total chlorine. When burned
as fuel in a small boiler, the contaminants may be emitted to the
ambient air at hazardous levels. This regulation establishes a
rebuttable presumption that used oil containing more than 1,000 ppm
total chlorine has been mixed with halogenated solvents and is a
hazardous waste. Rebutting the presumption requires the seller of oil
to prove that this chlorine is not due to halogenated solvents or
other hazardous halogenated organics. If the rebuttal is successful,
the oil can be sold as fuel up to a level of 4,000 ppm total chlorine.
Analytical techniques for determination of total chlorine were
evaluated or developed to provide regulatory agencies and the
regulated community with appropriate chlorine test methods. The
techniques evaluated included wet chemical titrations following oxygen
bomb combustion, disposable test kits, instrumental microcoulemetry,
and x—ray fluorescence spectrometry.
These candidate techniques were subjected to interlaboratory testing
to estimate their precision, accuracy, sensitivity, and susceptibility
to matrix effects. Information on ease of use and analysis costs was
also collected. Based on this study, test methods were prepared for
the most promising techniques and subjected to a formal collaborative
study to generate precision and accuracy data for each method. These
methods are to be proposed in the Federal Register as mandatory for
compliance with the existing used oil regulation.
5—119
-------
MICRC AVE ACID SAMPLE DECOMPOSITION FOR ELEMENTAL AN1 LYSIS
H.M. Kingston, Research Chemist, L.B. Jassie, Center for Analytical
Chemistry, Inorganic Analytical Research Division, National Bureau of
Standards, Gaithersburg, Maryland
ABSTRACT
Sample preparation is an essential step in achieving both accuracy and
precision in analysis of materials. It is also one of the most time
consuming portions of many analyses and has become the limiting step
for such multi—element techniques as ICP, XRF, and ICP—MS. Acid
digestion of biological and botanical samples can take from 4 to 48
hours using classical digestion techniques. Microwave acid digestion
requires only 10 to 15 minutes for the decomposition of many
biological and botanical samples, dramatically reducing preparation
time. Additionally, volatile elements such as selenium, phosphorus,
tellerium, vanadium and others can be quantitatively retained using
microwave decomposition in a sealed vessel prior to instrumental
analysis. To use the technique, however, its necessary to have an
understanding of the fundamental concepts controlling interactions
between microwave energy and acid sample molecules.
During the past several years, research has been conducted to
determine these fundamental relationships and to develop methods that
allow the analyst to predict the conditions that will be generated in
the microwave digestion, prior to programming and running the
equipment. This has been accomplished by measuring many of the
parameters necessary to allow calculation of the microwave power
absorption by the mineral acids. The basic theory will be presented
that describes the use of fundamental thermodynamic properties to
predict the microwave interactions with acids and samples.
The development of real—time monitors for temperature and pressure in
the microwave environment permits the investigation of closed vessel
digestion using microwave energy as the heat source. The digestion
uses Teflon (PFA) vessels. The microwave digestion technique has been
tested on all the major sample types including biological, botanical,
geological, alloy, and glassy samples and has demonstrated advantages
for each of these sample groups.
This work has opened many new applications. Becuase microwave
digestion is a well defined, precisely controlled system, it is
suitable for integration into an automated application. Acid
digestions have previously been too judgmental and variable for
automation to be practical. This microwave technique has direct
control of the power, acid temperature, and time of the digestion, and
has become structured to the point that it is possible to automate
sample decomposition prior to instrumental analysis. New microwave
transparent vessels, recently engineered , will permit the used of
higher temperatures and pressures, allowing the use of other reagents
and making this form of digestion even more efficient.
5-121
-------
CHARACTERIZATION OF MUNICIPAL AND HOUSEHOLD HAZARDOUS WASTES
Gary L. Mitchell, Vice President, Joyce K. Hargrove, Associate Staff
Scientist, SCS Engineers, Reston, Virginia; Anthony J. DiPuccio,
Project Director, SCS Engineers, Covington, Kentucky; Wayne Koser,
Resource Recovery Section, Michigan Department of Natural Resources,
Lansing, Michigan
ABSTRACT
The topics of this paper are the methodologies used to sample and
characterize municipal refuse and household hazardous wastes (HRW)
that are found in the municipal waste stream. Results of a major
characterization project will be used as examples. The paper will
provide specific examples of schedules, protocols, procedures and
problems. Special emphasis will be placed on HHW. Throughout, the
real world of waste characterization will be used to illustrate how
textbook guidance and local conditions are accommodated to yield
useful results. Statistical techniques used with the data will also
be presented.
The paper will be based on a project conducted by SCS Engineers for
the Michigan Department of Natural Resources. The project included
a year—long waste characterization effort in six counties throughout
Michigan. The paper will be divided into four major sub—divisions:
Introduction, Quantity, Composition, and Household Hazardous Wastes.
The Introduction will briefly describe the project and its
purposes. The general approach taken to coordinate with the host
counties, scheduling, logistics, personnel training, and quality
control will be presented.
The waste quantity aspects of the project will be presented briefly
in a discussion of the refuse weighing program. It will describe
the procedures and equipment used including color slides for
illustration. Field problems will be highlighted and a comparison
made between the program’s limited weighing effort and actual weight
data. Resulting per capita waste generation rates will also be
presented along with the statistical techniques applied to the data.
Waste composition will be a major element of the paper. The overall
characterization program will be described including both planning
and execution aspects. The waste categories utilized will be
presented along with a rationale for their selection and
identification of other categories that could be used for other
purposes. Actual field procedures for waste sorting will be
illustrated through the use of color slides.
Data collection methods will be summarized and the results
reported. Illustrations of tabulations will be included.
5—123
-------
Statistical procedures applied to the data will be identif led and
the rationale for their selection will be discussed. The percentage
of the sampled waste stream represented by each waste category was
of primary interest. Statistical parameters that were calculated
include standard division, upper and lower confidence limits, and
the application of the F—test to assess data variability.
Household hazardous wastes will be another major topic In the
paper. Project sponsors were Interested to know the types and
amounts of RUW in the wastes of each county. The definition of HHW
used was one developed by SCS in the preparation of the EPA report,
A Survey of Household Hazardous Wastes and Related Collection
Programs. The definition relates to environmental concerns rather
than to health and safety within the home and parallels the
definition of characteristic hazardous wastes. Field procedures
used to identify HHW will be described. Additionally, office review
of the field data by chemical and environmental engineers was used
to edit the field data. The types and quantities of HHW found will
be presented.
Two primary issues will be addressed in the paper. They are related
to overall waste consumption characterized and to household
hazardous wastes. The first issue is, “Can waste composition
estimates reasonably be based on national averages?” SCS experience
In tbls project and similar ones will be used to present and discuss
this lasue. The second Issue is, “What kinds and how much household
hazardous wastes are found In municipal refuse?”. Information
collected by S S during its previous EPA project on this topic plus
the field data from the Michigan waste characterization will provide
some of the most detailed data to answer that question.
INTRODUCTION
The purpose of this paper is to present methodologies used to sample
and characterize municipal refuse and household hazardous wastes
(1111W) that are found In the municipal waste stream. Result8 from a
major characterization project provide the basis for specific
examples of schedules, protocols, procedures and problems. The real
world of waste characterization will be used to illustrate how
textbook guidance and local conditions are accommodated to yield
useful results.
The overall goal of solid waste characterization is to determine,
with a specified degree of accuracy and precision, the quantity,
copoeltion andfor other chemical and physical characteristics of
solid waste. Specific objectives of waste characterization will
vary with individual projects; e.g. total waste quantities, amount
of newsprint, or presence of HHW.
5—124
-------
Quantity of solid waste generated in a particular locale is
important in projecting the life of a planned landfill. For
existing landfills this information can be used to reassess the life
of the landfill. Knowing the quantity of solid waste generated is
also useful in assessing the feasibility of various solid waste
management alternatives.
Solid waste composition is a major concern in resource recovery
facilities where composition of the waste significantly impacts the
practicality of such systems. Through waste characterization, key
factors such as problem wastes, and energy and moisture content of
wastes can be identified. Waste composition is also critical in
planning recycling and composting projects. In addition to
identifying desirable waste such as glass or aluminum for recycling
or biodegradable material for composting, waste characterization
will determine if there are sufficient quantities of the waste to
warrant the respective programs.
The goal of waste characterization also determines the degree of
accuracy and precision that is necessary. Accuracy requirements
depend more on the nature of the study than on the operation in
question. For example, waste characterization data from studies
performed elsewhere may be adequate for a preliminary feasibility
study, whereas a more extensive feasibility study or conceptual
design may require more accurate and precise estimates, based
in8tead on local conditions. During the final economic analysis of
resource recovery feasibility, accuracy and precision requirements
may ultimately be dictated by the buyers of the recovered material,
and by those providing financing for the project.
Waste characterization has typically been confined to: (1) use of
published data on solid waste quantity and composition, often
referred to as “national averages” or “typical characteristics” and
(2) limited field characterization involving little or no
statistical inference. The questionable results are often
compensated by conservative designs or revenue estimates which in
turn can impact operating efficiency and economic feasibility.
Thus, there is a need for development of statistically sound waste
characterization methods.
BACKGROUND
Marquette County’s waste stream assessment is one of the programs
undertaken by the Michigan Department of Natural Resources (MDNR) to
collect data and assess alternatives related to solid waste
management in the state. These alternatives include recycling,
composting, and energy recovery. Before the feasibility of waste
management alternatives can be determined, reliable information on
the quantity and composition of wastes generated is needed. In
addition, seasonal fluctuations, energy content, and the presence of
5-125
-------
household hazardous wastes are important. A.lthough limited waste
assessments were conducted in a few areas in Michigan prior to 1985,
the methodology used and data generated were found to be deficient
and not representative of waste streams. A definite need was found
to exist for thorough waste stream assessments in representative
areas across the state.
To meet this need, the Resource Recovery Section of the Community
Assistance Division within the Department of Natural Resources
initiated this project to collect data to determine the quantity and
composition of solid waste In six counties within the state. Funded
by the Clean Michigan Fund, the project provided specific
information about each of the six areas and prepared a guidebook
that could be used in other counties to conduct similar waste stream
assessment.
OBJECTIVES
The objective of this project was to collect and report current
field data regarding solid waste management in Marquette County.
The project focused on solid waste quantity and composition for
1986. AdditIonally, historical data on solid waste and Information
associated with Its management were obtained. The project was to
present a concise picture of the amount and composition of waste
generated and disposed within Marquette County.
Information obtained from this project may be used to:
o Assess the feasibility of resource recovery alternatives;
o Identify materials amenable to recycling, composting, and
incineration; and
o Assess the Impact of Michigan’s bottle bill on the
disposal of returnable containers.
Due to the growing concern of household hazardous wastes in solid
waste management, this project also Identified types and amounts of
HHW in wastes received at designated locations in Marquette County.
PROJECT APPROACH
The major disposal site in Marquette county is the Peninsula
Sanitation Transfer Station. Consequently, it was chosen as the
location for the weighing and waste composition program. Separate
assessments (one week each) were conducted during each of the four
seasons to Identify seasonal variations. Data were collected Monday
through Friday of each week. A second disposal site, the West
Marquette Landfill was added to the waste quantity investigation at
the request of the county. Limited field data were gathered at the
5—126
-------
landfill to better assess total solid waste quantities disposed
within the county, and to provide needed information on waste
generation from rural areas.
The weighing program involved obtaining a sample of vehicle weights
to determine the quantity of waste. Due to the small number of
vehicles using the transfer station, the weights of most of the
packers and other large vehicles using the site could be obtained
several times per week.
The composition and energy content aspects of the project were also
conducted at the transfer station. Randomly—selected loads of
refuse were sampled using the techniques described in the Waste
Stream Assessment Guidebook developed by SCS for MDNR. Trained
crews manually separated the refuse into designated categories. In
addition to typical waste categories, household hazardous wastes
were identified, segregated, and counted. Items considered as HEW
were completely described and their quantity estimated. Empty
containers were not considered HHW. Returnable beverage containers
were also noted. Soft drink and beer containers identif led as
returnable were counted by type of material: glass, aluminum, or
plastic.
One sample from each seasonal sort was collected for laboratory
analyses. These analyses included Btu content per unit weight,
moisture, ash, and volatiles. These analyses were chosen to provide
energy and residue information that will be useful in evaluating the
feasibility of recovering energy from the wastes.
WASTE QUANTITIES
Weighing Program
Two sites were chosen as the locations for the weighing/load count
portion of the solid waste assessment. The weighing/load count
program was conducted for one week, in each of the four seasons at
the transfer station. Load count information was collected for
three types of vehicles: packers, pickup trucks; and others. Load
weights were obtained for packers in all seasons and for others in
the summer and fall. Bad weather and operational problems did not
allow for weighing the other trucks in winter and spring, and small
vehicles were not weighed. Weighing was conducted with a portable
truck scale.
The vehicle to be weighed was driven onto a scale where each axle
was weighed separately and then added to calculate the gross weight
of the vehicle. After weighing, the vehicle unloaded at the
compactor. The empty vehicle was reweighed and the tare weight
recorded.
5—127
-------
All data were then entered into the Marquette County data base for
tabulation and calculations. The collection of weight data
underwent a two—stage review process. First, one (or both) SCS
field personnel checked the transfer of individual weights from
facility forms to the SCS forms. Second, the data received an
additional review off—site each day. This review included a
comparison with previously collected data, so that incorrectly
translated weights might be identified and corrected the following
day.
Statistical methods were applied to the waste quantity data.
Sumaations were made of the load counts and/or weight of waste
delivered by the three vehicle types (packers, pickups, and
others). The range of payloads was also identified for each
season. Likewise, the arithmetic mean (average) of payloads was
calculated for each type of vehicle during each season. Additional
statistical values calculated for weight data collected included:
standard deviation (SD), lover confidence limit (LCL), and upper
confidence limit (UCL).
Results of the load count and weighing program projected to an
annual basis are as follows:
Annual Loads Annual Ton
Peninsula Sanitation Transfer Station 8,567 12,984
West Marquette Landfill 15,379 5,070
The dramatic differences between the two sites are the number and
types of loads and the quantity of waste. The transfer station
serving the City of Marquette receives most of its waste from
packers. The ratio of packers to pickup is about 1 to 3. At the
landfill, about half of the waste is hauled by pickups, and the
packer—to—pickup ratio decreases to about 1 to 23.
SEASONAL VARIATIONS
Waste generation varies throughout the year. In general, less waste
than average is generated in winter months and more than average in
summer months. Typically, winter 18 the season of lowest waste
generation due to low levels of personal and commercial activity.
Summer and fall are often peak waste generation seasons related to
high levels of outdoor activities associated with landscaping,
travel, recreation, and construction. A graph of waste quantities
for the Peninsula Sanitation Transfer Station shown in Exhibit 1
indicates that winter was the season ‘when waste quantities from all
types of trucks were at their lowest.
5—128
-------
EXHIBIT 1
WASTE GENERATION AT THE PENINSULA SANITATION TRANSFER STATiON
Packers
Pk k-L s
Others
Season
1%6
Tons/Week
N)
LD
150
100
50
0
WINTER SPRING SUMMER FALL
- --
-------
The numbers of loads also changed from season to season, except for
packers which had a relatively constant frequency of just under 30
loads per week. Again, winter was the season with the fewest loads
from pickups and other trucks. Pickups peaked in spring and fall,
while spring was the peak season for loads from other trucks.
ESTIMATED GENERATION BATES
Per capita generation rates of solid waste are often estimated and
used for planning solid waste facilities such as transfer stations,
resource recovery facilities, and landfills. Development of
accurate generation rates depends upon reliable population and waste
quantity data. Neither are ever known with 100 percent accuracy for
coimiunities the size of Marquette County. However, reasonably
accurate information Is available to support the estimation of an
annual per capita solid waste generation rate.
Per capita waste generation rates were calculated for two areas in
Marquette County: the City of Marquette and the eight townships
served by the West Marquette Landfill. For the purposes of this
project, the population of the City of Marquette was estimated at
24,800. Annual waste quantities disposed at the transfer station
were estimated at 12,948 tons. Thus, the waste generation is
approxImately 2.9 pounds per person per day (365 days per year).
The eight townships served by the West Marquette Landfill have a
combined population of 12,100. The landfill received about 5,070
tons of waste annually, resulting in a generation rate of 2.3 pounds
per person per day (365 days per year).
WASTE COMPOSITION
Estimating the composition of waste in Marquette County Involved the
sampling of refuse delivered to the Peninsula Sanitation Transfer
Station and the manual sorting of each sample into numerous
components. Because the transfer station receives wastes
originating principally from the City of Marquette, the waste
composition data collected may be more representative of the urban
center than of the entire County. Sax pling and characterization
were performed during each of the four seasonal assessments.
Classification of waste was accomplished Monday through Friday of
each week using the same procedures and personnel from day to day
and season to season.
Vehicles were selected randomly by the SCS sort team leader for
sampling. The drivers of the vehicles were requested to leave
approximately one—quarter of each load near the sampling area as a
sample. The driver was interviewed as to type of waste (I.e.,
residential, commercial, or mixture), the area of the County where
the waste was collected, any unusual items in the load, and the
number of loads expected later that day. The SCS field crew
5-130
-------
manually collected 200 to 300 pounds of the load (the sample) and
placed it on one of two plastic sheets. SCS field personnel
supervised this operation to ensure the random selection of the
sample and the safety of the field crew.
Equipment used during the waste composition phase included a sort
box, plastic sheets, double—beam scale, 16 plastic trash cans,
shovel and push broom and safety equipment.
Refuse was manually separated into the following categories:
o Newsprint o Household Hazardous Wastes
o Corrugated Paper o Returnable Aluminum
o Office Paper o Returnable Glass
o Yard Waste o Returnable Plastic
o Textiles o Other Wastes
o Plastics
o Other Organics
o Glass
o Ferrous Metal
o Non—Ferrous Metal
o Other Inorganics
o Fines
The categories for sorting were identif led by the MDNR to best
evaluate the composition of solid wastes generated statewide. At
the beginning of the first seasonal assessment (the winter sort),
each category was explained and examples shown to the sort crew.
For consistency, the crew members became specialists in a discreet
number of categories. For example, one crew member would Identify
and remove from the sort box only newsprint, corrugated, office
paper, and textiles. Other members would specialize in other
categories. If questionable Items were found, the leader determined
the appropriate category.
The same procedures and crew members were used during the entire
waste composition program. This assured consistency from one sample
to the next and from season to season.
Actual waste sorting was methodical. Extremely large or heavy waste
Items were placed directly into the appropriate plastic sorting
container. Each item from the sample was placed in the sort box
until the box was full. Plastic bags of waste were torn open and
each Item of waste was manually segregated and placed in the proper
plastic container. These steps were repeated until the sample sort
was complete.
Upon completion of the sort, the plastic containers were moved to
the double—beam scale and weighed. The weights were recorded on a
field sort form. After weighing, each container was emptied at the
5—131
-------
working face of the landfill. For efficiency, this procedure was
performed concurrently with preparation of the area for the next
sample. This routine was repeated for up to 10 samples each day for
the five—day sort program during each season.
At the end of each day, weight percentages for each sample were
calculated and recorded on a daily summary sheet. The daily summary
included the mean, standard deviation and coefficient of variation.
These statistical parameters served as a basis for checking each
days’ sorting activities for completeness and accuracy. After
review of the data, it was entered into a portable computer and
stored for later evaluation.
A sIuhi Ary of the waste composition re8ults is shown In Exhibit 2.
Samples were taken only from packers. The average percentages by
weight of each category are shown for each seasonal sort. These
averages were themselves average to yield an estimated composition
of the City of Marquette’s residential and commercial waste. A
category termed “other wastes” was added to account for bulky,
infrequent items arriving at the disposal site, including white
goods (i.e., water heaters, stoves, freezers, etc.), tires and
furniture. White goods were not identified as a specific category
since no major appliances were recorded In the loads sampled and no
regular practice of separation occurred at the transfer station.
Seasonal Variations
Data In Exhibit 2 illustrate seasonal variations In waste
composition. For example, yard waste Increased to 15.3 percent in
spring compared to 0.3 percent in winter. The last column in
Exhibit 2 numerically compares seasonal variability. An analysis of
variance using the P—test was applied to waste composition data.
The F—test Ia a comparison of variances between data points In two
groups (e.g., the newsprint data for winter versus the newsprint
data for spring). Variances are calculated and checked against
values in F—test tables found in most statistics reference books.
These tables answered the question, “Do the data in these two groups
vary significantly from one another?” As with the LCL and UCL
calculations, the 95 percent confidence level was selected for the
F—teats.
HOUSEHOLD HAZARDOUS WASTES
Household hazardous wastes (HHW) was a sorting category during each
season. These wastes were identified using the definition developed
in the U.S. Environmental Protection Agency’s (EPA’s) first
nationwide survey of this topic conducted by SCS Engineers. Wastes
identified as originating from residential sources and likely to
fail one of the EPA’s hazardous waste characteristic tests
(ignitability, corrosivity, reactivity, or EP toxicity) were
5—132
-------
EXHIBIT 2
PENINSULA SANITATION TRANSFER STATION
WASTE COMPOSITION DATA FOR 1986
Waste Category
Percent by Weight
Varied
Significantly?
Winter
Spring
Summer
Fall
Average
Newsprint
9.2
4.9
5.3
65
6.5
No
Corrugated
2.5
7.0
7.0
10.1
6.7
Yes
Office Paper
3.1
2.7
2.9
3.5
3.0
No
Yard Waste
0.3
15.3
5.7
7.0
7.0
Yes
Textiles
2.1
2.5
1.8
3.8
2.6
Yes
Plastics
3.9
3.6
5.1
3.3
4.0
Yes
Other Organics
57.7
49.7
58.1
52.4
54.5
No
Subtotal
78.8
85.7
85.9
86.6
84.3
Glass
7.6
4.4
6.1
5.5
5.9
No
Ferrous Metal
0.6
4.3
4.7
4.3
.
3.5
NO
Non-Ferrous Metal
7.4
0.4
0.5
0.7
2.2
Yes
Other InorganIc
Fines
1.2
3.7
1.0
2.9
0.8
1.7
0.2
2.6
0.8
2.7
Yes
Yes
Subtotal
20.5
13.0
13.8
13.3
15.1
Household Hazardous
Wastes
Nil
Nil
Nil
Nil
Nil
N/A
Returnable Aluminum
Nil
Nil
Nil
Nil
Nil
N/A
Returnable Glass
Nil
Nil
Nil
Nil
Nil
N/A
Returnable Plastic
Nil
Nil
Nil
Nil
Nil
N/A
Other Waste*
0.7
1.3
0.3
0.1
0.6
Yes
Total
100.0
100.0
100.0
100.0
**
* Includes furniture, white goods, tires, and other large
or multi-material items.
** Sum of average composition may not be 100 percent due to rounding.
N/A — Not applicable.
5—133
-------
considered to be HHW. Through manual sort, items identified as
possible 1114W were completely described and their quantity
estimated. Empty containers of HUW products were not considered as
1111W for this study. The MDNR’s Hazardous Waste Division generally
agrees with this approach and it corresponds to the EPA ’s approach
regarding empty containers.
1411W were categorized as the following:
Household Cleaners: — drain cleaners, over cleaners, wood and
metal cleaners and poll8hes
kutomotive: — gasoline and oil additives, grease and
rust solvents, carburetor cleaners, air
conditioning refrigerant, starting fluids
Home Maintenance and
Improvement — paint thinners, strippers and removers,
adhesives and glues
Lawn and Garden — herbicides, pesticides, fungicides
Including wood preservatives
Note that not all examples of any generic type included in the list
are considered lifiW. For example, oven cleaners were included
because most would fail the corro8ivity teat. However, some
commercial oven cleaners would not fail this test.
Household hazardous wastes were a small portion of the waste on a
weight basis. The largest quantity of RHW identified occurred
during the fall sort. During that sampling week, 6.6 pounds of 1111W
were found in a total weekly sample weight of over 7,491 pounds.
Thus 1411W amounted to 0.09 percent (by weight) of the wastes sampled
that week.
The other seasonal sorts resulted in small percentages of 1111W. The
winter sort included 3.0 pounds of 1111W out of more than 3,784 pounds
of waste sampled, representing 0.08 percent of the waste. In the
spring, 2.1 pounds of 11 .11W were identified, representing 0.02 percent
of the week’s sample weight. During the summer sort, no 1111W were
found. The types and amounts of 1111W identified during each seasonal
sort are shown In Exhibit 3.
Overall, about 0.06 percent of the waste sampled during the
four—week survey was considered to be HHW. This projects to about
7.5 tons per year disposed at the transfer station. The annual
percentage and total projected weight of 1111W represent the 1111W that
routinely enter the solid waste stream. Other nationwide studies
report that many households tend to stockpile 1111W in basements and
garages. Thus, 1111W may be retained in homes for many years and not
5—134
-------
EXHiBIT 3
HOUSEHOLD HAZARDOUS WASTES IN MARQUETTE COUNTY
Hoe
Pie m l cnance/
Household Auto.otIve Lawn/Garden Improvement
Ct e aAet$ oz. Products Oz. Products Oz. Products Oz. MisceLlaneous Oi.
WINIER 1986
Oven Cleaner 4 E(ectr cal Contact 28
RathroO Cleaner 16 Cleaner
Season Totals: 20 28
total HNU (01): 48
Total NHW (Ibs): 3
i.— ’ total Weekly Sample Ut. (Ibs): 3,784
Percent P 1 1W: 0.08
SPRING
1986
Waste
Oil
32
3030
BulLet
2
Season
TotaLs:
32
2
Total 1111W tot):
totaL 1111W CIba):
total Weekly Sample Ut. (Ibs):
Percent 1111W:
34
2.1
9.544
0.02
sUMNER 1986
There was no household hazardous waste in the samples for the sorting season.
Season TotaLs: 0
Total 1111W (oz): 0
Total 11MW CIba): 0
Total Weekly Sample Ut. (Lbs):
Percent 1114W: 0.0
-------
C..)
EXHIBIT 3 (continued)
Household
Cleaners
Oz.
Aut oaotlv
Products
Oz.
Lawn/Garden
Products
Oz.
Maintenance!
laprov.iueflt
Products
Oz.
Miscellaneous
01.
FALL 1986
Disinfectant
Spray
1
west. Oil
Chs4n Lub
96
2
Hou .hold Sotvsnt
lust InhIbitor
2
4
Ned cIne
1
Season Totals:
1
98
6
Total NNW (Os):
Total NNW (lb.):
Total Weekly Sample
Percent 1 11 1W:
WI.
106
7
(lb.): 7 491
0.09
ANNUAL TOTALS
TotaL 1111W (05):
Total 11MW (lb.):
total Sample Wt. (lb.):
Percent 1111W:
188
12
2O 819
0.06
-------
disposed of on a regular basis. Periodic collection days, such as
those sponsored by grant programs (e.g., the Clean Michigan Fund),
civic organizations, or local governments, provide convenient and
beneficial outlets f or disposal.
The status of waste oil as a hazardous waste is indefinite.
Currently, MDNR does not consider waste oil to be hazardous unless
specific samples fail one of the characteristic tests. Waste oil
was Included as a 1111W for this study. If it Is not hazardous, the
quantity of HH in the waste samples will decrease to 3.8 pounds,
but the presence will decrease to 0.01 percent.
Exhibit 4 shows the results of various efforts to quantify HHW.
Each characterization study was conducted using a different
definition of 1111W and each had a different purpose. As a re8ult,
the data may not be comparable. However, there appears to be some
consistency in HHW representing a very small percentage of solid
waste.
FINDINGS M D CONCLUSIONS
The general approach described in this paper can be used in
estimating waste quantities, composition, and generation rates in
other counties. Although actual results from Marquette County may
be used in other counties, considerable caution should be
exercised. The basis for comparing Marquette County data to other
counties should include (1) population and population density, (2)
geographic location, (3) number and size of cities, (4) affluence of
the residents, and (5) the bases of the local economy. Other
considerations include seasonal population changes in vacation areas
or seasonal changes due to agricultural or other activities.
The following findings resulted from this waste characterization
effort:
o Approximately 12,948 tons of solid waste are disposed
annually at the Peninsula Sanitation Transfer Station,
resulting in a per capita waste generation of 2.9 pounds per
person per day (365 days per year).
o Quantities of waste at the transfer station are highest in
the spring and summer and lowest in the winter.
o ApproxImately 5,070 tons of waste are delivered annually to
the West Marquette Landfill, resulting In a waste generation
rate of 2.3 pounds per day per person (365 days per year) for
the rural townships served by this site. Little to no
commercial waste are disposed at the landfill.
5—137
-------
EXHIBIT 4
HHW Characterization Efforts
Date, Location
Conducted By Ouantitv Sorted HHW %
1979, Los Angeles 155 tons < 1% total HW
County, CA, Los Angeles HHW < 20% of HW
County Sanitation
District
1984-85, Los Angeles 15,000 tons scanned 0.00147% total HW
County, CA, Los Angeles for comercial size
County Sanitation 11W containers
District
1983, Albuquerque, Opinion survey of 0.5%
NM, Albuquerque 386 households
Environmental Health
and Energy Department
1986, Six MIchigan 149 tons 0.08%
Counties, SCS Engineers Range - 0 to 0.3%
5-138
-------
o Waste quantities at the landfill are highest in the summer
and fall and lowest in the winter.
o Household hazardous wastes comprise about 0.06 percent by
weight or about 7.5 tons per year projected for the Peninsula
Sanitation Transfer Station.
o Four samples of waste analyzed for energy content yielded an
average of 4,318 Btu per pound, essentially equal to the
national average of 4,500 Btu per pound.
o About 84 percent of the waste stream (the organics) is
combustible.
Comparison of these findings with national averages result in the
following conclusions:
o Waste generation in the City of Marquette falls within the
nationwide range of 2.5 to 3.5 pounds per person per day for
residential and commercial wastes.
o Marquette waste composition is similar to wastes generated in
other parts of the nation. Significant quantities of
recyclables and compostables are potentially available for
recovery, including newsprint, corrugated, glass, and yard
wastes. Organized recycling efforts should be focused in the
City of Marquette.
o Household hazardous wastes are present, but in lower
concentrations than in other locations nationwide.
5—139
-------
ENVIRONMENThL SCREENING IIEThODS FOR TOTAL ORGANIC HALOGENS
John Whitechurch, Shelley L. Smyers, Dohrmann, Santa Clara, California
ABSTRP CT
Approximately two—thirds of the priority pollutants identified by the
USEPA are halogenated organic compounds. In response, many analytical
techniques have been tested to characterize these compounds in water,
waste oils, and soils from spill sites and other waste sites. USEPA
Method 9020 documents a very useful and well—established means of
monitoring total organic halogens (TOX) in various water samples.
(The method is applicable to both identifiable and unidentifiable
halogenated organic compounds.) This paper demonstrates the utility
of changing analytical parameters of a TOX analyzer for these more
challenging sample types.
In cases where chromatographic techniques are requisites for organic
halogen identification (EPA Methods 601, 602, 624), time lost due to
detector overrange, trap saturation, and/or sample carryover in
chrornatographic columns becomes critical. For concentrated samples, a
fast and easy means of predicting sample and dilution volumes is in
high demand. A TOX analyzer is extremely useful in this regard.
For non—liquid samples such as organics, soils and solid wastes,
however, GC and GC/MS techniques can not be considered. For these
matrices, a fast, accurate and inexpensive method for measuring
organic halides is much needed. A TDX analyzer performs accurately
and reliably in this capacity.
Finally, when a high proportion of a liquid sample is not
chromatographable due primarily to the nature of the sample itself, GC
and GC/NS become expensive and unsuitable tools. Although a TOX
analyzer does not identify compounds or provide quantitation of
individual components, it does provide accurate quantitation of
organohalogens including those which cannot be chromatographed.
5—141
------- |